I'm looking for a lossless audio codec to capture video. The codec must be
available from the video capture software(s), thus it must be integrated and
be available in my WindowsXP lists of codecs. It must be capable of
compressing 48khz sampling rate.

I use various software to capture/process HuffYUV video such as VirtualDub
1.5.9, Adobe Premiere 6.0, Sonic Foundry Vegas 4.0. Once the processing
done, I use TMPGenc Plus 2.59 to compress the high quality final to MPEG1.

I installed various lossless audio codecs such as FLAC 1.1.0, Monkey 3.97,
La 0.4 and Shorten 2.3b. These codecs work only in command-line or GUI
software. Their codecs don't appear in other video/audio processing
software. Thus can't be called by other software, so they are pretty much
useless for me.

I want to use a lossless audio as an "intermediate" codec. (much like I use
lossless HuffYUV codec for the video track). Presently I capture the audio
track of the video in PCM audio at 48kHz, 16-bit, Stereo. Multiplied this by
8 hours of capture + another 8 hours to process. At this my secondary 80Gig
HD runs out of space in no time.... 50% audio compression would satisfy me.

Joseph Brown wrote:

I'm looking for a lossless audio codec to capture video. The codec must be
available from the video capture software(s), thus it must be integrated and
be available in my WindowsXP lists of codecs. It must be capable of
compressing 48khz sampling rate.

I use various software to capture/process HuffYUV video such as VirtualDub
1.5.9, Adobe Premiere 6.0, Sonic Foundry Vegas 4.0. Once the processing
done, I use TMPGenc Plus 2.59 to compress the high quality final to MPEG1.

It might be relevant for you to look into another adobe program:
Audition.

I want to use a lossless audio as an "intermediate" codec. (much like I use
lossless HuffYUV codec for the video track). Presently I capture the audio
track of the video in PCM audio at 48kHz, 16-bit, Stereo. Multiplied this by
8 hours of capture + another 8 hours to process. At this my secondary 80Gig
HD runs out of space in no time.... 50% audio compression would satisfy me.

Not my turf, perhaps someone else can suggest something. Audio files do
process very poorly, they tend to be mathematically too chaotic and the
processing overhead may be problematic. 192 or 320 kbit mp3 comes to
mind as a possibly acceptable compromise.

Kind regards

Peter Larsen


--
************************************************** ***********
* My site is at: http://www.muyiovatki.dk *
************************************************** ***********

Joseph Brown wrote:
snip
I want to use a lossless audio as an "intermediate" codec. (much like I use
lossless HuffYUV codec for the video track). Presently I capture the audio
track of the video in PCM audio at 48kHz, 16-bit, Stereo. Multiplied this by
8 hours of capture + another 8 hours to process. At this my secondary 80Gig
HD runs out of space in no time.... 50% audio compression would satisfy me.

If 50% will do it for you then use the u-law codec. U-law, while not lossless in
the literal sense, will convert 16 bit linear PCM to 8 bit logarithmic scale
samples. If you need even more compression, then ADPCM will work great at that
high sample rate, and cut 75% off of 16 bit PCM.

--
Phil Frisbie, Jr.
Hawk Software
http://www.hawksoft.com

In article ,
Nomen Nescio ] wrote:
I'm looking for a lossless audio codec to capture video. The codec must be
available from the video capture software(s), thus it must be integrated
and
be available in my WindowsXP lists of codecs. It must be capable of
compressing 48khz sampling rate.

Aren't you asking for the impossible? A lossless audio codec that does
compression? That's an oxymoron. If there's compression, it can't be
lossless. Unless you mean pkzip-type of compression.



************************************************** *****************
** The only good velocity-switch is an inaudible velocity-switch **
************************************************** *****************

Phil Frisbie, Jr. wrote:
Joseph Brown wrote:
snip

I want to use a lossless audio as an "intermediate" codec. (much like
I use
lossless HuffYUV codec for the video track). Presently I capture the
audio
track of the video in PCM audio at 48kHz, 16-bit, Stereo. Multiplied
this by
8 hours of capture + another 8 hours to process. At this my secondary
80Gig
HD runs out of space in no time.... 50% audio compression would
satisfy me.


