🚀 go-pugleaf

In article <BuBn7u.BFu@news.cso.uiuc.edu> ja51359@uxa.cso.uiuc.edu (axelrod) writes:
>	I'm looking for a audio compression algorithm that will result
>in telephone quality reproduction.  I.E.  4Khz bandwidth, limited dynamic
>range, average S/N ratio.
>	I'm already familiar with using delta-fibonacci, delta-huffman
>techniques, but I'm looking for a more lossy algorithm that will give
>better compression results, more like on the order of 8:1 with 8-bit
>samples.
>	How is the quality of CELP compression?  I heard voices end up
>sounding robotic.  I'd like something that sounds natural.
>
>Thanks,
>Jeff
>

CELP does not sound robotic above 4 kbits/sec, it's pretty much phone-like.
LPC, on the other hand, is best left to the Daleks on Dr. Who.
 

ja51359@uxa.cso.uiuc.edu (axelrod) writes:
>	I'm looking for a audio compression algorithm that will result
>in telephone quality reproduction.  I.E.  4Khz bandwidth, limited dynamic
>range, average S/N ratio.
>	I'm already familiar with using delta-fibonacci, delta-huffman
>techniques, but I'm looking for a more lossy algorithm that will give
>better compression results, more like on the order of 8:1 with 8-bit
>samples.
>	How is the quality of CELP compression?  I heard voices end up
>sounding robotic.  I'd like something that sounds natural.

Our version of CELP, QCELP, sounds quite decent.  If things aren't
tuned just right, voices can get a "sharpness" to them.  To my ear it
sounds superior to standard telephone, and those I've called have been
unable to tell when I'm calling with the desk phone and with the
cellular phone unless we introduced plenty of noise into the system
(at which time the voice starts sounding somewhat "bubbly" as the
noise overwhelms our error correction).

We output 192 bits per 20 millisecond frame which works out to 1200
bytes per second, or 4.3 megabytes per hour of speech.  In addition,
we do voice activity detection and can produce half, quarter, and
eighth rate frames.  The voice activity factor of standard speech
works out to about 0.6 with this method, which means that the
resulting data is only 60% of the size of that where we force it to
stay in full rate mode, which gives about 2.6 megs for an hour of
speech, or 720 bytes per second.

Given your sampling rate of 8000 Hz with 8 bit samples, that would be
28.8 megs for an hour of speech, so we're doing around 7:1 without
even bothering with voice activity, about 11:1 with it, including
error correction.

We're doing all this in an ASIC, but it demonstrates that it's
possible to get what you want with a version of CELP.  At least it
might be worth looking into.
-- 
I never made a mistake in my life.  I thought I did once, but I was wrong.

I've had some comments from our vocoder team on the actual figures
involved - I was actually underspecifying the product (so much for my
marketing career)!

First, I mentioned that the frames were 192 bits each including error
correction.  Without any of the error correction, etc. crud it's 160
bits, or 1000 bytes per second, so that's 80% of my earlier figure.

Second, I claimed we could cut 40% of the size due to the variable
rates.  It's actually 50%.

That gives 8:1 compression at full rate, 16:1 at variable rate for
standard conversations.

Finally, apparently "voice activity detection" has nasty connotations
of those horrible off/on systems which always seem to be chopping off
parts of your words.  Under our version, we're always doing something,
if only at low rate, which sounds much nicer.  That really has nothing
to do with the actual output, but I didn't want to impugn the great
job our vocoder people have done.  I'll never be in marketing (thank
god!).



-- 
I dress this way to annoy you.

Here's some excellent info on CELP in general:

---------------------- start quote ------------------------

Joe Campbell says:

In case you'd like to experiment with CELP, you can obtain a software
implementation of the 4800 bps Fed Std CELP coder for free:

The U.S. DoD's Federal-Standard-1016 based 4800 bps code excited linear
prediction voice coder version 3.2 (CELP 3.2) Fortran and C simulation source
codes are now available for worldwide distribution at no charge (on DOS
diskettes, but configured to compile on Sun SPARC stations) from:

Bob Fenichel
National Communications System
Washington, D.C.  20305
1-703-692-2124
1-703-746-4960 (fax)

In addition to the source codes, example input and processed speech files
are included along with a technical information bulletin to assist in
implementation of FS-1016 CELP.  (An anonymous ftp site is being considered
for future releases.)

Copies of the actual standard "Federal Standard 1016, Telecommunications:
Analog to Digital Conversion of Radio Voice by 4,800 bit/second Code
Excited Linear Prediction (CELP)" are available for $2.50 each from:

GSA Rm 6654
7th & D St SW
Washington, D.C.  20407
1-202-708-9205

The following articles describe the Federal-Standard-1016 4.8-kbps CELP
coder (it's unnecessary to read more than one):

Campbell, Joseph P. Jr., Thomas E. Tremain and Vanoy C. Welch,
"The Federal Standard 1016 4800 bps CELP Voice Coder," Digital Signal
Processing, Academic Press, 1991, Vol. 1, No. 3, p. 145-155.

Campbell, Joseph P. Jr., Thomas E. Tremain and Vanoy C. Welch,
"The DoD 4.8 kbps Standard (Proposed Federal Standard 1016),"
in Advances in Speech Coding, ed. Atal, Cuperman and Gersho,
Kluwer Academic Publishers, 1991, Chapter 12, p. 121-133.

Campbell, Joseph P. Jr., Thomas E. Tremain and Vanoy C. Welch, "The
Proposed Federal Standard 1016 4800 bps Voice Coder:  CELP," Speech
Technology Magazine, April/May 1990, p. 58-64.

