 |
|
| If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
|
|
| If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
|
| Datasheet File OCR Text: |
| Digital Voice Systems, Inc. The Speech Compression Specialists AMBE-2020TM Vocoder Chip User's Manual Version 4.0 January 2003 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 January, 2003 Copyright , 2003 Digital Voice Systems, Inc 234 Littleton Road Westford, MA 01886 This document may not, in whole or in part be copied, photocopied, reproduced, translated, or reduced to any electronic medium or machine readable form without prior consent in writing from Digital Voice Systems, Incorporated. Every effort has been made to ensure the accuracy of this manual. However, Digital Voice Systems, Inc. makes no warranties with respect to the documentation and disclaims any implied warranties of merchantability and fitness for a particular purpose. Digital Voice Systems, Inc. shall not be liable for any errors or for incidental or consequential damages in connection with the furnishing, performance, or use of this manual or the examples herein. This includes business interruption and/or other loss which may arise from the use of this product. The information in this document is subject to change without notice. Trademarks AMBE-2020TM Vocoder Chip is a registered trademark of Digital Voice Systems, Inc. Other product names mentioned may be trademarks or registered trademarks of their respective companies and are the sole property of their respective manufacturers. All Rights Reserved Data subject to change DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 2 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 AMBE-2020TM Vocoder Chip END USER License Agreement 1.0 Preliminary Statements and Definitions 1.1 "END USER" shall mean the person and/or organization to whom the AMBE-2020TM Vocoder Chip was delivered or provided to as specified in the purchase order or other documentation. In the event that the END USER transfers his rights under this license to a third party as specified in section 2.2, then this third party shall become an "END USER". 1.2 Digital Voice Systems, Inc. (DVSI) has developed a voice coding method and algorithm (the "Technology") based on the Advanced Multi-Band Excitation ("AMBE") voice coder. The technology codes speech at bit rates of 2.4 to 9.6 kilobits per second (kbps) including error correction bits. 1.3 "AMBE Voice Compression Software" shall mean the speech coding software and/or firmware integrated into the AMBE-2020TM Vocoder chip integrated circuit. 1.4 "Voice Codec" shall mean the AMBE-2020TM Vocoder Chip integrated circuit, the AMBE Voice Compression Software, firmware and associated documentation, including modifications, enhancements and extensions made by or for Digital Voice Systems, Inc. (DVSI) and including circuit diagrams, timing diagrams, logic diagrams, layouts, operating instructions and user manuals. 1.5 DVSI represents that it owns certain "Proprietary Rights" in the Technology and the AMBE Voice Compression Software, including patent rights in the Technology, and patent rights, copyrights, and trade secrets in the AMBE Voice Compression Software. 2.0 License Granted 2.1 Subject to the conditions herein and upon initial use of the AMBE-2020TM Vocoder Chip, DVSI hereby grants to END USER a non-exclusive, limited license to use the AMBE(R) Voice Compression Software in machine readable form solely on the AMBE-2020TM Vocoder Chip. Title to the AMBE(R) Voice Compression Software remains with DVSI. No license is granted for use of the AMBE(R) Voice Compression Software on other than the AMBE-2020TM Vocoder Chip. No license, right or interest in any trademark, trade name or service mark of DVSI is granted under this Agreement. 2.2 END USER shall not copy, extract, de-compile, reverse engineer or disassemble the AMBE(R) Voice Compression Software contained in the AMBE-2020TM Vocoder Chip. 3.0 Transfer of License 3.1 The END USER shall have the right to transfer the AMBE-2020TM Vocoder Chip and all rights under this Agreement to a third party by either (i) providing the third party with a copy of this Agreement or (ii) providing the third party with an agreement written by the END USER ( hereinafter "END USER Agreement") so long as the END USER Agreement is approved in writing by DVSI prior to transfer of the AMBE-2020TM Vocoder Chip. The END USER Agreement shall contain comparable provisions to those contained herein for protecting the Proprietary Information from disclosure by such third party. Third parties shall agree to accept all the terms and conditions under either Agreement or the END USER Agreement. 4.0 Term and Termination 4.1 This Agreement is effective upon initial delivery of the Voice Codec and shall remain in effect until terminated in accordance with this agreement. 4.2 This Agreement shall terminate automatically without notice from DVSI if END USER fails to comply with any of the material terms and conditions herein. END USER may terminate this Agreement at any time upon written notice to DVSI certifying that END USER has complied with the provisions of Section 3.3. 4.3 Upon termination of this Agreement for any reason, END USER shall: (i) return all AMBE2020TM Vocoder Chip purchased or acquired, or in Licensee's possession, to DVSI; (ii) have no further rights to any AMBE(R) Voice Compression Software or the Technology without a separate written license from DVSI; (iii) discontinue all use of the AMBE-2020TM Vocoder Chip; 5.0 Payments 5.1 In consideration of the materials provided as part of the Voice Codec, and in consideration of the license and rights in the AMBE Voice Compression Software granted by DVSI, and in consideration of DVSI's performance of its obligations hereunder, END USER agrees to pay to DVSI the fee specified in DVSI's invoice. 6.0 Proprietary Notices 12.0 Governing Law 6.1 END USER shall not remove any copyright or proprietary notice on the AMBE-2020TM Vocoder Chip or on the AMBE Voice Compression Software. 12.1 This Agreement is made under and shall be governed by and construed in accordance with the laws of the Commonwealth of Massachusetts, except that body of law governing conflicts of law. If any provision of this Agreement shall be held unenforceable by a court of competent jurisdiction, that provision shall be enforced to the maximum extent permissible, and the remaining provisions of this Agreement shall remain in full force and effect. 7.0 Proprietary Information 7.1 The parties agree that the AMBE Voice Compression Software shall be considered Proprietary Information. 7.2 Except as otherwise provided in this Agreement, END USER shall not use, disclose, make, or have made any copies of the Proprietary Information, in whole or in part, without the prior written consent of DVSI. 8.0 Limited Warranty 8.1 DVSI warrants the Voice Codec to be free from defects in materials and workmanship under normal use for a period of ninety (90) days from the date of delivery. 8.2 Except as stated in Section 7.1, the Voice Codec is provided "as is" without warranty of any kind. DVSI does not warrant, guarantee or make any representations regarding the use, or the results of the use, of the Voice Codec with respect to its correctness, accuracy, reliability, correctness or otherwise. The entire risk as to the results and performance of the Voice Codec is assumed by the END USER. After expiration of the warranty period, END USER, and not DVSI or its employees, assumes the entire cost of any servicing, repair, replacement, or correction of the Voice Codec. 8.3 DVSI represents that, to the best of its knowledge, it has the right to enter into this Agreement and to grant a license to use the AMBE Voice Compression Software to END USER. 8.4 Except as specifically set forth in this Section 7.0, DVSI makes no express or implied warranties including, without limitation, the warranties of merchantability or fitness for a particular purpose or arising from a course of dealing, usage or trade practice, with respect to the Voice Codec. Some states do not allow the exclusion of implied warranties, so the above exclusion may not apply to END USER. No oral or written information or advice given by DVSI or its employees shall create a warranty or in any way increase the scope of this warranty, and END USER may not rely on any such information or advice. The limited warranties under this section 7.0 give END USER specific legal rights, and END USER may have other rights which vary from state to state. 9.0 Limitation of Liability 9.1 In no event shall DVSI be liable for any special, incidental, indirect or consequential damages resulting from the use or performance of the Voice Codec whether based on an action in contract, tort (including negligence) or otherwise (including, without limitation, damages for loss of business profits, business interruption, and loss of business information), even if DVSI or any DVSI representative has been advised of the possibility of such damages. 9.2 Because some states do not allow the exclusion or limitation of liability for consequential or incidental damages, the above limitations may not apply to END USER. 9.3 DVSI's maximum liability for damages arising under this Agreement shall be limited to 20% (twenty percent) of the fees paid by END USER for the particular Voice Codec which caused the damages or that is the subject matter of, or is directly related to, the cause of action. 10.0 Taxes 10.1 All payments required under Section 4.0 or otherwise under this Agreement are exclusive of taxes and END USER agrees to bear and be responsible for the payment of all such taxes (except for taxes based upon DVSI's income) including, but not limited to, all sales, use, rental receipt, personal property or other taxes which may be levied or assessed in connection with this Agreement. 11.0 Export 11.1 United States export laws and regulations prohibit the exportation of certain products or technical data received from DVSI under this Agreement to certain countries except under a special validated license. As of November 30, 1999 the restricted countries are: Libya, Cuba, North Korea, Iraq, Serbia, Taliban in Afghanistan, Sudan, Burma, Yugoslavia and Iran. The END USER hereby gives its assurance to DVSI that it will not knowingly, unless prior authorization is obtained from the appropriate U.S. export authority, export or re-export, directly or indirectly to any of the restricted countries any products or technical data received from DVSI under this Agreement in violation of said United States Export Laws and Regulations. DVSI neither represents that a license is not required nor that, if required, it will be issued by the U.S. Department of Commerce. Licensee shall assume complete and sole responsibility for obtaining any licenses required for export purposes. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 3 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 1. Product Introduction 1.1 1.2 1.3 2 General Information Advantages Typical Applications 6 6 6 7 8 8 8 8 9 10 11 12 12 12 13 15 16 17 17 17 18 19 19 19 21 21 22 23 24 26 26 26 27 27 28 28 29 AMBE-2020TM Application Design Overview 2.1 Basic Operation 2.2 Initial Design Considerations 2.2.1 A/D - D/A Overview 2.2.2 Vocoder Front End Requirements 2.2.3 Channel Interface Overview 2.2.4 Speech and FEC Rate Selection Overview 3 Hardware Information 3.1 Special Handling Instructions 3.1.1 Storage 3.2 3.3 3.4 Pin Descriptions Clock and Reset Timing Associated Chip Delay 3.5 Crystal / Oscillator Usage 3.5.1 TTL Clock Source 3.5.2 Crystal Oscillator 3.6 3.7 3.8 3.9 4 Package Description Normal Operating Conditions Absolute Maximum Ratings Electrical Characteristics and Requirements Channel Interface 4.1 4.2 Overview Serial Configuration Selection 4.3 Channel Serial Mode 4.3.1 Low Level Timing for Passive and Active Serial Mode 5 Channel Data Format 5.1 Framed Format 5.2 Framed Input Format 5.2.1 Framed Input: Word 0 : Header 5.2.2 Framed Input: Word 1 : Power Control ID 5.2.3 Framed Input: Word 1 : Control Word 1 5.2.4 Framed Input: Words 2-6 : Rate Information 5.2.5 Framed Input: Word 7 : Unused in Input DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 4 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 5.2.6 5.2.7 5.2.8 5.2.9 5.2.10 Framed Input: Word 8 : Unused in Input Framed Input: Word 9 : Unused in Input Framed Input: Word 10 : DTMF Control Framed Input: Word 11 : Control Word 2 Framed Input: Words 12-23 : Channel Data 29 29 30 30 31 31 32 32 32 33 33 33 33 33 35 36 36 36 37 38 38 38 39 41 41 41 42 43 43 43 43 44 47 47 48 52 54 57 58 5.3 Framed Output Format 5.3.1 Framed Output: Word 0 : Header 5.3.2 Framed Output: Word 1 : Power Control ID 5.3.3 Framed Output: Word 1 : Control Word 1 5.3.4 Framed Output: Words 2-6 : Rate Information 5.3.5 Framed Output: Word 7 : Bit Error Rate 5.3.6 Framed Output: Word 8 : Soft Decision Distance 5.3.7 Framed Output: Word 9 : Detected Bit Errors 5.3.8 Framed Output: Word 10 : DTMF Control 5.3.9 Framed Output: Word 11: Control Word 2 5.3.10 Framed Output: Words 12-23 : Channel Data 5.4 Unframed Serial Format 5.4.1 Unframed Serial Output Format 5.4.2 Unframed Serial Input Format 6 A/D-D-A Interface 6.1 6.2 6.3 7 A/D-D/A Overview Configuring the A/D-D/A Interface using CODEC_SEL[1-0] Low Level A/D-D/A Timing Special Functions 7.1 7.2 7.3 7.4 Hardware vs. Software Selection Note Coding Rate Selection Voice Activation Detection (VAD), Comfort Noise Insertion (CNI) Dual Tone Multiple Frequency, Detection and Generation 7.5 Normal Power and Power Saving Modes 7.5.1 Standard Sleep Mode 7.5.2 Power Down 7.6 8 Slip Enable Appendices 8.1 8.2 8.3 8.4 8.5 8.6 Example: AD73311 Usage Example: Texas Instruments TLV320AIC10 Usage Configuring the AD73311 for 3-Volt Operation Interfacing to the Texas Instruments PCM3500 