If 50% will do it for you then use the u-law codec. U-law, while not
lossless in the literal sense, will convert 16 bit linear PCM to 8 bit
logarithmic scale samples. If you need even more compression, then ADPCM
will work great at that high sample rate, and cut 75% off of 16 bit PCM.


....or, as others have suggested, just use an MP3 codec at some high
bitrate (like 320kbits/sec). That will shrink the audio by
approximately 80% and have a completely unnoticible effect on the audio
quality, even after multiple iterations.

Cheers,

C

"SomeGuyOnTheInternet" wrote in message
...
In article ,
Nomen Nescio ] wrote:
I'm looking for a lossless audio codec to capture video. The codec
must be
available from the video capture software(s), thus it must be
integrated
and
be available in my WindowsXP lists of codecs. It must be capable of
compressing 48khz sampling rate.

Aren't you asking for the impossible? A lossless audio codec that does
compression? That's an oxymoron. If there's compression, it can't be
lossless. Unless you mean pkzip-type of compression.

Maybe you should do a simple web search before you speak.

http://flac.sourceforge.net/comparison.html

"SomeGuyOnTheInternet" wrote in message
...
In article ,
Nomen Nescio ] wrote:
I'm looking for a lossless audio codec to capture video. The codec
must be
available from the video capture software(s), thus it must be
integrated
and
be available in my WindowsXP lists of codecs. It must be capable of
compressing 48khz sampling rate.

Aren't you asking for the impossible? A lossless audio codec that does
compression? That's an oxymoron. If there's compression, it can't be
lossless. Unless you mean pkzip-type of compression.

There are certainly lossless compression schemes, like your zip example.
GIF compression can also be lossless if the source image has =256 colors.
DVD-Audio uses the MLP algorithm, Meridian Lossless Packing to get some
measure of compression, maybe around 50%.

Lossless compression will inherently compress less than lossy compress
(obviously). Also, the compression rate of all lossless schemes is data
dependent. In the best case, like for example an all-black image file, you
can get huge compression ratios. On the other hand, if the data is totally
random, you might end up with no compression, or even a slightly larger file
due to the compression scheme overhead. In general, the more "random" the
data looks, the less you can compress it. Try zipping a JPEG image to see
what I mean.

As for the OP's question, the MLP I mentioned is the only truly lossless
algorithm I know of for audio. And I don't know if it's available as a
Windows codec. Try searching the web, or you could start he
http://www.meridian-audio.com/

As for using u-law, I wouldn't recommend it as there are certainly losses in
the signal-to-noise ratio (despite the best efforts of the exponential
companding). MP3 or WMA with very high bit-rate would be my suggestion.

SomeGuyOnTheInternet wrote:

Aren't you asking for the impossible? A lossless audio codec that does
compression? That's an oxymoron. If there's compression, it can't be
lossless. Unless you mean pkzip-type of compression.

Please take care to use the wording bitreduction when it is about
bitreduction.


--
************************************************** ***********
* My site is at: http://www.muyiovatki.dk *
************************************************** ***********

"Peter Larsen" wrote in message
...
SomeGuyOnTheInternet wrote:

Aren't you asking for the impossible? A lossless audio codec that does
compression? That's an oxymoron. If there's compression, it can't be
lossless. Unless you mean pkzip-type of compression.

Please take care to use the wording bitreduction when it is about
bitreduction.

Now that's a good one. Invent a word and then tell everyone to take care to
use it.

Aren't you asking for the impossible? A lossless audio codec that does
compression? That's an oxymoron. If there's compression, it can't be
lossless. Unless you mean pkzip-type of compression.

YES, there are many bit-accurate lossless audio compressors out there.
Pkzip-type file compressors are not optimized for audio.

Usually, lossless audio compressors provide 2:1 compression ratio (50%).
Yes, they do exist.

The most popular a
FLAC from http://flac.sourceforge.net/
Mokey audio from http://www.monkeysaudio.com/
La from http://www.lossless-audio.com/
and may others found he http://flac.sourceforge.net/comparison.html

What I'm looking is for a DLL lossless audio codec that can be integrated in
my WinXP OS and be available and called from any application. I haven't
found any yet.

