What kind method ADPCM? What specific value quantized width ADPCM? sho
|
|
|
Top Searches for this datasheet
E2D0078-39-91 Semiconductor
What kind method ADPCM? What specific value quantized width ADPCM? should perceive relationship between rates synthesis length between rates sound quality? Q4.1: sinusoidal waves produced ADPCM? Q4.2: sinusoidal waves ADPCM used produce output voltage Vp-p? Q4.3: What frequencies sinusoidal waves? does voice synthesis require low-pass filter? should cut-off frequency low-pass filter determined? there inexpensive external filter configuration voice synthesis? Which chip's substrate electric potential, level? should multiple power supply pins (AVDD, DVDD forth) connected? Q10: Which EPROM used MSM6650 MSM6376 evaluation boards? Q11: voice output over lines? Q12: voice quality adversely affected deviation master oscillation frequency? Q13: What kind recording tape should recording original sound when asking voice analysis? Q14: does sound volume change when playback starts from part phrase? Q15: normal playback obtained when playback starts from part phrase? Q16: What analog flash memory?
1/28
Semiconductor What kind method ADPCM?
First basic methods mentioned, followed ADPCM method.
methods (Delta Modulation) method certain quantity predetermined each cycle sampling added subtracted express voice waveform. other words, addition encoded while subtraction encoded Thus, rapid changes voice waveform with respect step width cannot covered this method. Figure 10.1 illustrates this case. (Adaptive Delta Modulation) method, quantized width adjusted according rapid changes improve response voice waveform. Encoding based this means. value continues certain period, quantized width enlarged quicker response. Figure 10.1 illustrates such quick response. ADPCM method ADPCM (Adaptive Differential Pulse Code Modulation) method such that basic width quantization adjusted adaptive rapid changes each cycle sampling encode each signal three four bits data. This provides higher response voice waveform. example, case four-bit ADPCM, upper stands polarity (increase decrease) data, while lower three bits determine multiplier factor basic width quantization (D). (The value depends correlation with past data.) This means that about pieces data [(about pieces) (for bits)] changed time. Therefore, approximately pieces data available. Figure 10.2 illustrates this case. Furthermore, three-bit ADPCM, upper stands polarity (increase decrease) data, while lower bits determine multiplier factor basic width quantization (D). This means that about pieces data [(about pieces) (for bits)] changed time. Therefore, pieces data available. ADPCM method allows achievement high-quality sound relatively simple configuration, easy creation voice data. Note, however, that there compatibility between ADPCM data created different manufacturers. (Note) When playback starts from part phrase, normal waveforms cannot reproduced because ADPCM algorithm employed. A14.
2/28
Semiconductor
Figure 10.1 Methods
Encoding
Time
method
Dn+2
Dn+1
Time
Encoding
method
Figure 10.2 4-bit ADPCM
An+2 Dn+2 An+1 Dn+1 Time
About pieces data Eight pieces data
Encoding Polarity Multiplier factor (An)
3/28
Semiconductor What specific value quantized width ADPCM?
quantized width important factor determining sound quality implemented ADPCM method, know-how owned individual manufacturers. Hence, detailed data ADPCM method (OKIADPCM) cannot disclosed. OKIADPCM method compatible with ADPCM method conforming ITU-T (former CCITT).
4/28
Semiconductor
should perceive relationship between rates synthesis length between rates sound quality? rate length synthesis
rate indicates degree information compression, many bits data required synthesis second. Thus, rate depends sampling frequency amount data sample, determined following formula: rate (kbps) sampling frequency (kHz) amount data sample (bits) Example Sampling frequency: kHz, data: 4-bit ADPCM rate (kbps) (kHz) (bits) (kbps) Example Sampling frequency: kHz, data: 4-bit ADPCM rate (kbps) (kHz) (bits) (kbps) length synthesis depends memory capacity rate, shown following formula:
Synthesis length (seconds) 1.024 (memory capacity) (Kbits) rate (kbps) 1.024 (memory capacity) (Kbits) (seconds) sampling frequency (kHz) data amount sample (bits)
Example Sampling frequency: kHz, 4-bit ADPCM, Memory capacity: Kbits
Synthesis length (seconds) 1.024 Kbits (kHz) (bits) 16.4 (seconds)
rate sound quality lower rate results longer synthesis length. response voice waveform becomes lowered, with deteriorated sound quality.