Codec Expanded Tone Detection and Generation Soft Decision Decoding DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 5 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 1. Product Introduction 1.1 General Information Digital Voice Systems Inc.'s AMBE-2020TM Vocoder Chip is an extremely flexible, high-performance, single chip, speech compression coder. It provides superior voice quality at low data rates. It provides a real-time, half-duplex implementation of the standard-setting AMBE voice compression software algorithm. DVSI's patented AMBE voice compression technology has been proven to outperform CELP, RELP, VSELP, MELP, ECELP, MP-MLQ, LPC-10, and other competitive technologies. Numerous evaluations have shown its ability to provide performance equal to today's digital cellular systems at under half the data rate. The AMBE voice compression algorithm is used in applications throughout the world, including the next generation of digital mobile communication systems. The AMBE-2020TM Vocoder chip provides a high degree of flexibility in selecting the speech and FEC (Forward Error Correction) data rates. The user can separately select these parameters in 50 bps increments for total rates from 2.0 kbps to 9.6 kbps. Typically for higher error rate channels, the user will apportion a greater percentage of the total bit rate to FEC coding. The AMBE-2020TM voice coder maintains natural voice quality and speech intelligibility at rates as low as 2.0 kbits/sec. The AMBE algorithm's low complexity allows it to be fully integrated into a low cost, low power integrated circuit, the AMBE2020TM Vocoder Chip. The AMBE-2020TM Vocoder Chip offers similar features to DVSI's AMBE-1000TM Vocoder Chip allowing it to be incorporated into systems already designed for the AMBE-1000TM and is interoperable with other DVSI products. The AMBE2020TM Vocoder Chip delivers improved performance and enhanced modes such as 4.0 kbps toll quality speech and convolutional FEC coding. Along with these enhancements the AMBE-2020TM Vocoder Chip employs a control interface along with the variable data rates and FEC selections. 1.2 Advantages * * * * * * * * Superior Voice Quality Low Cost No External Memory Required Robust to Bit Errors & Background Noise Variable Data Rates - 2.0 kbps to 9.6 kbps Variable FEC Rates - 50 bps to 7.2 kbps Very Low Power (65mW @ 3.3V, 0.11mW Deep Sleep) Compact Single Chip Solution: 100 pin TQFP High Quality Low Data Rate Speech Coding DVSI's Half Duplex AMBE Voice Coder Supports Data Rates of 2.0 kbps to 9.6 kbps in 50 bps increments User Selectable Forward Error Correction rates Viterbi Decoder (rate 1/4 or more) 16 Level Soft Decision Decoding Voice Activity Detection (VAD) / Comfort Noise Insertion Single and Dual Tone (DTMF) Detection and Generation Power-Down Mode Minimal algorithmic processing delay DTMF detection and regeneration with North American call progress tones Features * * * * * * * * * * * DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 6 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 1.3 Typical Applications * * * * * * * Satellite Communications Digital Mobile Radio Secure Communications Cellular Telephony and PCS Voice Multiplexing Voice Mail Multimedia Applications DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 7 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 2 AMBE-2020TM Application Design Overview 2.1 Basic Operation In its simplest model, the AMBE-2020TM can be viewed as two separate components, the Encoder and the Decoder. The Encoder receives an 8kHz sampled stream of speech data (16-bit linear, 8-bit Alaw, or 8-bit ulaw) and outputs a stream of channel data at the desired rate. Conversely the Decoder receives a stream of channel data and synthesizes a stream of speech data. The timing for the interfaces for the AMBE-2020TM Encoder and Decoder are fully asynchronous. Figure 2-A Basic Operation AMBE-2020 8kHz Speech Data AMBE-2020 Compressed Data @ 2000-9600bps Encoder Decoder 8kHz Speech Data Typically the speech interface is an external A/D-D/A chip. The format of the incoming and outgoing speech data streams are coupled, that is to say they must be the same format (16-bit linear, 8-bit Alaw, or 8-bit law). The channel interface is commonly (but not limited to) an 8 or 16 bit microprocessor or other suitable `glue logic' hardware capable of performing the rudimentary formatting functions between the AMBE-2020TM channel format and the format of the system channel. Optional functions of the chip, such as voice activation/detection, power mode control, data/FEC rate selection, etc. are controlled either through hardware control pins (see Section 5) and/or through the decoder command interface (see Section 4) Data sent into the decoder for function control purposes is distinguished from the data to be decoded into speech through a channel format which is described in Section 4. 2.2 Initial Design Considerations * * * Choice of A/D-D/A chip. Choice of Channel Interface. Speech and FEC Rates. Some of the initial design considerations the application engineer will face are the following: 2.2.1 A/D - D/A Overview The choice of the A/D-D/A chip is critical to designing a system with superior voice quality. Given that Alaw and law companding chips are already incorporating some compression to reduce the number of bits per sample, it is recommended that, when possible, a 16-bit linear device be used for maximum voice quality. When choosing a device, pay particular attention to Signal to Noise ratios and Frequency Responses of any filters that may be present on the analog front end of these chips. The Alaw and law interfaces are also provided for the design engineer who is trying to fit to pre-existing conditions or is under other cost type restraints. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 8 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 2.2.2 Vocoder Front End Requirements In order to ensure proper performance from the voice coder, it is necessary for the vocoder front end to meet a set of minimum performance requirements. For the purposes of this section the vocoder front end is considered to be the total combined response between microphone/speaker and the digital PCM interface to the vocoder software, as shown in Figure 2-B. This includes any analog electronics plus the A-to-D and D-to-A converters as well as any digital filtering performed prior to the voice encoder or after the voice decoder. Microphone Vocoder Front End Analog Input Gain Analog Input Filter Digital Input Gain Analog Input Reference Pt. A-to-D Convertor Digital Output Reference Pt. AMBE+TM Voice Encoder Speaker Analog Output Gain Analog Output Filter D-to-A Convertor Digital Output Gain Analog Output Reference Pt. Digital Input Reference Pt. AMBE+TM Voice Decoder Analog Speech Digital Speech (8 kHz sampling) Figure 2 - B. Vocoder Front End The AMBE+TM voice encoder and decoder operate with unity (i.e. 0 dB) gain. Consequently the analog input and output gain elements shown in Figure 2 are only used to match the sensitivity of the microphone and speaker with the A-to-D converters and D-to-A converters, respectively. It is recommended that the analog input gain be set such that the RMS speech level under nominal input conditions is 25 dB below the saturation point of the A-to-D converter (+3 dBm0). This level, which equates to -22 dBm0, is designed to provide sufficient margin to prevent the peaks of the speech waveform from being clipped by the Ato-D converter. +2 dB -1 dB -2 dB -18 dB -35 dB -40 dB 0 200 400 3000 3400 4000 4600 freq (Hz) 8000 Figure 2 - C. Front End Input Filter Mask DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 9 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 The voice coder interface requires the A-to-D and D-to-A converters to operate at an 8 kHz sampling rate (i.e. a sampling period of 125 microseconds) at the digital input/output reference points. This requirement necessitates the use of analog filters at both the input and output to eliminate any frequency components above the Nyquist frequency (4 kHz). The recommended input filter mask is shown in Figure 2 - C, and the recommended output filter mask is shown in Figure 2 - D. For proper operation, the shaded zone of the respective figure should bound the frequency response of the front-end input and output. +2 dB -1 dB -2 dB -18 dB -40 dB -60 dB 0 200 3000 3400 4000 4600 freq (Hz) 8000 Figure 2 - D. Front End Output Filter Mask This document assumes that the A-to-D converter produces digital samples where the maximum digital input level (+3 dBm0) is defined to be +/- 32767, and similarly, that the maximum digital output level of the D-to-A converter occurs at the same digital level of +/- 32767. If a converter is used which does not meet these assumptions then the digital gain elements shown in Figure 2 should be adjusted appropriately. Note that these assumptions are automatically satisfied if 16 bit linear A-to-D and D-to-A converters are used, in which case the digital gain elements should be set to unity gain. Also note that the vocoder requires that any companding which is applied by the A-to-D converter (i.e. alaw or ulaw) should be removed prior to speech encoding. Similarly any companding used by the D-to-A converter must be applied after speech decoding. An additional recommendation addresses the maximum noise level measured at the output reference points shown in Figure 2-B with the corresponding inputs set to zero. DVSI recommends that the noise level for both directions should not exceed -60 dBm0 with no corresponding input. In addition the isolation from cross talk (or echo) from the output to the input should exceed 45 dB which can be achieved via either passive (electrical and/or acoustic design) or active (echo cancellation and/or suppression) means. 2.2.3 Channel Interface Overview The channel interface is meant to be flexible to allow for easy integration with the system under design. The basic hardware unit of the interface is a serial port. The serial port can run in passive or active modes. In passive mode, all of the channel interface control signals are inputs to the AMBE2020TM chip. In active mode, only the TX_DATA_STRB is an output from the AMBE2020TM chip. All other signals are inputs. Under normal operation, every 20ms, the encoder outputs a frame of coded bits, and the decoder needs to be delivered a frame of coded bits. There is some formatting of the data for both the encoder and the decoder. The primary purpose of the formatting is to provide alignment information for the encoded bit stream. The data has two formats, Framed and Unframed. Serial mode can run in either Framed or Unframed mode. The Framed and Unframed modes are explained in full detail in Section 4, but essentially the two formats are trying to achieve the same function, to provide positional information regarding the outgoing and incoming coded data streams. In Framed mode each 20msecs of output data from the encoder is preceded by a known structure (each packet corresponds to 20ms of speech data input into the encoder). This structure also embeds some status type flags, meant for local control DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 10 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 purposes, within it. The only data from the Framed format that is typically sent across the transmission channel under design are the actual encoded bits at the desired rate. In Framed mode, it is the responsibility of the designed system to pass enough information along with the encoded bits such that the Framed format needed by the decoder can be reconstructed on the other side. This extra information, or overhead, is going to be very specific to the system under design, but at a minimum needs to pass enough information to reliably reconstruct the 20msec frame structure at the other end for the decoder. In Unframed mode the data coming out of the encoder can be thought of as a continuous stream of voice data with the framing information embedded within the encoded bits. One advantage of this type of set-up is that the system does not have to add any bandwidth for overhead to the channel. The disadvantage is that the decoder needs 10-12 incoming frames in order to gain synchronization with the data stream before it can properly synthesize the speech waveform. Also, the Unframed mode only commits a single bit per frame to maintain data alignment. In higher error rate channels the performance will be improved by adding more bits per frame to the alignment information (which is more easily performed when using Framed mode). Additional flexibility is given to the channel interface to the encoder and decoder by allowing the AMBE-2020TM Vocoder Chip to run in Passive or Active modes. In Passive mode, data strobes are provided by an external source, while in Active mode, data strobes are provided by the AMBE-2020TM Vocoder Chip. Serial interfaces can be run in Passive or Active modes. See Section 4 for full details and timing for both Framed and Unframed data. 2.2.4 Speech and FEC Rate Selection Overview The total coded bit rate is the sum of two components, the Speech Data bit rate and the Forward Error Correction (FEC) Data bit rate. The addition of FEC data to the speech data allows the decoder to be able to correct a limited amount of errors within each frame should they arrive corrupted. If the channel is expected to have more errors then more bits should be dedicated to FEC. At the same time, voice quality will increase if the number of speech bits is increased. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 11 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 3 Hardware Information 3.1 Special Handling Instructions The AMBE-2020TM uses the TM320LC541B-66 core. For greater details on handling, electrical, packaging, or timing specs please refer to the TMS320-C54x data sheet at http://www.ti.com/sc/pcsheets/sprs039c/sprs039c.pdf. Although the AMBE-2020TM Vocoder Chip incorporates input protection circuitry, to avoid damage from the accumulation of a static charge, industry standard electrostatic discharge precautions and procedures must be employed during handling and mounting. The 100 pin TQFP package design of the AMBE-2020TM Vocoder Chip allows it to be mounted by infrared reflow, vaporphase reflow or equivalent processes. The peak package body temperature must not exceed 220C. The AMBE-2020TM Vocoder Chip requires baking before mounting, if any of the following conditions exist: * * * * Humidity indicator card (included in packaging) shows exposure to > 20 % when read at 23C + 5C Devices were not shipped in a package designated as "moisture controlled." Not mounted within 168 hours of receipt, at factory conditions of <30C and <60% RH If the device has not been stored at < 20% RH DVSI's recommended bake out procedures: * * For low-temperature device containers: 192 hours at 40C + 5C / -0C and < 5% Relative Humidity For high-temperature device containers : 24 hours at 125C + 5C. 3.1.1 Storage To insure maximum shelf life in long term storage, AMBE-2020TM Vocoder Chips should be kept in a moisture controlled package at <40C and <90% Relative Humidity DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 12 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 3.2 Pin Descriptions Pin Descriptive Name CHAN_SEL1 Pin Direction Input Notes Channel Interface Selection Pins: Use these bits to select the channel interface type (framed, unframed active, passive) according to Table 4-A. See full description in section 4.2. Used with CHAN_SEL1 to select channel operation mode A/D-D/A Select Pins see Table 6-A to select the interface. Pin Number 77 75 85 84 74 73 72 71 70 86 CHAN_SEL0 CODEC_SEL1 CODEC_SEL0 RATE_SEL4 RATE_SEL3 RATE_SEL2 RATE_SEL1 RATE_SEL0 VAD_EN Input Input Input Input Input Input Input Input Input Coding Rate Select Pins: Use these bits to select the voice and FEC rates according to section 7.2. The coding rates are also selectable using the Control Word interface described in section 5.2.4. Voice Activation Detection Enable Pin. Active HIGH. See Section 7.3. VAD can also be enabled/disabled using the Control Word interface as described in section 5.2.9. Encoder/Decoder switch. Switch on to enable Encoder and disable Decoder (must be used in conjunction with Encoder Enable function in Framed Input. See Section 5.2.9). Switch off to enable Decoder and disable Encoder. Standard Sleep Enable Pin. Active HIGH. See Section 7.5.1. Slip Control Enable Pin. Active HIGH. See Section 7.6. Clock Input 1. 16.384 MHz input. See Section 3.3 Output from internal oscillator for the crystal. If the internal oscillator is not used this pin should be unconnected. AMBE-2020TM Reset pin. Active LOW. See Section 3.3 Encode Packet Ready: Following a reset, this signal will have a high to low transition to indicate the first packet is ready. The next packet will be ready approximately 20 msec later. See Note 1. Soft decision decoding enable. When high it enables 4 bit soft decision error decoder. Keep low when not in use. Baud Rate Selector for unframed serial mode See Table 4-B. Channel Receive Data to AMBE-2020TM Channel Transmit Data from AMBE-2020TM Channel Receive Clock Channel Transmit Clock Channel Transmit Data Strobe Channel Receive Data Strobe Frame synchronization pulse for A/D data. Should be connected to CODEC_TX_STRB Frame synchronization pulse for D/A data. Should be connected to CODEC_RX_STRB PCM Data from A/D Converter to AMBE-2020TM 24 83 82 68 67 69 20 ENCODER_EN SLEEP_EN SLIP_EN X2/CLKIN X1 RESETN EPR Input Input Input Input Input Input Output 79 80 81 32 42 28 34 38 30 29 37 31 SOFT_EN BAUD_SEL0 BAUD_SEL1 CHAN_RX_DATA CHAN_TX_DATA CHAN_RX_CLK CHAN_TX_CLK CHAN_TX_STRB CHAN_RX_STRB CODEC_RX_STRB CODEC_TX_STRB CODEC_RX_DATA Input Input Input Input Output Input Input I/O Input Input Input Input DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 13 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Pin Number 41 27 33 51 8,11,12,23, 36,39,44,45, 46,47,48, 49,54,57, 64,76,87,90 1,9,10,25, 26,35,40,50, 52,53,56, 63,65,88,89 2,3,4,5,6,7, 13,14,15,16, 17,18,19,21, 22,43,55,58, 59,60,61,62, 66,78,91,92, 93,94,95,96, 97,98,99, 100 Pin Descriptive Name CODEC_TX_DATA CODEC_RX_CLK CODEC_TX_CLK CLOCK_MODE Pin Direction Output Input Input Input Notes PCM Data from AMBE-2020TM to D/A Converter A/D Serial clock. Should be connected to CODEC_TX_CLK D/A Serial clock Should be connected to CODEC_RX_CLK If high enables crystal oscillator option for clock source. If low then external oscillator option is selected. See Section 3.5 for details. VDD Power Supply Voltage GND Power Ground No Connection These pins must remain unconnected Note 1: The AMBE-2020TM expects an encoder packet to be read approximately every 20 msec. Following the initial reset, wait for EPR to go low and read the initial packet (t0). 20 msec later, the next packet (t1) should be ready. For packet t1 and all following packets use the procedure below: 1) Wait for slightly less than 20 msec. 2) Assert CHAN_TX_STB and read word on CHAN_TX_DATA. 3) If transmitted word not 0x13EC, discard it and repeat step 2. 4) If transmitted word 0x13EC, read 23 more words (rest of packet). The EPR signal should not be used as an interrupt. The EPR is only valid for the first high to low transition. A new packet should be ready every 20 msec after the initial EPR high to low transition. A packet read should take place every 20 msec. If there is a delay in the read (i.e. a packet is missed), it is recommended that the device be reset. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 14 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 3.3 Clock and Reset Timing To reset the AMBE-2020TM chip, the reset signal must be held low for a minimum of 50 s. The recovery time from reset is approximately 95 msec. In other words, 95 msec after the rising edge of the reset signal the AMBE-2020TM starts processing PCM samples. The first packet will be ready after 252 PCM samples are read. Figure 3-A X2/CLKIN and CLKOUT Timing Diagram tw(CIH) tc(CI) tw(CIL) tr(CI) tf(CI) ~ ~ X2/CLKIN ~ ~ td(CIH-CO) tp Unstable tc(CO) tw(COH) tf(CO) tw(COL) tr(CO) ~ ~ CLKOUT Table 3-A X2/CLKIN and CLKOUT Timing Parameters Reference Parameter Cycle time, X2/CLKIN Integer PLL multiplier N (N=4) tc(CI) Fall time, X2/CLKIN tf(CL) Rise time, X2/CLKIN tr(CL) Pulse duration, X2/CLKIN low tw(CIL) Pulse duration, X2/CLKIN high tw(CIH) Transitory phase, PLL lock-up time tP Cycle time, CLKOUT (typical is tc(CI)/4) tc(CO) Delay time, X2/CLKIN high/low to CLKOUT high/low td(CIH-CO) Fall/Rise time, CLKOUT (typical is 2 ns) tf(CO)/tr(CO) Pulse duration, CLKOUT low tw(COL) Pulse duration, CLKOUT high tw(COH) CLKOUT is shown for reference only it is not connected. H=7.629 ns Min 20 Max 400 4 4 Units ns ns ns ns ns s ns ns ns ns ~~ ~~ 6 6 50 15 4 H-4 H-4 16 H H * * DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 15 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 X2/CLKIN ~ ~ tsu(RS) tw(RSL) RESETN ~ ~ tsu(INT) th(RS) CLKOUT ~ ~ Figure 3-B Hardware Reset Timing Diagram Figure 3-C Hardware Reset Timing DiagramTable 3-B Reset Timing Parameters Reference th(RS) tw(RSL) tsu(RS) tsu(INT) Parameter Hold time, RS after CLKOUT low Pulse duration, RS low Setup time, RS before X2/CLKIN low Setup time, INTn, NMI, RS before CLKOUT low Min 0 50 5 10 Max Units ns s ns ns * CLKOUT is shown for reference only it is not connected. 3.4 Associated Chip Delay The associated delay due to the coding/decoding algorithm is shown below Encoder Delay Algorithmic Delay = 32 ms Encoder Processing Delay = 11.5 ms Decoder Delay Algorithmic Delay = 10 ms Decoder Processing Delay = 7.5 ms Total Delay = 32 ms + 11.5 ms + 1 ms* +10 ms +7.5 ms = 62 ms Frame Processing Delay = 11.5 ms (encoder) + 1 ms* + 7.5 ms = 20 ms * 1ms of idle time between encode and decode sequence. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 16 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 3.5 Crystal / Oscillator Usage The AMBE-2020TM Vocoder Chip has an input clock frequency of 16.384 MHz. Two options are outlined below in providing this signal. The CLOCK_MODE pin 51 must be set appropriately for the option used. The following points should be noted when designing any printed circuit board layout: * * Keep the crystal and external capacitors as close to the X2/CLKIN and X1 pins as possible to minimize board stray capacitance. Keep X2/CLKIN and X1 away from high frequency digital traces (example CLKOUT) to avoid coupling. 3.5.1 TTL Clock Source Figure 3-D X2/CLKIN and X1 with TTL Clock Source 16.384 MHz TTL/CMOS Clock Source X2/CLKIN (pin 68) If CLOCK_MODE pin is low then a TTL/CMOS source is used as the clock input. Connect X2/CLKIN and X1 as follows: AMBE-2020 Unconnected X1 (pin 67) 3.5.2 Crystal Oscillator The Crystal Oscillator option is selected with CLOCK_MODE pin set to a high level. To use the crystal oscillator, connect the crystal across X2/CLKIN and X1 along with one external capacitor from each of these pins to ground. Recommended values for C1 and C2 is 10 pF. Figure 3-E X2/CLKIN and X1 with Crystal Oscillator C1 = 10 pF X2/CLKIN (pin 68) 16.384 MHz AMBE-2020 X1 (pin 67) C2 = 10 pF DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 17 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 3.6 Package Description 100 pin TQFP (Thin Quad Flat Pack) All Dimensions are in millimeters Figure 3-F Package Dimensions 16.20 / 15.80 SQ 14.20 / 13.80 SQ 75 76 12.00 TYP 51 50 100 1 12 All Around o 26 25 1.40/1.60 mm 0.5 mm 0.17 / 0.27 1.35 /1.45 1.60 MAX 0.05 0.25 Detail Gage Plane 0.13NOM 0 - 7 MAX 0.45 / 0.75 Not Drawn to Scale DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 18 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 3.7 Normal Operating Conditions Table 3-C Normal Operating Conditions Normal Operating Conditions Operating Voltage Operating Case Temperature Range Storage Temperature Range 3.3V -40C to 100C -55C to 150C 3.8 Absolute Maximum Ratings Stresses in excess of the Absolute Maximum Ratings can cause permanent damage to the device. These are absolute stress ratings only. Functional operation of the device is not implied at these or any other conditions in excess of those given in the operational sections of the data sheet. Exposure to Absolute Maximum Ratings for extended periods can aversely affect device reliability. Table 3-D Absolute Maximum Ratings Absolute Maximum Ratings Voltage Range on any Pin with Respect to Ground -0.3V to 4.6V 3.9 Electrical Characteristics and Requirements Table 3-E Recommended Operating Conditions Parameter DVDD VSS VIH VIL IOH IOL Device Supply Voltage Supply Voltage, GND High-level input voltage, I/O Low-level input voltage High-level output current Low-level output current Min 3 2.5 -0.3 Nom 3.3 0 Max 3.6 DVDD + 0.3 0.8 -300 1.5 Unit V V V V A mA DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 19 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Table 3-F Electrical Characteristics over Recommended Operating Case Temperature Range (Unless Otherwise Noted) Parameter VOH High-level output voltage VOL Low-level output voltage IX Input current in high impedance (VI = VSS to VDD) CI Input capacitance Test Conditions VDD = 3.3 V, IOH = MAX IOL = MAX VDD = MAX, VI = VSS to VDD MIN 2.4 TYP MAX Unit V 0.4 -10 10 10 10 V A pF pF CO Output capacitance All values are typical unless otherwise specified. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 20 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 4 4.1 Channel Interface Overview The Channel Interface is the general term used for the interface for the compressed bits coming from the encoder and the compressed bits going to the decoder. This same interface is also used to output status information from the encoder and decoder such as whether a DTMF tone has just been detected in the speech input, or whether the decoder has detected and synthesized a frame of silence. Additionally, this interface is used to perform more complex control operations on both the encoder and decoder (usually at start-up). These control functions include speech data and FEC rate control. It is important to realize that not all data being output from the AMBE-2020TM is intended for transmission over the channel. Status type of data is typically only useful at the `local' end. In most voice transmission systems, the actual encoded bits are extracted from the channel formatting, combined into the systems transmission stream, sent over the transmission path, extracted from the transmission path at the receiving end, and reassembled into the AMBE-2020TM 's channel format for synthesis by the decoder. Figure 4-A Channel Interface Overview Typical Voice Frame Output From Encoder Overhead Data Voice Data (40 to 192 bits) (Header, Status) Transmission Channel System extracts relevant Voice Data bits and formats them for transmission over Channel System Overhead Voice Data (40 to 192 bits) System extracts relevant Voice Data bits and formats them for Input into the Decoder adding Header and Control Information Typical Voice Frame Input to the Decoder Overhead Data Voice Data (40 to 192 bits) (Header, Control) DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 21 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 4.2 Serial Configuration Selection The hardware interface to the Channel Interface is configured as a serial interface based exclusively on the hardware settings of CHAN_SEL[1-0]. See Table 4-A. Table 4-A Channel Interface Selection Table Interface Select Pins CHAN_SEL1 (pin 77) 0 0 1 1 BAUD_SEL1 (pin 81) 0 0 1 1 Port Type Active Framed Active Unframed Passive Framed Passive Unframed Table 4-B Unframed Bit per Word Selection Table Interface Select Pins Number of Voice Data Bits per Word 1 2 3 4 All transfers occur through a serial port. The serial port inputs and outputs a 16 bit word for every write and read strobe signal respectively. Serial mode can be framed or unframed. Within the unframed mode, the data is input and output in 16 bits words still but with only 1 to 4 voice data bits carried within each word. These four configurations can be seen in Table 4-B. See section 4.3 for all the details on the serial interface. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com BAUD_SEL0 (pin 80) 0 1 0 1 CHAN_SEL0 (pin 75) 0 1 0 1 Page 22 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 4.3 Channel Serial Mode The signals in Table 4-C make up the serial channel interface. The serial channel mode transfers data in and out of the AMBE-2020TM using 16 bit words on the two data lines CHAN_RX_DATA and CHAN_TX_DATA. The selection of the framed or unframed format of this data is made using information in Table 4-A. Table 4-C Channel Serial Interface Pin Descriptions Pin Symbol EPR Pin Direction Out Pin Number 20 Description Encoder Packet Ready: Following a reset, this output signal will have a high to low transition to indicate that the encoder has a frame of data to output. The next packet will be ready approximately 20 msec later. See Note 1. Serial Data Input : 16 bits of channel data are input on CHAN_RX_DATA, synchronous to CHAN_RX_CLK, with each CHAN_RX_STRB pulse. Serial Input Clock : In coordination with CHAN_RX_STRB, CHAN_RX_DATA is latched by the AMBE-2020TM on the falling edges of CHAN_RX_CLK. Input (Write) Data Strobe : This signal indicates to the AMBE-2020TM when the data on CHAN_RX_DATA will be latched by CHAN_RX_CLK. See figure 4-B. Serial Data Output : 16 bits of channel data are output on CHAN_TX_DATA, synchronous to CHAN_TX_CLK, with each CHAN_TX_STRB pulse. Serial Output Clock : In coordination with CHAN_TX_STRB, the data on CHAN_TX_DATA is output by the AMBE-2020TM on the rising edges of CHAN_TX_CLK. Output (Read) Data Strobe : This signal indicates to the AMBE-2020TM when to bring the data to the CHAN_TX_DATA pin. See figure 4-B. CHAN_RX_DATA CHAN_RX_CLK CHAN_RX_STRB CHAN_TX_DATA In In In Out 32 28 30 42 CHAN_TX_CLK In 34 CHAN_TX_STRB Active Output/ Passive Input 38 Note 1:The AMBE-2020TM expects an encoder packet to be read approximately every 20 msec. Following the initial reset, wait for EPR to go low and read the initial packet (t0). 20 msec later, the next packet (t1) should be ready. For packet t1 and all following packets use the procedure below: 1) Wait for slightly less than 20 msec. 2) Assert CHAN_TX_STB and read word on CHAN_TX_DATA. 3) If transmitted word not 0x13EC, discard it and repeat step 2. 4) If transmitted word 0x13EC, read 23 more words (rest of packet). A new packet should be ready every 20 msec after the initial EPR high to low transition. A packet read should take place every 20 msec. If there is a delay in the read (i.e. a packet is missed), it is recommended that the device be reset. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 23 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 4.3.1 Low Level Timing for Passive and Active Serial Mode Figure 4-B Low Level Timing for Passive and Active Serial Mode tc(SCK) tw(SCK) CHAN_RX_CLK th(FSR) tsu(FSR) CHAN_RX_STRB tw(SCK) tsu(DR) th(DR) CHAN_RX_DATA MSB 1 tr(SCK) tf(SCK) ~~ ~ ~ 2 15 LSB 16 tc(SCK) CHAN_TX_CLK td(FSX) th(FSX) CHAN_TX_STRB th(FSX)H tw(SCK) td(DX) tr(SCK) tdis(DX) CHAN_TX_DATA ~~ ~ ~ ~ tw(SCK) tf(SCK) th(DX) MSB 1 2 15 LSB 16 Table 4-D Switching Characteristics Over Recommended Operating Conditions for Serial Port Receive Reference Parameter Hold time, FSR after CLKR falling edge Hold time, DR after CLKR falling edge 3.3 Volts Min Max 6 6 Units ns ns th(FSR) th(DR) DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 24 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Table 4-E Switching Characteristics Over Recommended Operating Conditions for Serial Port Receive Reference Parameter Delay time, DX valid after CLKX rising (Passive Mode) Cycle time, serial port clock 3.3 Volts Min Max 25 Units ns td(DX) * ns 6H tc(SCK) 6 ns Fall time, serial port clock tf(SCK) 6 ns Rise time, serial port clock tr(SCK) Pulse duration, serial port clock low/ high 3H ns tw(SCK) ns Setup time, FSR before CLKR falling edge 6 tsu(FSR) Setup time, DR before CLKR falling edge 6 ns tsu(DR) 15 ns Delay time, CLKX rising to FSX td(FSX) 15 ns Delay time, CLKX rising to DX (Active Mode) td(DX) Hold time, FSX after CLKX falling edge 6 ns th(FSX) 2H-5 ns Hold time, FSX after CLKX rising edge th(FSX)H 20 ns Disable time, CLKX rising to DX tdis(DX) Hold time, DX valid after CLKX rising edge -5 ns th(DX) Note: H = 7.629 ns; however, do not operate serial clocks any faster than 2.048 MHz. Thus tc(SCK) should be a minimum of 488.3 nsec. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 25 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 5 Channel Data Format The channel interface is responsible for outputting the compressed data from the encoder and inputting compressed data to the decoder. In addition to these most basic functions the channel interface is also capable of reporting certain events, such as the detection of a DTMF tone. The channel interface can also control certain selectable functions of the AMBE-2020TM, such as the voice coding rate. This chapter will describe how the AMBE-2020TM uses the channel interface to multiplex these capabilities. There are two formats to the data, Framed and Unframed, both of which operate in serial mode. Generally speaking the Unframed mode is used only when the connection between the AMBE-2020TM and the channel under design is relatively direct, and the designer wants to simplify the extraction of the relevant voice data. In this mode configuration is accomplished using hardwired pins. In most cases, when a controller is present between the AMBE-2020TM and the channel, the system designer will find that using the Framed format is more flexible. 5.1 Framed Format The Framed format is a 24 by sixteen-bit word format for a total of 48 bytes or 384 bits. Every 20 milliseconds either the encoder outputs 24 sixteen-bit words, or the decoder expects to receive 24 words depending which mode the chip is in (encode only or decode only). The format of the input and output frames are detailed below. The first 12 sixteen bit words are made up of header, ID and status or control information. The remaining 12 sixteen bit words make up the encoded data bit field. These 12 words, or 192 bits, will be fully populated with relevant voice data only when the AMBE-2020TM is operating in a 9600bps mode (9600 bits/sec / 50 frames/sec = 192 bits/frame). Otherwise, when the data rate is less than 9600bps, the coded voice bits are filled starting from the MSB of the first word in the field, leaving any unused bits as zeros. It is important to note here that even when the AMBE-2020TMis operating at less than 9600bps, all 384 bits of the Framed format (including any unused trailing zeros) must be transferred out of the encoder and into the decoder. 5.2 Framed Input Format In Figure 5-A we see the format of the Framed input. Keep in mind that even though the channel data in this Framed input format is closely associated with the decoder, the control information will apply to both encoder and decoder functions. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 26 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Figure 5-A Basic Framed Input Format Word # 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 (12) 16 bit words of overhead (192 bits) Description Header always set to 0x13EC Power Control ID (8bits) Control Word 1 (8 bits) - see Table 5-B Rate info 0 Rate info 1 See Tables 5-C and 5-D Rate info 2 Rate info 3 Rate info 4 Unused in Input Unused in Input Unused in Input DTMF Control - see Tables 5-E and 5-F Control Word 2 - see Table 5-G Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data 24 sixteen-bit words = 48 bytes = 384 bits 20 ms frame 5.2.1 Framed Input: Word 0 : Header The decoder uses the header information to synchronize with the beginning of each 20 millisecond frame. this 16 bit word MUST be 0x13EC. 5.2.2 Framed Input: Word 1 : Power Control ID Set the 8-bit Power Control ID field to 0x00 for normal use. For Power Down Mode, set this value to 0x55. This causes the AMBE-2020TM to enter low power mode. To exit low power mode, the device must be reset through the hardware. Table 5-A Framed Input : Power Control ID Values Summary ID Type Description 0x00 (12) 16 bit words of data (192 bits) Voice Data This ID value instructs AMBE-2020TM to operate in normal fashion 0x55 Low Power Mode When this mode is activated the AMBE-2020TM Vocoder Chip will go into a mode which conserves power, where no voice packets are being processed. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 27 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 5.2.3 Framed Input: Word 1 : Control Word 1 Table 5-B Control Word 1 Format 7: MSB Lost Frame Indicator 6 Unused 5 Control Word 1 - 8-bits (See Table 1-F) 4 3 2 Unused Unused Unused 1 CNI 0 : LSB Unused Use the 8-bit control word to set various functions. Unused Lost Frame Indicator : Setting the Lost Frame Indicator bit to a 1 will cause the AMBE-2020TM decoder to construct the voice frame using the parameters from the previous frame. This is an effective way to mask the effects of short periods of data loss. This bit should be set by the user when channel data is lost or corrupted. It works by replacing the frame of corrupted data with the previous, intact, frame. Comfort Noise Insertion (CNI) : Setting the CNI bit will cause the decoder to output a frame of comfort noise. This bit is used with systems that are capable of discontinuous transmission (DTX). 5.2.4 Framed Input: Words 2-6 : Rate Information Rate Info 0, Rate Info 1, Rate Info 2, Rate Info 3, Rate Info 4 The initial rate of the AMBE-2020TM is set through the hardware pins RATE_SEL[4-0] (see Section 7.2 and Tables 7-A and 7-B) after resetting the device. The coding rate can be modified for both the encoder and the decoder by sending a framed packet to the input channel interface. The Rate Selector control field (described in section 5.2.9) determines where to apply the rate change. Particular attention must be paid to this field when using the BER reporting fields (section 5.3). The AMBE-2020TM used these five words to set the source and FEC coding rates. Tables 5-C and 5-D list predefined value for various source and FEC rates. These are only a representation of the most common rates that are requested. Please contact DVSI for additional rate information if the desired rates are not listed. The software configurations in Table 5-C are compatible for use with the AMBE-1000TM (using AMBETM technology). If compatibility is not an issue, use the software codes in Table 5-D to select speech and FEC rates to optimize use of the AMBE-2020TM (using AMBE+TM technology). DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 28 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Table 5-C Rate Selection Using Rate Info 0-4, compatible w/ AMBE-1000TM (AMBE) Rate Info 0 0x9030 0x902f 0x9348 0x9243 0xab50 0x934b 0xab60 0xab5b 0x9348 0x923e 0xab53 0xab58 0xbf9b 0xab5d 0xbfc0 0xab16 Rate Info 1 0x0000 0x0000 0x0000 0x0080 0x0000 0x0080 0x0000 0x0080 0x2030 0x2800 0x2c00 0x3000 0x0080 0x3400 0x0000 0xe400 Rate Info 2 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 Rate Info 3 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 Rate Info 4 0x4330 0x6930 0x6f48 0x5348 0x3950 0x3950 0x7960 0x6860 0x7060 0x7460 0x5680 0x4490 0x49a0 0x31a0 0x72c0 0x67c0 Speech Rate (bps) 2400 2350 3600 3350 4000 3750 4800 4550 3600 3100 4150 4400 7750 4650 9600 4850 FEC Rate (bps) 0 50 0 250 0 250 0 250 1200 1700 2250 2800 250 3350 0 4750 Total Rate (bps) 2400 3600 4000 4800 6400 7200 8000 9600 Table 5-D Rate Selection Using Rate Info 0-4, AMBE-2020TM only (AMBE+) Rate Info 0 0x0028 0x5048 0x5250 0x1030 0x5360 0x5250 0x5048 0x1030 0x6b80 0x5250 0x5258 0x7fa0 0x5250 0x7fc0 0x5048 0x1030 Rate Info 1 0x0000 0x0000 0x0000 0x0001 0x0000 0x2010 0x0001 0x0005 0x0000 0x0001 0x0009 0x0000 0x0005 0x0000 0x000e 0x000e Rate Info 2 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x180c 0x0000 0x0000 0x1e0c 0x0000 0x2010 0x0000 0x4010 0x681a Rate Info 3 0x0000 0x0000 0x0000 0x341a 0x0000 0x0000 0x2412 0x3018 0x0000 0x542a 0x4127 0x0000 0x6834 0x0000 0x6a2e 0x511b Rate Info 4 0x6428 0x3948 0x4150 0x6750 0x6c60 0x7460 0x6860 0x7360 0x6c80 0x5280 0x7390 0x52a0 0x72a0 0x69c0 0x65c0 0x76c0 Speech Rate (bps) 2000 3600 4000 2400 4800 4000 3600 2400 6400 4000 4400 8000 4000 9600 3600 2400 FEC Rate (bps) 0 0 0 1600 0 800 1200 2400 0 2400 2800 0 4000 0 6000 7200 Total Rate (bps) 2000 3600 4000 4800 6400 7200 8000 9600 5.2.5 Framed Input: Word 7 : Unused in Input Should be set to 0x0000 5.2.6 Framed Input: Word 8 : Unused in Input Should be set to 0x0000 5.2.7 Framed Input: Word 9 : Unused in Input Should be set to 0x0000 DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 29 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 5.2.8 Framed Input: Word 10 : DTMF Control Use this word to set DTMF tones. See Table 5-F for a list of tones and their corresponding values. Table 5-E DTMF Control Format DTMF Control - 16-bits 15: MSB 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0: LSB DTMF Amplitude DTMF Digit Detect/ Generate Table 5-F DTMF Codes for Digit Detect/ Generate DTMF Code 0x80 0x84 0x88 0x81 0x85 0x89 0x82 0x86 0x8A 0x87 0x83 0x8B 0x8C 0x8D 0x8E 0x8F 0xff DTMF Digit 1 2 3 4 5 6 7 8 9 0 * # A B C D Inactive Frequency 1 (Hz) 1209 1336 1477 1209 1336 1477 1209 1336 1477 1336 1209 1477 1633 1633 1633 1633 N/A Frequency 2 (Hz) 697 697 697 770 770 770 852 852 852 941 941 941 697 770 852 941 N/A To generate no tones, set the DTMF Code to 0xff. Dial, ring, and busy tones are standard North American call progress tones. An expanded list of tones and values can be found in the Appendices, Section 8.4. DTMF Amplitude The DTMF Amplitude runs from 3 to -60 dBm0. The value is a signed byte. So 0x03 = 3, 0x00 = 0, 0xff = -1, 0xc4 = -60. 