...or, as others have suggested, just use an MP3 codec at some high
bitrate (like 320kbits/sec).

I noticed that when I load an AVI video with a 48kHz 16-bit PCM audio track,
the highest mp3 that VirtualDub will allow me to encode is 48kHz, stereo,
128 kbps. It can't go higher, even with the "show all formats" enabled.

I installed the extra mp3 codecs as well as Fraunhofer IIS .mp3 Producer
Professional which allows extra bitrates to appear. They don't appear
everywhere.

SO, is there any lossless DLL audio codec that can be integrated in
my WinXP OS and be available and called from any application? I haven't
found any yet.

On Tue, 25 Nov 2003 17:15:28 -0500, "Joseph Brown"
wrote:

...or, as others have suggested, just use an MP3 codec at some high
bitrate (like 320kbits/sec).

I noticed that when I load an AVI video with a 48kHz 16-bit PCM audio track,
the highest mp3 that VirtualDub will allow me to encode is 48kHz, stereo,
128 kbps. It can't go higher, even with the "show all formats" enabled.

I installed the extra mp3 codecs as well as Fraunhofer IIS .mp3 Producer
Professional which allows extra bitrates to appear. They don't appear
everywhere.

SO, is there any lossless DLL audio codec that can be integrated in
my WinXP OS and be available and called from any application? I haven't
found any yet.


If I were to summarize the replies so far, it would be that...

1. There lossless compressors for audio, but none that work as codecs.
2. Using a lossy compression codec at a sufficiently high bit rate is so
close to as good as lossless that you shouldn't worry about the difference.

My only question on that point would be whether the codecs in question take
all 24 bits into account during encoding/decoding, or if they only deal with
16 bits.

Joseph Brown wrote:
Usually, lossless audio compressors provide 2:1 compression ratio (50%).
Yes, they do exist.

The most popular a
FLAC from http://flac.sourceforge.net/
Mokey audio from http://www.monkeysaudio.com/
La from http://www.lossless-audio.com/
and may others found he http://flac.sourceforge.net/comparison.html

What I'm looking is for a DLL lossless audio codec that can be integrated in
my WinXP OS and be available and called from any application. I haven't
found any yet.

You did not look hard enough http://mediaxw.sourceforge.net/Enchanced/about.htm

Media XW says it installs codecs for Monkey's Audio and FLAC that can be used by
Windows programs.


--
Phil Frisbie, Jr.
Hawk Software
http://www.hawksoft.com

On Tue, 25 Nov 2003 23:40:19 GMT, "Phil Frisbie, Jr."
wrote:

Joseph Brown wrote:
Usually, lossless audio compressors provide 2:1 compression ratio (50%).
Yes, they do exist.

The most popular a
FLAC from http://flac.sourceforge.net/
Mokey audio from http://www.monkeysaudio.com/
La from http://www.lossless-audio.com/
and may others found he http://flac.sourceforge.net/comparison.html

What I'm looking is for a DLL lossless audio codec that can be integrated in
my WinXP OS and be available and called from any application. I haven't
found any yet.

You did not look hard enough http://mediaxw.sourceforge.net/Enchanced/about.htm

Media XW says it installs codecs for Monkey's Audio and FLAC that can be used by
Windows programs.

Well, you have to look pretty hard, eh? How is it even possible, from that
page, to tell what mediaxw does or does not support.

You did not look hard enough
http://mediaxw.sourceforge.net/Enchanced/about.htm

Media XW says it installs codecs for Monkey's Audio and FLAC that can be
used by
Windows programs.

I downloaded the .exe installer and the .msi then installed.

The codecs don't appear in the ACM (Audio Codec Manager), therefore can't be
called by any windows program.

I found a lossless codec called LiteWave that's available in ACM form at:
http://www.clearjump.com/products/LiteWave.html
but it's not free... I e-mailed them and I'll wait for then to see how much
it costs.