5/28
Semiconductor Recording/Playback Time Quick Reference Table
following quick reference table recording/playback times calculated basis each voice synthesis method.
Voice synthesis method Sampling frequency (kHz) 10.6 12.8 16.0 ADPCM 32.0 10.6 12.8 16.0 32.0 10.6 12.8 16.0 32.0 (Reference) Memory capacity
length (bit)
rate
(kbps) 256Kbit 512Kbit 1Mbit 1.5Mbit 2.0Mbit 3.0Mbit 4.0Mbit 8.0Mbit Maximum playback time (second) 16.0 21.2 25.6 32.0 42.4 51.2 64.0 128.0 12.0 15.9 19.2 24.0 31.8 38.4 48.0 96.0 32.0 42.4 51.2 64.0 84.8 102.4 128.0 256.0 12.0 10.0 12.6 16.0 1118
6/28
Semiconductor Q4.1: A4.1 sinusoidal waves produced ADPCM?
Tables 10.1.1 10.1.4 show ADPCM data 4-bit sinusoidal waves, Table 10.1.5 shows ADPCM data 3-bit sinusoidal waves. There four kinds 4bit data corresponding four, six, twelve samples depending repeated data count period sinusoidal wave. Figure 10.3 shows model output waveform with supply voltage (VDD) five volts.
Q4.3, Table 10.2 determining frequency sinusoidal wave frequencies.
Leading transient data
Repeat data period ("n" integer)
Trailing transient data
Output voltage VP-P
Figure 10.3 Sinusoidal Waveform Output ADPCM
7/28
Semiconductor
Table 10.1.1 Four-sample Sinusoidal Wave Output Voltage ADPCM (4-bit) Data Simusoidal wave frequency (fsam kHz)
Output voltage (VDD (2.0 VP-P) (2.6 VP-P) (3.0 VP-P) Leading transient data Repeat data Trailing transient data
(3.5 VP-P)
8/28
Semiconductor
Table 10.1.2 Six-sample Sinusoidal Wave Output Voltage ADPCM (4-bit) Data Simusoidal wave frequency 1.33 (fsam kHz)
Output voltage (VDD (0.85 VP-P) (1.0 VP-P) Leading transient data Repeat data Trailing transient data
(1.5 VP-P)
(2.0 VP-P)
(2.6 VP-P)
(3.5 VP-P)
(4.25 VP-P)
9/28
Semiconductor
Table 10.1.3 Ten-sample Sinusoidal Wave Output Voltage ADPCM (4-bit) Data Simusoidal wave frequency (fsam kHz)
Output voltage (VDD (1.0 VP-P) (2.0 VP-P) Leading transient data Repeat data Trailing transient data
(3.0 VP-P)
(4.0 VP-P)
10/28
Semiconductor
Table 10.1.4 Twelve-sample Sinusoidal Wave Output Voltage ADPCM (4-bit) Data Simusoidal wave frequency (fsam kHz)
Output voltage (VDD (1.0 VP-P) (2.0 VP-P) Leading transient data Repeat data Trailing transient data
(3.0 VP-P)
(4.0 VP-P)
11/28
Semiconductor
Table 10.1.5 Six-sample Sinusoidal Wave Output Voltage ADPCM (3-bit) Data Simusoidal wave frequency 1.33 (fsam kHz)
Output voltage (VDD (1.0 VP-P) (2.0 VP-P) Leading transient data Repeat data Trailing transient data
(3.0 VP-P)
12/28
Semiconductor Q4.2: A4.2 Q4.3: A4.3
sinusoidal waves ADPCM used produce output voltage Vp-p? Data covered Tables 10.1.1 10.1.5 cannot generated. What frequencies sinusoidal waves? [kHz] stands frequency sinusoidal wave resulting from ADPCM data covered Tables 10.1.1 10.1.5, while fSAM [kHz] stands sampling frequency. samples ADPCM data repeated, frequency resulting sinusoidal wave given formula below. fSAM/n (Common four-bit three-bit data.) actual values, Table 10.2. Table 10.2 Frequency Synthesized Sinusoidal Wave
Sampling Repeated frequency data samples (Hz)
Frequency (Hz) synthesized sinusoidal wave samples samples samples samples
13/28
Semiconductor does voice synthesis require low-pass filter?