5.2.9 Framed Input: Word 11 : Control Word 2 Table 5-G Control Word 2 Format Control_Word 2 - 16-bits (See Table 1-F) Bit 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 Decoder Output Volume Control Unused 0 Unused 0 VAD Unused 0 SL EE RIS DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 30 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Rate Information Selector (RIS) : Use these 2 bits to select which part(s) of the vocoder will be affected by the rate selection words. Table 5-H Rate Information Selection Codes value 0x0 0x1 0x2 0x3 Area Controlled Encoder and Decoder Encoder only Decoder only Neither Encoder nor Decoder Encoder Enable (EE): Set this bit to 1 AND set the Encoder Enable Pin (ENCODER_EN, pin 24, see section 3.2) to 1 to enable the encoder and disable the decoder on a frame by frame basis. Set this bit to 0 to enable the decoder and disable the encoder. This must be set to the desired function in every packet sent to continue use. Sleep (SL): Set this bit to 1 to enter Sleep mode on a frame by frame basis. Sleep is a low power mode, not to be confused with Power Down Mode. The sleep function must be enabled in every packet to continue in sleep mode. Set this bit to 0 to exit Sleep mode. VAD: In order to set the Voice Activity Detector on, set the VAD to 1. To turn the Voice Activity Detector off, set the VAD to 0. 5.2.10 Framed Input: Words 12-23 : Channel Data This is the field that contains the actual coded bits. Input of the data begins with the MSB of the first word in this field and continues through with the final bit output being the LSB of the final word. If the data rate selected is less than 9600bps then the unused bits in each frame are zero and populate the end of the field. As is noted in the Channel Interface definitions, these unused bits must still be clocked out of the AMBE-2020TM. The packet must always consist of 24 word 5.3 Framed Output Format The format for framed output data is shown in Figure 5-B. Only the bits in the Channel data field are transmitted along with framing information (data used to locate the start of each frame for proper reconstruction at the decoder) over the channel. The first 192 bits provide overhead information which is sometimes useful to the host but is generally not transmitted over the channel. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 31 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Figure 5-B Basic Framed Output Format Word # 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 (12) 16 bit words of overhead (192 bits) Description Header always set to 0x13ec Power Control ID (8bits) Control Word 1 (8 bits) - see Table 5-J Rate info 0 Rate info 1 Rate info 2 Rate info 3 Rate info 4 Bit Error Rate Soft Decision Distance Detected Bit Errors In Current Frame DTMF Control - see Table 5-L Control Word 2 - see Table 5-N Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data Channel Data 24 sixteen-bit words = 48 bytes = 384 bits 20 ms frame 5.3.1 Framed Output: Word 0 : Header The header is a 16 bit word that begins each valid frame corresponding to 20 milliseconds of speech. This field will always be 0x13EC. 5.3.2 5.3.3 Framed Output: Word 1 : Power Control ID Framed Output: Word 1 : Control Word 1 Table 5-J Control Word 1 Format 7: MSB Unused 6 Unused 5 Control Word 1 - 8-bits (See Table 1-A) 4 3 2 Decoder Silence Detect Unused Unused 1 Encoder Silence Detect 0: LSB Encoder DTMF Detect The encoder will always use 0x00 in the 8-bit field of an output frame. This 8-bit control word indicates the activity of various functions. (12) 16 bit words of data (192 bits) Decoder Frame Repeat Decoder Frame Repeat: When the Decoder Frame Repeat flag is set to 1, the decoder is reporting that the last frame decoded was a repeat of the previous frame. Decoder Silence Detect: When the Decoder Silence Detect flag is set to 1, the decoder is reporting that the last frame decoded was a comfort noise frame. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 32 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Encoder DTMF Detect: The Encoder DTMF Detected Flag will be set to a 1 when the encoder detects a DTMF tone. Encoder Silence Detect: The Encoder Silence Detected Flag will be set to 1 when no voice activity is detected. The Silence Detect option is controlled by and can be disabled by the VAD as described in section 5.2.9. 5.3.4 Framed Output: Words 2-6 : Rate Information Rate Info 0, Rate Info 1, Rate Info 2, Rate Info 3, Rate Info 4. Words 2-6 in the packet indicate the rate at which the AMBE-2020TM encoder is operating . These words are output. See tables 5-C and 5-D for corresponding values. 5.3.5 Framed Output: Word 7 : Bit Error Rate This status field is used for the decoder to report bit error information. The 16 bit number output is used to compute the BER using the following calculation % BER = (Word 7) / (32768) x 100 This status field works in conjunction with the RIS bits in control word 2. Every time the RIS bits are set to any value other than 0x3, this field is reset. The BER is calculated only when using FEC provided by the AMBE-2020TM. 5.3.6 Framed Output: Word 8 : Soft Decision Distance This status field is used in conjunction with hardware pin 79 SOFT_EN. This field is used to report the soft decision distance for the most recent frame. This status field works in conjunction with the RIS bits in control word 2. Every time the RIS bits are set to any value other than 0x3, this field is reset. The Soft decision Distance is calculated only when using Soft Decision decoding provided by the AMBE-2020TM. 5.3.7 Framed Output: Word 9 : Detected Bit Errors This status field is used to report the number of detected bit errors in the current frame. This status field works in conjunction with the RIS bits in control word 2. Every time the RIS bits are set to any value other than 0x3, this field is reset. The number of bit errors is calculated only when using FEC provided by the AMBE-2020TM. 5.3.8 Framed Output: Word 10 : DTMF Control This word corresponds to the DTMF Detection capabilities of the vocoder. It uniquely identifies specific tones recognized by the encoder. See table 5-M for a list of tones and their corresponding values. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 33 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Table 5-K DTMF Tone Detection Parameters DTMF Tone Detection Requirement Minimum Input Level Minimum Signal to Noise Distortion ratio Minimum Frequency Tolerance Maximum Frequency Tolerance Value -25 dBm0 Description An input signal shall not be rejected as a DTMF tone if its amplitude is greater than -25 dBm0 (maximum sinusoid dBm0 is defined as+3.17 dBm0). In order for an input signal to correspond to a valid DTMF tone, the ratio of inband to out-ofband energy must be greater than 15dB. Inband energy is defined to be the energy in frequency components within 3.5% of the two frequencies defined by the DTMF frequencies. Out-of-band energy is defined to be the total energy minus in the inband energy. An input signal shall not be rejected as a DTMF tone if both of its principal frequency components are within 1.5% of the frequencies needed for the DTMF tone. An input signal shall be rejected as a DTMF tone if either of its principal frequency components are outside 3.5% of the frequencies needed for the DTMF tone. An input signal does not correspond to a valid DTMF tone if the energy contained within the low frequency band is more than 10 dB greater than the energy contained in the high frequency band. An input signal shall not be rejected as a DTMF if energy contained within the low frequency band is less than 8 dB greater than the energy contained in the high frequency band. Each low and high frequency band is limited to 3.5% of the frequencies needed for the DTMF tone. An input signal does not correspond to a valid DTMF tone if the energy contained within the high frequency band is more than 10 dB greater than the energy contained in the low frequency band. An input signal shall not be rejected as a DTMF if energy contained within the high frequency band is less than 4 dB greater than the energy contained in the low frequency band. Each low and high frequency band is limited to 3.5% of the frequencies needed for the DTMF tone. An input signal shall not be rejected as a DTMF tone as long as its time duration is greater than 45 mS. In addition a minimum of two frames will be transmitted of the DTMF tone if a valid tone is detected. The duration of a tone is defined by the points at which the envelope is 20 dB below its peak value. 15 dB 1.5% 3.5% Normal Twist Range 8-10 dB Reverse Twist Range 4-10 dB Minimum Tone Duration 45 mS Table 5-L DTMF Control Format DTMF Control - 8-bits (See Table 1-F) 15: MSB 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0: LSB DTMF Amplitude DTMF Digit Detect/ Generate DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 34 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Table 5-M DTMF Codes for Digit Detect/ Generate DTMF Code 0x80 0x84 0x88 0x81 0x85 0x89 0x82 0x86 0x8A 0x87 0x83 0x8B 0x8C 0x8D 0x8E 0x8F 0xff DTMF Digit 1 2 3 4 5 6 7 8 9 0 * # A B C D No Tone Frequency 1 (Hz) 1209 1336 1477 1209 1336 1477 1209 1336 1477 1336 1209 1477 1633 1633 1633 1633 N/A Frequency 2 (Hz) 697 697 697 770 770 770 852 852 852 941 941 941 697 770 852 941 N/A If no tones are detected , the DTMF code is set to 0xFF. Dial, ring, and busy tones are standard North American call progress tones. An expanded list of tones and values can be found in the Appendices, Section 8.4. DTMF Amplitude The DTMF Amplitude runs from 3 to -60 dBm0. The value is a signed byte. So 0x03 = 3, 0x00 = 0, 0xff = -1, 0xc4 = -60. 5.3.9 Framed Output: Word 11: Control Word 2 Table 5-N Control Word 2 Format Control Word 2 - 16-bits (See Table 1-A) Bit 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 Decoder Output Volume Control Unused 0 Unused Unused Unused 0 0 0 SL EE RIS Rate Information Selector (RIS) : This field is used to indicate which part of the device is affected by words 2-6, the rate control words. Table 5-O Rate Information Selection Codes value 0x0 0x1 0x2 0x3 Area Controlled Encoder and Decoder Encoder only Decoder only Neither Encoder nor Decoder DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 35 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Encoder Enable (EE): A 1 in this field AND the Encoder Enable Pin (ENCODER_EN, pin 24, see section 3.2) set to 1 indicates the encoder is enabled and the decoder is disabled. A 0 in this field indicates the decoder is enabled and the encoder is disabled. Sleep (SL): A 1 in this field indicates the device has been put into sleep mode. Decoder Output Volume Control: Indicates the current decoder volume. 5.3.10 Framed Output: Words 12-23 : Channel Data This is the field that contains the actual coded bits. Output of the data begins with the MSB of the first word in this field and continues through with the final bit output being the LSB of the final word. If the data rate selected is less than 9600bps then the unused bits in each frame are zero and populate the end of the field. As is noted in the Channel Interface definitions, these unused bits must still be clocked out of the AMBE-2020TM. The packet must always consist of 24 words. 5.4 Unframed Serial Format The Unframed Format for the channel data is useful for applications which desire minimal glue logic between the AMBE2020TM and the channel hardware. The use of minimal hardware in place of a microcontroller can be realized using this data format. Another distinct difference in this data format is that framing information (data which carries the positional information relating to the coded bits) is embedded into the data stream itself. Using this data format, the system designer need only transfer the coded data itself. A single bit from each frame is `borrowed' from the voice data to embed the framing information. Keep in mind that this `borrowed' bit reduces the effective voice coding rate quality by 50 bits per second. For example, a system with no FEC running at 2450 bps in Unframed mode will sound equivalent to one running at 2400 bps in Framed mode. The designer should also be aware that it takes approximately 15 frames (300 milliseconds) for the decoder to attain synchronization with the incoming stream before it can output synthesized speech. Systems which are attempting to save power by shutting down transmission during periods of silence, and then resuming during periods of speech can not handle this 300 millisecond delay for each synchronization, and thus should use Framed mode with a more sophisticated framing method. The 16 bit per word format, pictured in Section 5.2, is maintained in this mode but only a fraction of the full 16 bits is used to transfer the coded data. The user selects whether 1, 2, 3 or 4 bits will be transferred in each word based on pins BAUD_SEL[0:1] Table 4-B. IMPORTANT: The voice coding data rate selected must be evenly divisible by the number of voice data bits per word selected. 