I also found Ogg Vorbis in ACM from a German site:
http://www.pctip.ch/downloads/dl/24655.asp

This codec appears in the ACM windows codec manager! and can be called from
virtaully any software. But the problem is that it's lossy. I tried
encoding a video clip with the audio track set at 320kbps with this coded
and I noticed a slight drift...

Isn't there any plans to port FLAC, Monkey or La or any other open source
lossless audio codec to ACM?

The most widely supported lossless codec is FLAC, which isn't saying much
since the only program I know that can use it for any real-time process is
WinAmp.

I don't mind real-time or not processing. All I'm looking for is for an ACM
version of FLAC. That is that it could be called by *any* windows
audio/video program supporting windows Audio Codec Manager (ACM) such as
CoolEdit, SoundForge, VirtualDub, Premiere, Vegas, TMPGEnc, RealMedia
Producer, etc.

Is there any ACM FLAC ? or any plans to port it to ACM?

Joseph Brown wrote:
Is there any ACM FLAC ? or any plans to port it to ACM?

The people at http://www.openacm.org/ seem to be planning to.

By the way, doesn't Microsoft know that the acronym "ACM" already
means something fairly important in the computer world? :-)

- Logan

On Wed, 26 Nov 2003 03:20:33 GMT, Logan Shaw
wrote:

Joseph Brown wrote:
Is there any ACM FLAC ? or any plans to port it to ACM?

The people at http://www.openacm.org/ seem to be planning to.

By the way, doesn't Microsoft know that the acronym "ACM" already
means something fairly important in the computer world? :-)

- Logan

FWICT, they do that on purpose. Soon after the advent of eXtreme Programming
came Windows XP. After they put eXtreme Programming under the umbrella of
Agile development, Microsoft released a product called Agile. Microsoft also
calls their SQL server "SQL Server".

Hmmm... I wrote an MPEG Layer 2 ACM about 6 years ago. I should rummage
through all my old software and post it for people to download.

MP2 provides excellent quality at 384kbps. Please post it. One good place
to post iy would be: http://www.softpedia.com/

I thought only QDesign ever made the Mpeg-1 layer 2 ACM codec found he
http://www.broadcast.co.uk/product_i...roducts_id/116

"Joseph Brown" wrote in message
news
[snip]

I want to use a lossless audio as an "intermediate" codec. (much like I
use
lossless HuffYUV codec for the video track). Presently I capture the
audio
track of the video in PCM audio at 48kHz, 16-bit, Stereo. Multiplied this
by
8 hours of capture + another 8 hours to process. At this my secondary
80Gig
HD runs out of space in no time.... 50% audio compression would satisfy
me.

Perhaps I am missing something here... but 48kHz, 16 bit stereo would work
out to only a little over 5Gb for 8 hours worth of audio. However, your
HuffYUV video stream would have filled up your 80Gb drive a long while
before you came anywhere near the 8 hour mark! 50% compression would only be
saving you around 2.5Gb, which is only a bit over 3% of the total drive
size. I don't really understand the point!? ;(

FLY135 wrote:

"Peter Larsen" wrote in message
...
SomeGuyOnTheInternet wrote:

Aren't you asking for the impossible? A lossless audio codec that does
compression? That's an oxymoron. If there's compression, it can't be
lossless. Unless you mean pkzip-type of compression.

Please take care to use the wording bitreduction when it is about
bitreduction.

Now that's a good one. Invent a word and then tell everyone to take care to
use it.

I just paraphrased an AES recommendation.

--
************************************************** ***********
* My site is at: http://www.muyiovatki.dk *
************************************************** ***********

On Wed, 26 Nov 2003 11:59:50 +0100, Peter Larsen
wrote:

FLY135 wrote:

"Peter Larsen" wrote in message
...
SomeGuyOnTheInternet wrote:

Aren't you asking for the impossible? A lossless audio codec that does
compression? That's an oxymoron. If there's compression, it can't be
lossless. Unless you mean pkzip-type of compression.

Please take care to use the wording bitreduction when it is about
bitreduction.

Now that's a good one. Invent a word and then tell everyone to take care to
use it.

I just paraphrased an AES recommendation.

I believe the recommendation is due to the frequent confusion when referring
to data compression as compression, and the other party thinks you mean
dynamic range compression.