low-pass filter (LPF) only passes frequency components below certain frequency input signal. Synthesized voice output issued converter. Therefore, stepwise waveform output, shown below.
Figure 10.4 Voice Output Waveform from Converter waveform contains high-frequency noise components (metal-like sound). Moreover, sampling theorem indicates that only frequency components below half sampling frequency valid output. high-frequency noise components removed LPF, following output waveform obtained. (Smooth sound)
Figure 10.5 Output Waveform
14/28
Semiconductor should cut-off frequency low-pass filter determined?
ideal low-pass filter completely cuts frequency components above certain frequency. reality however, output frequency components above certain frequency attenuated slope. Conventionally, such slope expressed "dB/oct". example, slope indicates that, when frequency doubled octave), output becomes (1/4). frequency which attenuation started referred cut-off frequency (*1). optimum cut-off frequency depends sampling frequency, attenuation characteristics, frequency components source voice. conventional measure design degree attenuation half sampling frequency (fSAM). Table 10.3 lists reference values. optimum value varies significantly with frequency components source voice, recommended that cut-off frequency determined auditory way. Table 10.3 Relationship between Filter's Attenuation Characteristics, Cut-off Frequency Sampling Frequency
Filter's attenuation characteristics (dB/oct)
Cut-off frequency fSAM 0.33 fSAM 0.35 fSAM 0.38 fSAM
Gain fSAM (dB) -8.7 -12.5
Figure number
more abrupt attenuation characteristics, further efficient high-frequency noise removal effective signal component output. increase number component devices results less cost effective performance. Figure 10.6 shows filter's frequency characteristics (fSAM).
15/28
Semiconductor
Gain (dB)
Frequency (Hz)
Figure 10.6 Filter's Frequency Characteristics (fSAM) cut-off frequency strictly defined frequency which attenuation begins. Butterworth type low-pass filter, cut-off frequency defined point total characteristics". Chebyshev type low-pass filter, cut-off frequency defined "first point beyond maximum ripple amplitude passband". Roughly, however, cut-off frequency defined frequency which attenuation begins.
16/28
Semiconductor there inexpensive external filter configuration voice synthesis?
active filters (filters using active devices) classified Butterworth, Bessel Chebyshev type filters. These filters selected according purposes applications. Butterworth type active filter focuses flatness passband. characteristics attenuation response inferior those Bessel Chebyshev type active filters. applications where severe flatness passband required case with used voice synthesis, Chebyshev type active filter recommended where allowing ripples, abrupt attenuation characteristics attained active filter composed smaller number parts. Chebyshev type active filter designed selecting appropriate ripple amplitude attenuation characteristics. frequency characteristics speaker itself does reach desired cut-off frequency, filter required.
Configuration LPF, consisting stage, containing active device shown Figure 10.7, transfer characteristics given formula below.
where Cut-off frequency 2pfo)
Figure 10.7 Consisting Stage Figure 10.8 plots (jw). shown Figure 10.8, frequency characteristics shows attenuation dB/oct above value observed.
Gain (dB)
w/wo
Figure 10.8 Frequency Characteristics Consisting Stage 17/28
Semiconductor Figure 10.9 illustrates circuit configuration second-order Chebyshev type filter.