5.4.1 Unframed Serial Output Format The Unframed output format contains 1 to 4 bits within each 16 bits serial output word. The formats which contain more than one bit each word the MSB of the data bits is considered first in the transmission. In Unframed mode, only the coded voice data bits are output. None of the superfluous information that exists in framed mode is available in this mode. The number of words that need to be transferred out of the encoder for each 20 millisecond frame will be the number of bits per frame divided by the number of bits per word. So a system coding at 4800 bps with 3 bits per word will need to read 32 ([4800 /50] /3 = 32) words each frame. The serial clock rate is computed by1/ [20 msec /(32 x 16)] = 25.6 kHz. If passive unframed mode is selected, the data strobe will be computed by 1/[20msec / 32] = 1.6 kHz. Table 5-P Unframed Serial Output Data Format Bits per Word See Table 4-A 1 bit per Word Format 2 bits per Word Format 3 bits per Word Format 4 bits per Word Format Data 15 14 13 12 11 10 9 8 7 Unused 6 5 4 3 2 1 0 D msb D msb D msb D msb 0 D D D 0 0 D D 0 0 0 D 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 36 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 5.4.2 Unframed Serial Input Format The Unframed Input format contains 1 to 4 bits within each 16 bits serial output word. For the formats which contain more than one bit each word the MSB of the data bits is considered first in the transmission. In Unframed mode, the header data from Framed mode is dropped and each 16 bit write contains 1 to 4 coded voice data bits. The number of words that need to be transferred into the decoder for each 20 millisecond frame will be the number of bits per frame divided by the number of bits per word. So a system coding at 4800 bps with 3 bits per word will need to write exactly 32 ( [4800 /50] /3 = 32) words each frame. Table 5-Q Unframed Serial Input Data Format Bits per Word See Table 4-A 1 bit per Word Format 2 bits per Word Format 3 bits per Word Format 4 bits per Word Format Data 15 14 13 12 11 Control Offset 10 9 8 7 6 5 Control Data 4 3 2 1 0 D msb D msb D msb D msb 0 D D D 0 0 D D 0 0 0 D 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 37 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 6 A/D-D-A Interface 6.1 A/D-D/A Overview The interface from the analog world of speech to the AMBE-2020TM is typically an A/D-D/A chip. Selection of the A/DD/A chip should be made carefully, with a preference given to 16 bit linear devices. Additionally, consideration should be given for signal to noise ratios and filtering characteristics typically built into many such devices. Generally speaking, the flatter the frequency response over the voice spectrum (20-4000Hz) the better the overall system will sound. The AMBE-2020TM Vocoder Chip operates with a speech data sample rate of 8kHz for both the A/D and D/A interfaces. This 8kHz data is input and output using a serial port on the AMBE-2020TM. In order to simplify the process of configuring the interface to the A/D-D/A chip, a number of preset configurations can be chosen through the CODEC_SEL[1-0] pins shown in Table 6-A. These preset configurations control companding and sampling rate as well as the sequence of programming words for the programmable devices, specifically the AD73311. 6.2 Configuring the A/D-D/A Interface using CODEC_SEL[1-0] In order to simplify the process of configuring the A/D-D/A interface certain preset configurations are available to the user. Selection of these preset modes is made through the 2 hardware pins CODEC_SEL[1-0]. In Table 6-A, the 2 digit binary value for CODEC_SEL[1-0] corresponds to the levels present on the hardware pins, with a 0 corresponding to GND, and a 1 corresponding to VCC. Table 6-A CODEC_SEL[1-0] : A/D-D/A Hardware Configuration Values A/D-D/A Type Generic 16 bit Linear 8Khz Analog Devices 73311 32kHz Generic law 8kHz Generic Alaw 8kHz CODEC_SEL[1-0] pins 85,84 00b 01b 10b 11b DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 38 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 6.3 Low Level A/D-D/A Timing Figure 6-A Low Level Timing for A/D-D/A Low Level Timing for A/D-D/A tc(SCK) tw(SCK) CODEC_RX_CLK th(FSR) tsu(FSR) CODEC_RX_STRB tsu(DR) th(DR) CODEC_RX_DATA MSB 1 tw(SCK) tf(SCK) ~ tr(SCK) ~~ ~ 2 7/15 LSB 8/16 tc(SCK) CODEC_TX_CLK td(FSX) th(FSX) CODEC_TX_STRB th(FSX)H tw(SCK) td(DX) ~ tw(SCK) tf(SCK) tr(SCK) tdis(DX) CODEC_TX_DATA ~~ ~ th(DX) MSB 1 2 7/15 LSB 8/16 Table 6-B Switching Characteristics Over Recommended Operating Conditions for Serial Port Receive Reference Parameter Hold time, FSR after CLKR falling edge Hold time, DR after CLKR falling edge 3.3 Volts Min Max 6 6 Units ns ns th(FSR) th(DR) DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 39 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Table 6-C Switching Characteristics Over Recommended Operating Conditions for Serial Port Receive Reference Parameter Delay time, DX valid after CLKX rising (Passive Mode) Cycle time, serial port clock Fall time, serial port clock Rise time, serial port clock Pulse duration, serial port clock low/ high Setup time, FSR before CLKR falling edge Setup time, DR before CLKR falling edge Delay time, CLKX rising to FSX Delay time, CLKX rising to DX (Active Mode) Hold time, FSX after CLKX falling edge Hold time, FSX after CLKX rising edge Disable time, CLKX rising to DX Hold time, DX valid after CLKX rising edge -5 6 2H-5 20 3H 6 6 15 15 3.3 Volts Min Max 25 6H 6 6 Units ns ns ns ns ns ns ns ns ns ns ns ns ns td(DX) tc(SCK) tf(SCK) tr(SCK) tw(SCK) tsu(FSR) tsu(DR) td(FSX) td(DX) th(FSX) th(FSX)H tdis(DX) th(DX) * Note: H = 7.629 ns; however, do not operate serial clocks any faster than 2.048 MHz. Thus tc(SCK) should be a minimum of 488.3 nsec. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 40 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 7 7.1 Special Functions Hardware vs. Software Selection Note Many of the functions of the AMBE-2020TM can be accessed through both a hardware and software interfaces to the device. The following hardware inputs, CHANN_SEL[1-0], RATE_SEL[4-0], CODEC_SEL[1-0], VAD_EN, ENCODER_EN, and SLEEP_EN, are only accessed for input during the first 200 microseconds after a hardware reset on RESETN. For predictable operation these signals must remain stable over this time period. After this initialization period the functions that these pins access can only be reconfigured through the channel interface described in section 5.2. Changes on these pins after the 200 microseconds initialization period after reset are ignored, unless another reset is performed. 7.2 Coding Rate Selection The Voice coding rate as well as the FEC coding rate can be selected individually on the AMBE-2020TM. These rates are selected by using Rate Info words as described in section 5.2.4, or through hardware pins RATE_SEL[4-0] subject to the restrictions in section 7.1. The five input pins RATE_SEL[4-0] give 32 preconfigured voice/FEC rates. If rates other than these are desired then the Rate Info Words (in the Framed channel interface) can be used to configure voice and FEC rates in 50 bps increments. Table 7-A Rate Selection Using Rate Info 0-4, compatible w/ AMBE-1000TM (AMBETM) RATE_SEL4 Pin RATE_SEL3 Pin RATE_SEL2 Pin RATE_SEL1 Pin RATE_SEL0 Pin Speech Rate (bps) FEC Rate (bps) Total Rate (bps) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 1 0 0 0 1 0 0 1 1 0 1 0 0 0 0 1 1 1 1 0 0 0 1 1 1 1 1 1 0 1 0 0 0 0 1 0 1 1 1 1 0 1 1 0 0 0 1 0 1 0 0 2400 2350 3600 3350 4000 3750 4800 4550 3600 3100 4150 4400 7750 4650 9600 4850 0 50 0 250 0 250 0 250 1200 1700 2250 2800 250 3350 0 4750 2400 3600 4000 4800 6400 7200 8000 9600 DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 41 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Table 7-B Rate Selection Using Rate Info 0-4, AMBE-2020TM only (AMBE+TM)* RATE_SEL4 Pin RATE_SEL3 Pin RATE_SEL2 Pin RATE_SEL1 Pin RATE_SEL0 Pin Speech Rate (bps) FEC Rate (bps) Total Rate (bps) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1 0 1 1 0 1 0 1 1 1 0 0 1 0 0 1 0 0 0 0 1 1 1 1 1 1 0 0 1 1 0 1 0 1 1 1 0 0 0 0 1 1 0 1 0 0 0 1 1 1 0 1 0 0 1 1 0 2000 3600 4000 2400 4800 4000 3600 2400 6400 4000 4400 8000 4000 9600 3600 2400 0 0 0 1600 0 800 1200 2400 0 2400 2800 0 4000 0 6000 7200 2000 3600 4000 4800 6400 7200 8000 9600 *Note: AMBE+ is only used for speech rates at 3600 bps and higher. Any rates below 3600 bps use a modified algorithm similar to the AMBE algorithm. 7.3 Voice Activation Detection (VAD), Comfort Noise Insertion (CNI) The Voice Activation Detection (VAD) algorithm along with the Comfort Noise Insertion (CNI) feature of the AMBE2020TM chip performs useful functions in systems trying to convert periods of silence, that exist in normal conversation, to savings in system bandwidth or power. With the VAD functions enabled, periods of silence will be denoted by the encoder in two ways. First, the encoder will output a silence frame (in-band). This silence frame contains information regarding the level of background noise, which allows the corresponding decoder to synthesize a "Comfort Noise" signal at the other end. The comfort noise is intended to give the listener the feeling that the call is still connected, as opposed to producing absolute silence, which can give the impression that, the call has been "dropped". Second, the Encoder Silence Detected flag is set in Control Word 1 of the Framed Output format described in section 5.3.3. VAD can be enabled in one of two ways. A logic high signal on the hardware pin VAD_EN (pin 86), subject to the restrictions of section 7.1, enables the VAD function. Once the AMBE2020TM has begun operating, control word 2 is used to enable/disable VAD as described in section 5.2.9. If the VAD features are being used to reduce transmit power during times of conversational silence, DVSI recommends that a silence frame be transmitted at the start of the period and approximately each 500-1000 milliseconds thereafter. This is to ensure that the parameters regarding the levels of background noise are transmitted to the decoder for the smoothest audible transitions between synthesized speech and synthesized silence. The silence threshold value is -25 dBm0 in the VAD algorithm. Each frame that exceeds this level will be classified as voice. If the frame level is less than -25 dBm0 the voice/silence decision will be determined based upon various adaptive thresholds. The synthesis of a Comfort Noise frame by the decoder is not dependant on VAD being enabled. The decoder will produce a comfort noise frame if it receives an in-band silence frame (produced only by an encoder with VAD enabled). DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 42 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 7.4 Dual Tone Multiple Frequency, Detection and Generation The AMBE-2020TM Vocoder Chip is capable of detecting, transmitting, and synthesizing DTMF tones. DTMF features are always enabled. Detection of a DTMF tone by the encoder sets the Encoder DTMF Detected Flag described in section 5.3.3. The DTMF tone detected along with amplitude information is placed in the DTMF Control Word as described in section 5.3.8. Additionally, the encoder passes the DTMF data in-band (within the regular voice data bits) so that normal DTMF tones pass seamlessly from the encoder to the decoder for synthesis. The decoder synthesizes a DTMF tone in response to reception of an in-band DTMF tone frame or by setting the DTMF Control word as described in section 5.2.8. 7.5 Normal Power and Power Saving Modes Power savings can be achieved during times of longer inactivity of the AMBE-2020TM chip by placing it into one of three available Low Power Modes. The chip can be placed into low-power and stand-by modes via hardware or software Control Words. In low power modes the A/D-D/A port will be disabled, concurrently halting any processing of voice frames in either direction. Depending on the low power state selected, either a Wake Up Control Word or a hardware reset on RESETN is necessary to return the AMBE-2020TM to normal operation. 7.5.1 Standard Sleep Mode The standard sleep is the only low power mode that can be entered into either through hardware or software. The AMBE2020TM Chip can be placed into Standard Sleep mode either by setting SLEEP_EN (pin 83) high, subject to the restrictions of section 7.1, or through software by using Control Word 2 with bit 3 set to 1 as described in section 5.2.9. SLEEP_EN should be tied high if you plan to configure the A/D-D/A chip from Standard Sleep mode upon power-up or reset. When using software SLEEP_EN with A-law or -law codecs, it is important to note that if packets are sent to the decoder while it is in sleep mode, noise will be heard at the output. It is recommended that no packets be sent to the decoder until it is commanded to wake up. 7.5.2 Power Down Power Down provides the lowest power usage of the sleep modes, the only drawback to this mode is the necessity of a hardware reset on RESETN (pin 69) to resume normal operation. Table 7-C Summary of Power Saving Modes Power Consumption Sleep Mode Enter State via Return to Normal Operation via N/A Control Word RESETN Wake Up Time Crystal Normal Operation Standard Sleep Power Down N/A SLEEP_EN pin at reset OR Control Word Control Word N/A N/A 95 ms. 24 mW 0.11 mW 3V CMOS TTL Approx. 