"Peter Larsen" wrote in message
...
FLY135 wrote:

"Peter Larsen" wrote in message
...
SomeGuyOnTheInternet wrote:

Aren't you asking for the impossible? A lossless audio codec that
does
compression? That's an oxymoron. If there's compression, it can't
be
lossless. Unless you mean pkzip-type of compression.

Please take care to use the wording bitreduction when it is about
bitreduction.

Now that's a good one. Invent a word and then tell everyone to take
care to
use it.

I just paraphrased an AES recommendation.

The problem with the term "bit reduction" is that it's usage to describe
lossy compression is almost ubiquitous, as a search of the web will show. I
doubt that hardly anyone would have trouble grasping that "lossless
compression" is well... lossless. Although now I'm thinking that wasn't the
distinction you were trying to make.

With the advent of widespead personal computing (not to mention multimedia)
I would imagine that the more people associate compression with
bit-reduction than the dynamic range reduction. Much to the chagrin of the
AES.

PS. The "inventing a word" thing was sarcastically referring to the lack of
a hyphen or space in "bit reduction". In reflection... It was a dumb joke.

Joseph Brown wrote:

I'm looking for a lossless audio codec to capture video. The codec must be
available from the video capture software(s), thus it must be integrated and
be available in my WindowsXP lists of codecs. It must be capable of
compressing 48khz sampling rate.

snip

I installed various lossless audio codecs such as FLAC 1.1.0, Monkey 3.97,
La 0.4 and Shorten 2.3b. These codecs work only in command-line or GUI
software. Their codecs don't appear in other video/audio processing
software. Thus can't be called by other software, so they are pretty much
useless for me.

I don't know much about current Windows codecs but wrote such a thing
for Quicktime a few years ago. There were some necessary workarounds
that would have made it a chancy product to support commercially. The
difficulty was that the audio media handlers would allocate and read
chunks of data based on a fixed, preregistered compression factor.
Since a lossless codec can't guarantee compression it was necessary to
register the compression factor as 1. As a result during live capture
no disk space or io activity was actually saved. When recompressing an
existing quicktime file it WAS possible to fool the handler into
writing variable length audio blocks, but during playback of such
files the handler would still present the codec with fixed blocks,
typically padded with subsequent video frames. Excess disk activity
would occur when the video media handler would read these frames a
second time, particularly when playing from a CD. So until audio
tracks are structured like video tracks, with variable length frames,
seamless lossless compression will be difficult.


I want to use a lossless audio as an "intermediate" codec. (much like I use
lossless HuffYUV codec for the video track). Presently I capture the audio
track of the video in PCM audio at 48kHz, 16-bit, Stereo. Multiplied this by
8 hours of capture + another 8 hours to process. At this my secondary 80Gig
HD runs out of space in no time.... 50% audio compression would satisfy me.

50% is about average for continuous analog-mastered music, the 2-3
LSBs being mostly noise. When these bits are significant 3x
compression is more likely, to 5x in softer passages. Of course,
lossless compression is particularly effective on high quality
narration tracks.

FLY135 wrote:

Please take care to use the wording bitreduction when it is about
bitreduction.

Now that's a good one. Invent a word and then tell everyone to take
care to use it.

I just paraphrased an AES recommendation.

The problem with the term "bit reduction" is that it's usage to describe
lossy compression is almost ubiquitous, as a search of the web will show.

And?

I doubt that hardly anyone would have trouble grasping that
"lossless compression" is well... lossless. Although now I'm
thinking that wasn't the distinction you were trying to make.

The "data guys" means avoiding repeating the same information multiple
times in a file, thereby making it smaller when they say compression and
look for arc, zoo, zip, lha, rar to do it with. The "audio guys" mean
keeping the darn trombone under control and look for something orban,
valley people or dbx or altec valvy to do it with. This is two uses of
the same magic term. That must be enough.

*peg, Atrac, windowsmedia, whatever is all about making files smaller by
discarding data that "don't matter".

With the advent of widespead personal computing (not to mention multimedia)
I would imagine that the more people associate compression with
bit-reduction than the dynamic range reduction. Much to the chagrin of the
AES.