Figure 10.9 Second-order Chebyshev Type Filter following gives transfer characteristics circuit.
gain because voltage follower consisting operational amplifier used. Assuming that equal following expressions obtained:
2Qwo
Design high-order Chebyshev type filter Such even-order filters fourth- sixth-ones resolved into second-order elements. Such odd-order filters third- fifth-ones resolved into second- first-order (passive filter consisting stage) elements. example, fourth-order filter resolved into second-order elements shown Figure 10.10. Determining allows fourth-order filter readily built.
Second-order filter INPUT Second-order filter OUTPUT
Figure 10.10 Implementation Fourth-order
18/28
Semiconductor
Chebyshev type filter incurs ripple passband. Attenuation characteristics vary with permissible ripples, settings each stage changed. Table 10.4 lists values Chebyshev type LPF. Table 10.4 Values Chebyshev Type
Ripple=0.1dB Ripple=0.2dB Ripple=0.25dB Ripple=0.3dB Ripple=0.5dB
1.231418
0.8637210
Second order Third order
1.8204497 0.7673593 1.2999029 1.3409276 0.9694057 0.5*
1.5351966 0.7966418 1.1889612 1.4595033 0.8146341 0.5*
1.4539722 0.8092536 1.3911667 0.8210811 1.1569921 1.5080264 1.1321861 1.5524768 0.7672227 0.5* 0.7292773 0.5*
1.0688535 1.7061895 0.6254565 0.5*
Fourth order
1.1532699 2.1829303 0.7892557 0.6188010
1.0948338 2.4350125 1.0779389 2.5361100 1.0648159 2.6279020 1.0312704 2.9405542 0.7011094 0.6458968 0.6744223 0.6572494 0.6532428 0.6677803 0.5970024 0.7051102 1.0570753 3.7068586 1.0466301 3.8756825 1.0385110 4.0283601 1.0177347 4.5449633 0.7472558 1.0009079 0.7324054 1.0359319 0.7207553 1.0678979 0.6904832 1.1778056 0.4614106 0.5* 0.4369509 0.5* 0.4171291 0.5* 0.3623196 0.5*
Fifth order
1.0931318 3.2820141 0.7974460 0.9145215 0.5389143 0.5*
Sixth order
1.0627261 4.6329012 0.8344903 1.3315707 0.5131875 0.5994600
1.0382299 5.2689021 1.0311242 5.5204164 1.02555981 5.7474076 1.0114459 6.5128456 0.8030621 1.4917187 0.7938542 1.5556533 0.7866630 1.6135959 0.7681212 1.8103772 0.4603216 0.6259511 0.4440628 0.6370268 0.4310754 0.6472924 0.3962290 0.6836390
value 0.5* indicates first-order stage.
Example design following gives example designing fifth-order Chebyshev type LPF. permissible ripple Figure 10.11 provides circuit.
OUTPUT
INPUT
Second-order filter first stage
Second-order filter second stage
First-order filter third stage
Figure 10.11 Fifth-order Chebyshev Type
19/28
Semiconductor
values covered Table 10.4. Data fifth order ripple provides following values. second-order filter first stage 1.0177347 4.5449633 second-order filter second stage 0.6904832 1.1778056 first-order filter third stage 0.3623196 values above used calculate constants. When cut-off frequency kHz, constants determined follows.
Second-order filter first stage RF1=51 2850 (Hz) 2Qw0 2pf0 9953 (pF)
2qn2pf0
(pF)
Second-order filter second stage =2800 0.6904832 1933 (Hz) 1.1778056 3463 (pF) 1933 (pF) 2pw0 1.1778056 1933
First-order filter third stage RF3=68 1014 (Hz) 2pf0 2308 (pF)
20/28
Semiconductor
value changed reflect actual capacitor value. values have been multiplied 1.5, value should divided 1.5. Selecting appropriate capacitor values leads final determination filter constants. Figure 10.12.