For U.S. FED-STD-1016 (4800 bps CELP) _realtime_ DSP code
and information about products using this code, contact:

John DellaMorte
DSP Software Engineering
165 Middlesex Tpk, Suite 206
Bedford, MA  01730
1-617-275-3733
1-617-275-4323 (fax)
dspse.bedford@channel1.com

DSP Software Engineering's code can run on a DSP Research's Tiger 30 board
(a PC board with a TMS320C3x and analog interface suited to development work)
or on Intellibit's AE2000 TMS320C31 based 3" by 2.5" card.

DSP Research                Intellibit
1095 E. Duane Ave.          P.O. Box 9785
Sunnyvale, CA  94086        McLean, VA  22102-0785
(408)773-1042               (703)442-4781
(408)736-3451 (fax)         (703)442-4784 (fax)
-- 
.............................................................................
; Joe Campbell  N3JBC  jpcampb@afterlife.ncsc.mil  74040.305@compuserve.com ;
; My opinions are mine!   Happiness = Reality - Expectations, Click & Clack ;
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Thanks, Joe!

He also told me in email that if you wish it to also be available via
anonymous ftp, tell Bob Fenichel.

--
Dan Frankowski                dfrankow@cs.umn.edu


------------------- end quote ---------------------
-- 
Sauron is alive in Argentina!

In <rdippold.716147581@qualcom> rdippold@qualcom.qualcomm.com (Ron Dippold) writes:

>Here's some excellent info on CELP in general:

>---------------------- start quote ------------------------

>Joe Campbell says:

>In case you'd like to experiment with CELP, you can obtain a software
>implementation of the 4800 bps Fed Std CELP coder for free:

>The U.S. DoD's Federal-Standard-1016 based 4800 bps code excited linear
>prediction voice coder version 3.2 (CELP 3.2) Fortran and C simulation source
>codes are now available for worldwide distribution at no charge (on DOS
>diskettes, but configured to compile on Sun SPARC stations) from:

>Bob Fenichel
>National Communications System
>Washington, D.C.  20305
>1-703-692-2124
>1-703-746-4960 (fax)

>In addition to the source codes, example input and processed speech files
>are included along with a technical information bulletin to assist in
>implementation of FS-1016 CELP.  (An anonymous ftp site is being considered
>for future releases.)

>Copies of the actual standard "Federal Standard 1016, Telecommunications:
>Analog to Digital Conversion of Radio Voice by 4,800 bit/second Code
>Excited Linear Prediction (CELP)" are available for $2.50 each from:

>GSA Rm 6654
>7th & D St SW
>Washington, D.C.  20407
>1-202-708-9205

>The following articles describe the Federal-Standard-1016 4.8-kbps CELP
>coder (it's unnecessary to read more than one):

>Campbell, Joseph P. Jr., Thomas E. Tremain and Vanoy C. Welch,
>"The Federal Standard 1016 4800 bps CELP Voice Coder," Digital Signal
>Processing, Academic Press, 1991, Vol. 1, No. 3, p. 145-155.

>Campbell, Joseph P. Jr., Thomas E. Tremain and Vanoy C. Welch,
>"The DoD 4.8 kbps Standard (Proposed Federal Standard 1016),"
>in Advances in Speech Coding, ed. Atal, Cuperman and Gersho,
>Kluwer Academic Publishers, 1991, Chapter 12, p. 121-133.

>Campbell, Joseph P. Jr., Thomas E. Tremain and Vanoy C. Welch, "The
>Proposed Federal Standard 1016 4800 bps Voice Coder:  CELP," Speech
>Technology Magazine, April/May 1990, p. 58-64.

>For U.S. FED-STD-1016 (4800 bps CELP) _realtime_ DSP code
>and information about products using this code, contact:

>John DellaMorte
>DSP Software Engineering
>165 Middlesex Tpk, Suite 206
>Bedford, MA  01730
>1-617-275-3733
>1-617-275-4323 (fax)
>dspse.bedford@channel1.com

>DSP Software Engineering's code can run on a DSP Research's Tiger 30 board
>(a PC board with a TMS320C3x and analog interface suited to development work)
>or on Intellibit's AE2000 TMS320C31 based 3" by 2.5" card.

>DSP Research                Intellibit
>1095 E. Duane Ave.          P.O. Box 9785
>Sunnyvale, CA  94086        McLean, VA  22102-0785
>(408)773-1042               (703)442-4781
>(408)736-3451 (fax)         (703)442-4784 (fax)
>-- 
>.............................................................................
>; Joe Campbell  N3JBC  jpcampb@afterlife.ncsc.mil  74040.305@compuserve.com ;
>; My opinions are mine!   Happiness = Reality - Expectations, Click & Clack ;
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

>Thanks, Joe!

>He also told me in email that if you wish it to also be available via
>anonymous ftp, tell Bob Fenichel.

>--
>Dan Frankowski                dfrankow@cs.umn.edu


>------------------- end quote ---------------------
>-- 
>Sauron is alive in Argentina!

Does this code also allow lossy music compression/decompression ?

I'm still looking for a lossy WAV compressor/decompressor !

Best regards Stefan Hartmann.
email to: leo@zelator.in-berlin.de


-- 
 
*************************************************************
*  Stefan Hartmann       This is how to contact me:         *
*  EMAIL: leo@zelator.in-berlin.de                          *
*  Phone : ++ 49 30 344 23 66      FAX : ++ 49 30 344 92 79 *
*************************************************************