Yes. In this newsgroup we could try to avoid confusing one another, it
must be enough that we confuse innocent bystanders.

PS. The "inventing a word" thing was sarcastically referring to the lack of
a hyphen or space in "bit reduction". In reflection... It was a dumb joke.

Not really, it was to the point and relevant because it must be
somebody's question to all of this.


Kind regards

Peter Larsen

--
************************************************** ***********
* My site is at: http://www.muyiovatki.dk *
************************************************** ***********

On Wed, 26 Nov 2003 05:09:42 GMT, Steve Jorgensen wrote:
On Wed, 26 Nov 2003 03:20:33 GMT, Logan Shaw
wrote:

Joseph Brown wrote:
Is there any ACM FLAC ? or any plans to port it to ACM?

The people at http://www.openacm.org/ seem to be planning to.

By the way, doesn't Microsoft know that the acronym "ACM" already
means something fairly important in the computer world? :-)

- Logan

FWICT, they do that on purpose. Soon after the advent of eXtreme Programming
came Windows XP. After they put eXtreme Programming under the umbrella of
Agile development, Microsoft released a product called Agile. Microsoft also
calls their SQL server "SQL Server".

the next major open source project competing against MSFT ought to name
its product "pointy-haired software".

On Wed, 26 Nov 2003 11:05:48 GMT, Steve Jorgensen
wrote:

I believe the recommendation is due to the frequent confusion when referring
to data compression as compression, and the other party thinks you mean
dynamic range compression.

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.
The other bit-reduction is making the sample-width smaller by one or
more bits. And this affects the sound as it reduces the dynamic range
of the sound (every bit you chop off, you loose 6 dB).

Confusing huh?

cheers

-martin-
--
filmmaker/DP/editor,
Sydney, Australia

http://www.pictocrime.com

On Thu, 27 Nov 2003 10:36:17 +1100, Martin Heffels
wrote:

On Wed, 26 Nov 2003 11:05:48 GMT, Steve Jorgensen
wrote:

I believe the recommendation is due to the frequent confusion when referring
to data compression as compression, and the other party thinks you mean
dynamic range compression.

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.
The other bit-reduction is making the sample-width smaller by one or
more bits. And this affects the sound as it reduces the dynamic range
of the sound (every bit you chop off, you loose 6 dB).

Confusing huh?

Yup - I hadn't thought of that one. I guess we just have to spell out what we
mean from the following:

Dynamics compression
Lossless data compression
Lossy data compression

Someone should come up with a standard shorthand for these things.

"Steve Jorgensen" wrote in message
...
Dynamics compression
Lossless data compression
Lossy data compression

Someone should come up with a standard shorthand for these things.

If somebody just states what they mean in the terminology you just used,
then it's perfectly clear. When someone says bit reduction, now that's not
so clear.

Martin Heffels wrote:

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.

And there is no bit reduction, all the bits are there, they are just
written differently.

The other bit-reduction is making the sample-width smaller by one or
more bits. And this affects the sound as it reduces the dynamic range
of the sound (every bit you chop off, you loose 6 dB).

There are fewer bits afterwards, so this is bitreduction.

Confusing huh?

No.

-martin-


Kind regards

Peter Larsen

--
************************************************** ***********
* My site is at: http://www.muyiovatki.dk *
************************************************** ***********

"Peter Larsen" wrote in message
...
Martin Heffels wrote:

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.

And there is no bit reduction, all the bits are there, they are just
written differently.

Huh? Are you trying to say that lossless compression is not bit reduction?

FLY135 wrote:

"Peter Larsen" wrote in message
...

Martin Heffels wrote:

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.

And there is no bit reduction, all the bits are there, they are just
written differently.

Huh? Are you trying to say that lossless compression is not bit reduction?

That happens to be what I have said all along.

--
************************************************** ***********
* My site is at: http://www.muyiovatki.dk *
************************************************** ***********

It uses fewer bits to describe more bits. Of course you can't compress
data without using fewer bits (that's like saying you can eat without
eating).


FLY135 wrote:

"Peter Larsen" wrote in message
...