0.01 INPUT 2200 3300
OUTPUT
Figure 10.12 Designed Filter (Fifth Order) Then, filter characteristics Figure 10.12 plotted. Transfer characteristics second-order filters first second stages expressed formula below.
Transfer characteristics first-order filter third stage expressed formula below.
example, equal filter first stage, following expression.
jw0) 4.5449633j (1)2
following gives absolute value voltage ratio. 4.545 13.15 (dB) absolute value 13.15 frequency 2850 Figure 10.13 provides plotted dashed curve. alternate dot-dash line alternate dot-dash line cover second third stages, respectively. solid line provides total characteristics. capacitor values have been approximated, total characteristics indicate that maximum ripple cut-off frequency about kHz.
21/28
Semiconductor
Gain (dB)
Frequency (Hz)
Figure 10.13 Frequency Characteristics Designed Filter (Fifth Order) Figures 10.14 10.15 provide constants frequency characteristics designed thirdorder Chebyshev type LPF. solid line Figure 10.15 provides total characteristics. capacitor values have been approximated, characteristics indicate that maximum ripple cut-off frequency 2.56 kHz.
6600 INPUT
OUTPUT 3300
Figure 10.14 Designed Filter (Third Order)
22/28
Semiconductor
Gain (dB)
Frequency (Hz)
Figure 10.15 Frequency Characteristics Designed Filter (Third Order)
23/28
Semiconductor Which chip's substrate electric potential, level?
Table 10.5 lists voice chip products available from their substrate potential. Table 10.5 Chip products their substrate potential
chip product model MSM6658A-XXX MSM6656A-XXX MSM6655A-XXX MSM6654A-XXX MSM6653A-XXX MSM6652A-XXX MSM6650 MSM9805-XXX MSM9803-XXX MSM9802-XXX MSM9836-XXX MSM6376 MSM6379 MSM6378A MSM6588 MSM6722 MSA180 MSC1157 Chip's substrate pontential
24/28
Semiconductor
should multiple power supply pins (AVDD, DVDD forth) connected? such MSM6588 having multiple supply pins including AVDD, DVDD, AGND DGND. same power supply should used shown Figure 10.16. Connect AVDD, DVDD, AGND, DGND separately each printed board wiring pattern.
Figure 10.16 Typical Connection MSM6588 Power Supply following connection should made.
Power supply analog signals Power supply digital signals
DVDD DVDD'
separation supply pins AVDD (analog system) DVDD (digital logic system) contributes sound quality enhancement. Ideally stable potential from separate power sources should supplied. reality supply from separate power sources results potential differences analog digital systems. Since latching damage same power supply must used.
DVDD' DVDD AVDD MSM6588 DGND AGND AVDD
Power supply DVDD DVDD' AVDD
25/28
Semiconductor Q10:
Which EPROM used MSM6650 MSM6376 evaluation boards? Table 10.6 lists EPROMs (including ROMs) that applied evaluation boards. Table 10.6 EPROMs Applicable Evaluation Boards
MSM6650 Evaluation Board
Mbits
M5M27C401, MBM27C4001, MSM27C401 (OTP), TC574000, mPD27C4001, devices with same layout
MSM6650 Evaluation Board MSM6376 Evaluation Board
Mbit
M5M27C101, MBM27C1001, MSM271000, TC571000, mPD27C1001, devices with same layout
Q11:
voice output over lines? Figure 10.17 covers MSM6650 family (LPF output).
Note: Resistors dividing must adjusted that maximum line output value mVp-p LINE
MSM6650 family
AOUT
Figure 10.17 Typical Connection Line Output MSM6650 Family
26/28
Semiconductor Q12:
voice quality adversely affected deviation master oscillation frequency? master oscillation frequency within range recommended operating conditions (quaranteed range), will malfunction. However, speech speed pitch change with deviation from typical value. possible correct deviation with voice analysis editing tool before performing voice analysis. What kind recording tape should recording original sound when asking voice analysis? Send following recording tapes with original sound. Open Reel Tape (tape speed 19cm/sec, normal winding) (Digital Audio Tape) commercialized cassette tape (Mini Disk) because adversely affect quality analyzed sound. does sound volume change when playback starts from part phrase? When playback starts from part phrase, normal waveforms cannot reproduced because ADPCM algorithm employed.