65mW 36 mW 0.11 mW DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 43 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 7.6 Slip Enable In any real time communication system, clock skew issues must be anticipated to keep the flow of data smooth from one end of the system to the other. The SLIP_EN (pin 82) signal allows the encoder of the AMBE-2020TM to react to small slips in the encoder channel signals. When the AMBE-2020TM is in active mode, the chip produces the signals internally for the transfer of data. Because this timing will then likely drive the transmission channel, the necessity of controlling slip becomes a moot point. Any time the AMBE-2020TM encoder channel is in one of the passive modes and the channel timing is asynchronous to the A/D-D/A clock (very rarely are these two interfaces coupled) then the SLIP_EN pin should be set active high. The AMBE-2020TM Vocoder chip process speech in voice frames that are approximately 20 ms in duration. When configured appropriately the chip provides a slip control feature that automatically adjusts the frame size to either 160 or 161 speech samples per frame. This slip control feature allows the vocoder chip to compensate for drift between the frame and sample rate clocks on the order of approximately 0.6% (6,000 ppm.) The vocoder chip also accepts Slip Control Packets that extend the range of allowable frame sizes to be between 159 and 161 samples per voice frame. When properly used these Slip Control Packets provide the designer with additional flexibility in dealing with clock drift. There are three recommended methods for using slip control on the AMBE-2020TM Vocoder Chip which are described below. The system designer should select the method that best meets the needs of their system configuration. Also included is some background information on the operation of the AMBE-2020TM in passive mode In order to help understand the Slip control feature here is a brief description on reading encoder packets from the AMBE2020TM in passive framed mode. When transmitting a packet, the AMBE-2020TM writes a Header = 0x13ec followed by 23 words of data, followed by 0xfffe into the transmit (i.e. output) buffer. The terminating word 0xfffe is written into the transmit buffer by the AMBE-2020TM at the end of each encoder packet. Normally in passive mode this terminating word is in the transmit buffer at the beginning of each transmission cycle (from the previous frame) and so it is the first word output whenever a packet is transmitted. If the encoder packet is ready, then the second output word will be the packet Header=0x13ec followed by 23 words of data. However if the packet is not ready then the AMBE-2020TM will continue to output the terminating word (0xfffe) until the packet is ready and placed in the transmit buffer. At this point the full 24 word packet beginning with the Header will be output on subsequent transmissions. This process continues for each packet transmission which occurs nominally every 20 ms, provided that each 24 word packet (Header + 23 data words) is read in full. If the full 24 words of the packet are not read from the AMBE-2020TM, then the chip's transmit buffer will contain some words left over from a previous packet. Generally these words would be 0x0000 for lower data rates which don't use the last words of the packet. This case should be avoided hence: It is recommended that in passive framed mode the system always read packets by requesting words from the AMBE-2020TM until a packet Header is received and then continuing to request 23 additional output words from the AMBE-2020TM until a total of 24 words beginning with the Header word = 0x13ec are received. Any words output by the AMBE-2020TM prior to the Header should be ignored by the system (except for monitoring as discussed in Method 3 below). DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 44 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Word # 0 1 2 3 4 5 6 7 8 9 10 11 12-23 Value 0x13ec 0x07xx 0x12c9 0x009f 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 Description Header Slip Control Packet ID, xx=CWD1 (See table 5-B) Typical value xx=00 Slip Control indicator for AMBE-2020TM Slip Control data Must be 0x0000 Must be 0x0000 Must be 0x0000 Must be 0x0000 Must be 0x0000 Must be 0x0000 Must be 0x0000 Must be 0x0000 Channel Data for AMBE(R) voice decoder Table 7D: Slip Control Packet Method 1 - Internal Slip Control This is the simplest method of slip control and is the default method provided slip control is enabled (SLIP_EN is high) on the chip. In this method the vocoder chip's internal slip adjustment of 160 to 161 sample per frame is used. In order to work properly, the system designer must set-up the sample rate and packet timing so that the following constraint is met: 160 < Average-Frame-Interval * (1 +/- Frame-Drift) * Sample-Rate * (1 +/- Sample-Drift) < 161 For example using an Average-Frame-Interval = 20 ms, and assuming 100 parts per million oscillator accuracy (i.e FrameDrift = Sample-Drift = .0001), then the above constraint requires: 8002 Hz < Sample-Rate < 8048 Hz. In practice Sample-Rate=8002 Hz would be preferred since it is closer to the nominal value of 8000 Hz. In another example the Sample-Rate = 8000 Hz. as provided by a PCM source. Again assuming 100 parts per million oscillator accuracy for both clocks (i.e. Frame-Drift = Sample-Drift = .0001), then the above constraint requires: 20.0041 ms. < Average-Frame-Interval < 20.1209 ms. This can be achieved by slightly decreasing the channel bit rate or adding an extra bit into the channel bit stream every several voice frames. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 45 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Method 2 - Extended Slip Control with Periodic Slip Control Packets The AMBE-2020TM can provide extended slip compensation through the insertion of Slip Control Packets. One method of using this capability is for the system to periodically insert these Slip Control Packets into the data stream sent to the AMBE-2020TM. Note that for this method slip control must be enabled (SLIP_EN is high) on the chip. This approach gives the designer a way to accommodate clock drift while providing very flexible frame-interval and sample-rate timing. Furthermore minimal system overhead is required. In this method a Slip Control Packet is generated by the system by setting the Control Words as shown in Table 7D above, where the Channel Data is the compressed voice data being sent to the AMBE decoder. In the Periodic Slip Control method such a Slip Control Packet is input into the vocoder chip every N frames. The value of N must be selected by the design engineer to meet the following constraint: 0.25 > (1/N) > Average-Frame-Interval * Sample-Rate * (Frame-Drift + Sample-Drift) For example with an Average-Frame-Interval = .02 (i.e. 20 ms) and a Sample-Rate = 8000, then with 100 parts per million oscillator accuracy (i.e. Frame-Drift = Sample-Drift = .0001), then above constraint equates to 4 < N < 31.25, and N=30 would be a reasonable selection. In this case the system would input the specified Slip Control Packet into the AMBE-20X0 vocoder chip every 30'th frame enabling the vocoder chip to adjust for the actual clock drift. ***RECOMMENDED*** Method 3 - Extended Slip Control with as needed Slip Control Packets The preferred method for using Slip Control Packets is to monitor the availability of data from the AMBE-2020TM vocoder chip and to only input Slip Control Packets into the data stream sent to the AMBE-2020TM as needed. Note that for this method slip control must be enabled (SLIP_EN is high) on the chip. This method provides compensation for the widest range of clock drift (1.2% or 12000 ppm), with the greatest flexibility in frame-interval and sample-rate timing. In this method the same Slip Control Packets shown in Table 7D are inserted into the data stream going to the AMBE-2020TM. However, unlike in the Periodic method, the packets are not input at regular intervals but are instead only input to the chip when needed. The recommended procedure for this Method of Slip Control is for the system to read a packet from the AMBE-2020TM at regular fixed frame intervals where the fixed interval must be within the range [19.875 - 20.125] ms for a sample rate = 8000 Hz. The system application should check each word output by the AMBE-2020TM and should continue requesting words from the chip until the packet Header followed by 23 data words are received. If the words received before the Header word consist of only a single termination word (0xfffe) then no further action is required. However if two or more termination words are received prior to the Header then the system should input a Slip Control Packet to the AMBE-2020TM on the next available transmission into the chip (i.e. the next packet going into the AMBE-2020TM decoder should be a Slip Control Packet). Once this Slip Control Packet is input into the AMBE-2020TM it will respond within 1-2 frames by advancing the time when packets are ready for transmission by 125 microseconds. Note that this procedure also may require a small amount of buffering in the system to account for the fact that the packet my not be ready for some small time (< 125 microseconds) after it is first requested. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 46 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 8 Appendices 8.1 Example: AD73311 Usage The following examples of A/D-D/A chips have been included to show connections necessary for interfacing to a number of popular chips. U1 CODEC_RX_CLK CODEC_TX_CLK CODEC_TX_STRB CODEC_RX_STRB CODEC_RX_DATA CODEC_TX_DATA AMBE-2020TM 27 33 37 29 31 41 14 17 18 16 19 U2 SCLK SDOFS SDIFS SDO SDI SE MCLK AVDD1 AVDD2 DVDD 20 15 3 9 12 5V 3.3 V SERIAL PORT ENABLE (HIGH) 16.384 MHz AD73311 REFERENCE DATA The AMBE-2020TM sends the following configuration sequence to the AD73311 when CODEC_SEL[1-0]=01b: 0x8113, // write 13 to CRB of AD73311, MCD=1 (DMCLK=MCLK/2),SCD=0(SCLK=DMCLK/8) 0x82f9, // write f9 to CRC of AD73311, 5VEN, REFOUT, all Power ON 0x8300, // write 0 to CRD of AD73311 0x8001 // write 1 to CRA of AD73311 (Enters Data Mode) DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 47 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 8.2 Example: Texas Instruments TLV320AIC10 Usage The Texas Instruments' TLV320AIC10 codec presents a simple low cost solution for use with DVSI's AMBE-2020TM vocoder chip. This application note provides information on interfacing these components. Figure 1 shows a sample block diagram interface, between the TLV320AIC10 codec and DVSI's AMBE-2020TM vocoder chip. U1 CODEC_RX_DATA CODEC_TX_DATA CODEC_TX_CLK CODEC_RX_CLK CODEC_RX_STRB CODEC_TX_STRB AMBE-2020TM 31 41 33 27 29 37 16.384 MHz 16 17 19 20 21 22 23 24 25 26 VDD 27 10 11 VDD VOICE_*RESET 12 13 U27 DOUT DIN SCLK MCLK FSD FS FLAG FC DCSI ALTIN M/*S M0 M1 *PWRDWN *RESET VDD 15 30 AVDD 45 34 3 2 1 48 47 9 8 43 38 7 6 5 4 DV DV DD DD 1 2 AV AV DD DD 1 2 AURXCP AURXM AURXFP INM INP OUTM OUTP VMID FILT DTXIM DTXIP DTXOM DTXOP DV DV SS SS 29 14 AV AV AV AV SS SS SS SS 33 40 42 46 TLV320AIC10 CODEC Configuration SERIAL DATA SCLK *SEE DETAIL BELOW Figure 1: AMBE-2020TM and TLV320AIC10 sample block diagram Configuration: To configure the AMBE-2020TM for operation with the TLV320AIC10, set the CODEC_SEL pins on the AMBE-2020TM vocoder chip to work with a generic 16 bit linear 8 kHz codec as follows: CODEC_SEL [1-0] (pins 85,84) = 00b DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 48 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Initialization Procedure: The control registers in the TLV320AIC10 codec must be initialized for proper operation. The recommended procedure is to initialize the TLV320AIC10 by writing data to its 4 control registers through the DCSI port, while the AMBE-2020TM is held in reset. The timing for the DCSI port is shown in Figure 2. Note that the Device Address (D14-D12) is normally set to 0 unless multiple codec devices are used in cascade. Be sure that the stop bit is at least 2 clock pulses long between data words as shown in the timing diagram. Shift the control words into the device 1 bit at a time at the rate of SCLK. Various configuration data can be used to control the operation of the TLV320AIC10 codec (see the data sheet for more information), however for reference the AMBE-2020TM has been tested with the TLV320AIC10 configured using the register values shown in Table 1. Once the TLV320AIC10 is configured, the AMBE-2020TM should be taken out of reset to begin communication with the codec. The logic connected to the DCSI port does not have to be disabled. The user can make adjustments to the configuration as needed (for example ADC and DAC gain). A reset to the TLV320AIC10 codec will reset all of the internal registers. As a result, the TLV320AIC10 must be reconfigured following a reset. Figure 2: TLV320AIC10 configuration timing via DCSI port Register Address (D11-D9) 0x1 0x2 0x3 0x4 Configuration Data (D7-D0) 0x11 0x08 0x01 0x00 Notes: D4=1: select AUXP AND AUXM for ADC (Handset) D5=0: enable antialiasing filter D0=1: select 16 BIT data Format for DAC D7=0: select normal Operation D4-D0=8: set Frequency Divider N=8 D7-D6=0: default operation D0=1: 16-Bit data format for ADC D7-D4=0: ADC input gain = 0 dB D3-D0=0: DAC output gain = 0 dB Gain values can be adjusted as needed. Table 1: Recommended TLV320AIC10 Configuration Data DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 49 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 VDD D_IN [15:0] LOAD LS_IN CLK_EN SCLK CLK SHIFT CNTL_DATA [15:0] AIC_LOAD Serial Out AIC_CTRL COUNT Q0 Q1 1 U5A 2 U6A 2 U8A 2 U7A 2 3 4 5 AND4 U9 1 AIC_LOAD 1 Q2 CE Q3 1 1 2 CLK CLR Figure 3: TLV320AIC10 Codec Configuration Detail DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 50 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 TLV320AIC10 Reference Schematic (Analog Section) C1 820pF R1 10K R2 20K VOICEOUTP R3 100K R4 100K R5 10K U1A AD8544 1 11 3.3VA R8 R9 10K 10K + R10 10K C5 10uF 4 + C3 200pF C4 2 nF 3 2 R6 20K VOICEOUTM 3.3VA 3.3VA 7 C2 10uF R7 220 J1 2 3 1 4 Handset 2 3 1 4 + + - 5 6 U1B AURXM C6 0.1uF R11 10K 3VA R13 5.6K R12 220K + C7 10uF VMID R14 4.7K + C8 10uF + AURXFP R15 10K AURXCP C9 10uF Title: Analog TLV320AIC10 Rev: 1 Document Number: Reference Materials: AMBE2020TM Vocoder chips Users Manual: http://www.dvsinc.com/literature.htm TLV320AIC10 Data Sheet: http://www-s.ti.com/sc/ds/tlv320aic10.pdf TLV320AIC10 EVM User's Guide: http://www-s.ti.com/sc/psheets/slwu003d/slwu003d.pdf Application Report - Understanding Data Converters: http://www-s.ti.com/sc/psheets/slaa013/slaa013.pdf DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 51 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 8.3 Configuring the AD73311 for 3-Volt Operation The Analog Devices AD73311 codec chip presents a simple low cost solution for use with DVSI's AMBE-2020TM Vocoder chip. This application note provides information on alternative methods of interfacing these components. AD73311AR codec (using a 3 volt supply) It may be desirable for the AD73311 AD/DA converter to be configured for a 3-volt supply voltage instead of a 5-volt supply voltage. The diagram and configuration procedure below outline the details necessary in order to send alternate configuration words to the CODEC. U2 CODEC_RX_CLK CODEC_TX_CLK CODEC_TX_STRB CODEC_RX_STRB CODEC_RX_DATA CODEC_TX_DATA AMBE-2020TM 27 33 37 29 31 41 2 3 U4A SN74AHC125 14 17 18 16 19 SCLK SDOFS SDIFS SDO SDI SE MCLK AVDD1 AVDD2 DVDD 20 15 3 9 12 3.3 Volts High 16.384 MHz 3.3 Volts U1 AD73311AR U8A SDI CONTROL 0/DSP 1/AMBE-2000 1 2 1 U3 SCLK 2 3 U5A SN74AHC125 FS 1 SDO DSP/Microcontroller Figure 2: The AMBE-2020TM with the AD73311 Configured for 3-volt operation. The objective of this circuit is to tri-state the output of the AMBE-2020TM CODEC_TX_DATA. This allows the DSP or Microcontroller to communicate with the AD73311 to send it the desired configuration. Configure the AMBE-2020TM for operation with the AD73311 Codec. Set the CODEC_SEL pins as shown. CODEC_SEL [1-0] (pins 85,84) = 01,b Hold the CODEC_TX_DATA lines on the AMBE2020 in tri-state (in the circuit set the SDI bit to 0,b) during power up. While the AMBE-2020TM lines are in tri-state, send the desired configuration words from the DSP/Microcontroller/logic to the AD73311 (Reset timing constraints for the AD73311 must be met). After sending the configuration words set the SDI Control bit to 1,b (SEE NOTE). This sets the CODEC SDI line for normal operation with the AMBE-2020TM. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 52 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 NOTE: The SDI Control bit must be set to 0 for approximately 365 msec following a hardware reset. Register Address (D10-D8) CRB 0x1 CRC 0x2 CRD 0x3 CRA 0x0 Configuration Data (D7-D0) 0x13 Notes: MCD = 1 Sets DMCLK=MCLK/2 SCD = 0 Sets SCLK=DMCLK/8 D6 = 1 Reference Out Enabled D5 = 1 Reference Power On D4 = 1 DAC Power On D3 = 1 ADC Power On D0 = 1 Power On Gain set to 0 D0 = 1 Puts CODEC in Data Mode 0x79 0x00 0x01 Table 2: Alternate Configuration Data for the AD73311 AD73311L codec (3 volt supply) The Analog Devices AD73311L is a low power 3 volt version of the AD73311. It is possible to use this part with the AMBE-2000TM or AMBE-2020TM vocoder chip utilizing the method described above for sending alternate configuration words to the AD73311 CODEC. Table 3 lists alternate control words for configuring the AD73311L for use with the AMBE-2000TM or AMBE-2020TM. Register Address (D10-D8) CRB 0x1 CRC 0x2 CRD 0x3 CRA 0x0 Configuration Data (D7-D0) 0x13 Notes: MCD = 1 Sets DMCLK=MCLK/2 SCD = 0 Sets SCLK=DMCLK/8 D6 = 1 Reference Out Enabled D5 = 1 Reference Power On D4 = 1 DAC Power On D3 = 1 ADC Power On D0 = 1 Power On Gain set to 0 D0 = 1 Puts CODEC in Data Mode 0x79 0x00 0x01 Table 3: Configuration Data for the AD73311L Additional Reference Material AMBE-2020TM vocoder chip Users Manual http://www.dvsinc.com/literature.htm Application Report - Understanding Data Converters: http://www-s.ti.com/sc/psheets/slaa013/slaa013.pdf AD73311 - Data Sheet http://www.analog.com/productSelection/pdf/AD73311_b.pdf AD73311L - Data Sheet http://www.analog.com/productSelection/pdf/AD73311L_a.pdf DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 53 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 8.4 Interfacing to the Texas Instruments PCM3500 Codec The Texas Instruments PCM3500 codec chip presents a simple low cost solution for use with DVSI's AMBE-2000TM or AMBE-2020TM vocoder chips. This application note provides information on alternative methods of interfacing these components. PCM3500 The block diagram in Figure 1 shows a sample interface between the PCM3500 codec and DVSI's AMBE-2000TM vocoder chip. The AMBE-2000TM or AMBE-2020TM CODEC_SEL bits (see AMBE-2000TM or AMBE-2020TM users manual) must be set for use with a generic 16 bit linear codec (CODEC_SEL1,0 - 00). U1 CODEC_RX_CLCK CODEC_TX_CLCK CODEC_TX_STRB CODEC_RX_STRB CODEC_RX_DATA CODEC_TX_DATA AMBE-2020 27 33 37 29 31 41 11 10 DOUT 8 9 U2 /BCK FS Vdd Vcc 13 24 3.3V M/S 6 Master/Slave DIN PCM3500 Figure 1: The AMBE-2020TM Vocoder with the PCM3500 CODEC There are two advantages to using the PCM3500 codec. The first is the single supply design. The PCM3500 supports a single power supply design from +2.7 v to +3.6 v. The second advantage lies in its simplicity. There are no complicated configuration schemes associated with the codec. For configuration information, please see the PCM3500 data sheet and the reference circuit in Figure 2. Notes On Analog Circuit Design The example circuit assumes that a telephone handset is going to be used in the circuit. Typically, handset microphones have a very small gain and the output is at moderately low levels (on the order of 50 millivolts peak to peak). The PCM3500 Voice Codec is designed for an analog input voltage of 2 volts peak to peak. The analog input in the reference design is amplified (Gain = 22) in order to bring the handset voltage to the level expected by the ADC. The output section is designed using a low pass filter design with a gain of 1. The filter is designed to allow the maximum amount of the voice signal to pass unimpeded. The output of the PCM3500 should be filtered for maximum voice quality. Capacitors C13 and C14 are for creating a very low noise DC bias signal. If a low noise DC bias is available elsewhere in the circuit, they are not needed. For optimum performance, the analog circuit should be adjusted for whatever input (and output) device is used. Please reference the PCM3500 data sheet for the analog requirements. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 54 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Application Information It is strongly recommended that the user review the Application Information provided in the Texas Instruments PCM3500 data sheet before finalizing any design. Additional Reference Material AMBE-2000TM or AMBE-2020TM vocoder chip Users Manual http://www.dvsinc.com/literature.htm Application Report - Understanding Data Converters: http://www-s.ti.com/sc/psheets/slaa013/slaa013.pdf PCM3500 Data Sheet http://www-s.ti.com/sc/ds/pcm3500.pdf PCM3500 Evaluation Board http://www-s.ti.com/sc/psheets/sbau028/sbau028.pdf DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 55 C1 47nF 3.3VA 1 C2 2.2uF R1 20K R2 2.1K R3 887 3 2 C6 1nF C7 100nF Analog In 2 4 1 C4 10uF R4 220 J1 2 3 1 4 Handset 2 3 1 4 + 11 U1A + - 1 2 C3 4.7uF 1 2 C5 4.7uF 1 2 3 4 5 U2 Vcom Vref1 Vref2 Vin AGND M/S TSC BCK FS DIN DOUT FSO PCM3500 Vcc AGND Vout AGND /PDWN Loop HPFD XTI XT0 SCKIO DGND Vdd 24 23 22 21 20 19 18 3.3 V Analog Out 3.3 V 3.3 V R5 220K 6 7 Y1 17 16 15 14 13 1 C11 0.1uF 2 3.3 V 1 2 C12 4.7uF 2 1 1 C8 22pf 4.096 MHz 2 CODEC_TX/RX_CLK CODEC_TX/RX_STRB CODEC_TX_DATA 2 1 8 9 10 11 12 1 2 C9 22pf C10 0.2uF R6 10K U1B 3.3VA 65+ 7 3.3VA R8 10K CODEC_RX_DATA 3.3VA R7 5.6K R9 4.7K + C14 10uF 10K R10 + C13 10uF Title PCM3500 Size B Date: Document Number 1 Monday, February 03, 2003 Sheet 1 of 1 Rev 1 Figure 2 8.5 Expanded Tone Detection and Generation The AMBE-2000TM is capable of detecting and generating single tones as well as dual tones. The single tones span from 156.25 Hz to 3812.5 Hz in 31.25 Hz steps. Tone Index (Decimal) 0-4 5 6 7 ......... 122 123-127 128 ............. 143 144 145 146 147-254 255 Tone Type Invalid Single Single Single ............. Single Invalid DTMF "1" .............. DTMF "D" Call Progress Call Progress Call Progress Invalid No Tone Freq #2 (Hz) N/A N/A N/A N/A ............ N/A N/A 1209 ............. 1633 440 480 620 N/A N/A Freq #1 (Hz) N/A 156.25 187.5 218.75 ............... 3812.5 N/A 697 ............ 941 350 440 480 N/A N/A AMBE-2020TM Vocoder Chip User's Manual Version 4.0 8.6 Soft Decision Decoding In modern communication systems the transmitter transmits information in the form of symbols. The demodulator takes the received signal and tries to decide which symbol was transmitted. In other words, the demodulator is making a decision based on the received signal. For instance, in a binary system, the symbols could be represented by a 1 and a 0. Because of interference and many other factors, these signals can be misinterpreted because of channel degradation. Significant improvement in FEC performance can be added by setting up the receiver so that the demodulator is making a finer estimation of the received energy prior to the decoder, this is called soft-decision decoding. The link below describes this in more detail. http://www.mathworks.com/access/helpdesk/help/toolbox/commblks/usersguide/tutor135.shtml The AMBE-2000/2020 utilizes a 4 bit soft decision decoder. The bits are defined as follows: Decision Value (Binary) Interpretation 0000 Most confident 0 Least confident 0 Least confident 1 Most confident 1 0111 1000 1111 Placing a logic high on pin 79 of either the AMBE-2000TM or AMBE-2020TM vocoder chips enables the soft decision error correction on the decoder. Enabling the soft-decision does nothing to the encoder packet. The packet will look like a normal encoded packet. The user must implement circuitry at the receive end of the channel for making a finer (4 bit) estimation of the received energy (see above link for an explanation of how this can be done). The AMBE-2000/2020 decoder packet structure is altered. The decoder expects each voice data bit of the encoded packet to be represented by 4 soft decision (SD) bits. The decoder will make the decision of whether or not a 1 or a 0 is represented by the SD bits. Table 1 is an example of a soft-decision decoder packet. DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 58 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 Table 1. Soft-Decision Decoder Packet Word # 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 (12) 16 bit words of overhead (192 bits) Power Control (8 bits) Description Header always set to 0x13EC Control Word 1 (8 bits) Rate info 0 Rate info 1 Rate info 2 Rate info 3 Rate info 4 Unused in Input Unused in Input Unused in Input DTMF Control Control Word 2 SD1 SD2 SD5 SD6 SD9 SD10 SD13 SD14 SD17 SD18 SD21 SD22 SD25 SD26 SD29 SD30 SD33 SD34 SD37 SD38 SD41 SD42 SD45 SD46 SD49 SD50 SD53 SD54 SD57 SD58 SD61 SD62 SD65 SD66 SD69 SD70 SD73 SD74 SD77 SD78 SD81 SD82 SD85 SD86 SD89 SD90 SD93 SD44 SD97 SD98 SD101 SD102 SD105 SD106 SD109 SD110 SD113 SD114 SD117 SD118 SD121 SD122 SD125 SD126 SD129 SD130 SD133 SD134 SD137 SD138 SD141 SD142 SD145 SD146 SD149 SD150 SD153 SD154 SD157 SD158 SD161 SD162 SD165 SD166 SD169 SD170 SD173 SD174 SD177 SD178 SD181 SD182 SD185 SD186 SD189 SD190 SD0 SD4 SD8 SD12 SD16 SD20 SD24 SD28 SD32 SD36 SD40 SD44 SD48 SD52 SD56 SD60 SD64 SD68 SD72 SD76 SD80 SD84 SD88 SD92 SD96 SD100 SD104 SD108 SD112 SD116 SD120 SD124 SD128 Sd132 SD136 SD140 SD144 SD148 SD152 SD156 SD160 SD164 SD168 SD172 SD176 SD180 SD184 SD188 SD3 SD7 SD11 SD15 SD19 SD23 SD27 SD31 SD35 SD39 SD43 SD47 SD51 SD55 SD59 SD63 SD67 SD71 SD75 SD79 SD33 SD87 SD91 SD95 SD99 SD103 SD107 SD111 SD115 SD119 SD123 SD127 SD131 SD135 SD139 SD143 SD147 SD151 SD155 SD159 SD163 SD167 SD171 SD175 Sd179 SD183 SD187 SD191 60 sixteen-bit words = 120 bytes = 960 bits DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com SD = 4 bits x 4 per word = 16 bit words of data x 48 words = 768 bits 20 ms frame Page 59 AMBE-2020TM Vocoder Chip User's Manual Version 4.0 History of Revisions Revision Number 1.0 1.1 1.2 Date of Revision November 1999 April 2000 August Description Initial Version Pin descriptions Serial Configuration Selection Channel Serial Interface Pin Descriptions Table 6-A CODEC_SEL[1-0]: A/D- D/A Hardware Configuration Values Added AD73311 to AMBE-2020TM connection diagram Pin Description : ENCODER_EN is pin 24. Corrections of Table 4A and 4B Clarification on VAD Expanded description of Control Word 1 Table 5-M added DTMF Code 0xff Channel and Codec Timing Diagrams and Tables Table 4B Corrected (pin 80 and 81) Added Note describing H Clarified/corrected the following pages. Updated Timing Diagrams and Tables for Channel and Codec Changed Pin Description CLK_I to X2/CLKIN and CLK_I2 to X1 Updated timing diagram and table for X2/CLKIN and RESETN Replaced ...set the DTMF Code to 0x00 to ...set the DTMF Code to 0xff Updated company address Removed description of Decoder Output Volume Control Modified SLEEP_EN Section 7.5.1 Modified description of EPR Added Detailed Explination of Slip Enable Control Added description of Decoder Silence Detect, Decoder Frame Repeat and Encoder DTMF Detect bits Various grammer changes Vad Threshold Explination Associated Delay 4.0 January 2003 Front End Requirements Added AD73311AR Configuration Words Added Tech Note 8.2 Expanded DTMF Tone Chart Pages 11-12 19 20 35 42 11 19 28 30 32 21, 22, 36, 37 19 22, 37 10, 11, 12, 13, 16, 17, 21, 22, 23, 27, 28, 29, 32, 34, 36, 37, 39 21, 22, 36, 37 11 13, 14 27 2 28 40 11, 12, 20 42, 43 30 43 41 14 9 46 47 30, 34 1.3 October 1.4 November 2000 2.0 2.1 2.2 January 2001 February 2001 February 2001 3.0 August 2001 DVSI Confidential Proprietary, Subject to Change Visit us at www.dvsinc.com Page 60 |
Price & Availability of AMBE-2020
 |
|
|
|
|
All Rights Reserved © IC-ON-LINE 2003 - 2022 |
| [Add Bookmark] [Contact Us] [Link exchange] [Privacy policy] |
|
Mirror Sites : [www.datasheet.hk]
[www.maxim4u.com] [www.ic-on-line.cn]
[www.ic-on-line.com] [www.ic-on-line.net]
[www.alldatasheet.com.cn]
[www.gdcy.com]
[www.gdcy.net] |