When playback starts from point sound volume changes normal waveforms cannot reproduced. Target products: MSM9841/MSM9842 and, MSM6588/MSM6688 When command used
Q13:
Q14:
Start Phrase
Q15:
normal waveforms reproduced when playback starts from part phrase? recording/playback continuously without interruption after dividing phrase recorded into many phrases with recording/playback time second second. reproduced sound useful though voices between divided phrases missing sound quality little degraded. quality reproduced sound should evaluated before practical using's demonstration board. Target products: MSM6588/MSM6688 When command used MSM9841/MSM9842
27/28
Semiconductor Q16: What analog flash memory?
priciple analog flash memory described below comparing digital flash memory analog flash memory. following waveform, digital values analog values shown, which sampled when analog signals input 8-bit converter. When sampled digital value written digital flash memory, 8-bit memory required. other hand, when analog flash memory used, only memory capacity needed digital flash memory required because analog value itself stored memory cell. analog flash memory innovative memory, which eliminates need converters requires small memory capacity because analog signals input analog value directly stored memory cell. Principle analog flash memory Difference between digital flash memory analog flash memory
Digital value 11111111 11000000 Analog value 5.00V 3.75V
00000000
0.00V
When digital flash memory used
When analog flash memory used
3.75V
Only transistor required transistors required Only memory capacity required
Conceptual diagram recording playback with analog flash memory
A-IN
Analog Flash
A-OUT
converters required
28/28
E2Y0002-29-62
NOTICE
information contained herein change without notice owing product and/or technical improvements. Before using product, please make sure that information being referred up-to-date. outline action examples application circuits described herein have been chosen explanation standard action performance product. When planning product, please ensure that external conditions reflected actual circuit, assembly, program designs. When designing your product, please product below specified maximum ratings within specified operating ranges including, limited operating voltage, power dissipation, operating temperature. assumes responsibility liability whatsoever failure unusual unexpected operation resulting from misuse, neglect, improper installation, repair, alteration accident, improper handling, unusual physical electrical stress including, limited exposure parameters beyond specified maximum ratings operation outside specified operating range. Neither indemnity against license third party's industrial intellectual property right, etc. granted connection with product and/or information drawings contained herein. responsibility assumed infringement third party's right which result from thereof. products listed this document intended general electronics equipment commercial applications (e.g., office automation, communication equipment, measurement equipment, consumer electronics, etc.). These products authorized system application that requires special enhanced quality reliability characteristics system application where failure such system application result loss damage property, death injury humans. Such applications include, limited traffic automotive equipment, safety devices, aerospace equipment, nuclear power control, medical equipment, life-support systems. Certain products this document need government approval before they exported particular countries. purchaser assumes responsibility determining legality export these products will take appropriate necessary steps their expense these. part contents contained herein reprinted reproduced without prior permission. MS-DOS registered trademark Microsoft Corporation.
Copyright 1999 Electric Industry Co., Ltd.
Printed Japan
Other recent searches
SN74F175 - SN74F175 SN74F175 Datasheet
SD-73929-006 - SD-73929-006 SD-73929-006 Datasheet
pkg-3 - pkg-3 pkg-3 Datasheet
NJW1167 - NJW1167 NJW1167 Datasheet
NJW1167V - NJW1167V NJW1167V Datasheet
KTX212E - KTX212E KTX212E Datasheet
HJIV04AA - HJIV04AA HJIV04AA Datasheet
CY7C1418BV18 - CY7C1418BV18 CY7C1418BV18 Datasheet
CY7C1420BV18 - CY7C1420BV18 CY7C1420BV18 Datasheet
CY3125 - CY3125 CY3125 Datasheet
Privacy Policy | Disclaimer
© 2012 Datasheet Archive