Martin Heffels wrote:


Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.

And there is no bit reduction, all the bits are there, they are just
written differently.


Huh? Are you trying to say that lossless compression is not bit reduction?

On Thu, 27 Nov 2003 14:41:15 +0100, Peter Larsen
wrote:

Martin Heffels wrote:

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.

And there is no bit reduction, all the bits are there, they are just
written differently.

That depends very much on which level of abstraction you happen to be looking
at. Certainly, the new data stream contains the same information as the old
one in a different representation, but the representation has fewer bits. If
one is talking about the representation, the number of bits has been reduced
even if the same amount of information is present.

Actually, to get technical, the compressed data stream has a bit less
information because it is dependent on more information encapsulated by the
compression algorithm about the context of sound.

On Thu, 27 Nov 2003 14:41:15 +0100, Peter Larsen
wrote:

Martin Heffels wrote:

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.

And there is no bit reduction, all the bits are there, they are just
written differently.

Of course there is bit-reduction. You end of with a file with less
bits then the original
That's what makes the term bit-reduction so confusing. Maybe we
should set-up a IEEE-committee to set a standard for the naming?

cheers

-martin-
--
filmmaker/DP/editor,
Sydney, Australia

http://www.pictocrime.com

On Sat, 29 Nov 2003 09:32:14 +1100, Martin Heffels
wrote:

On Thu, 27 Nov 2003 14:41:15 +0100, Peter Larsen
wrote:

Martin Heffels wrote:

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say that
there are ten zeros at this place. This leads to no reduction in
quality.

And there is no bit reduction, all the bits are there, they are just
written differently.

Of course there is bit-reduction. You end of with a file with less
bits then the original
That's what makes the term bit-reduction so confusing. Maybe we
should set-up a IEEE-committee to set a standard for the naming?

I can't tell if this was a serious suggestion or not, but it seems to me that
it probably should be. Currently, it takes at least 3 words to unambiguously
say what variant of compression you mean, and given the human tendency to
abbreviate, all those words will not be in the average statement made from one
engineer to another. It would be good to have some unambiguous, single-word
terms.

Martin Heffels wrote:

Of course there is bit-reduction. You end of with a file with less
bits then the original

Forgive the spelling, in danish it would spell: sofisteri.

That's what makes the term bit-reduction so confusing. Maybe we
should set-up a IEEE-committee to set a standard for the naming?

This is not complicated. If you have a file where the wordlengh varies
alongt the way instead of being constantly 16 bit, then you have a bit
reduced file.

cheers

-martin-
--
filmmaker/DP/editor,
Sydney, Australia

http://www.pictocrime.com

--
************************************************** ***********
* My site is at: http://www.muyiovatki.dk *
************************************************** ***********

"Peter Larsen" wrote in message
...
FLY135 wrote:

"Peter Larsen" wrote in message
...

Martin Heffels wrote:

Yep, one refers to bit-reduction as in what programs like WinZIP etc
do: if I got ten zeros in a row, I don't write ten zeros, but say
that
there are ten zeros at this place. This leads to no reduction in
quality.

And there is no bit reduction, all the bits are there, they are just
written differently.

Huh? Are you trying to say that lossless compression is not bit
reduction?

That happens to be what I have said all along.

Actually no, you haven't. You responded to a post that had nothing to do
with bit reduction in the context you are now using it, by telling him to
use the term bit reduction. Hence the confusion.

Peter Larsen wrote:

Martin Heffels wrote:

Of course there is bit-reduction. You end of with a file with less
bits then the original

That's what makes the term bit-reduction so confusing. Maybe we
should set-up a IEEE-committee to set a standard for the naming?


This is not complicated. If you have a file where the wordlengh varies
alongt the way instead of being constantly 16 bit, then you have a bit
reduced file.

Only if the average word length is smaller than 16 bits. That is an
implicit assumption. Implicit assumptions are the kind of thing that
allows one group of people to go around thinking that a term is clear
and unambiguous while other people (who do not share the implicit
assumptions) wonder which of the several possible meanings is the
intended one.

The same thing applies, in my mind, to the term "bit reduction".
Unless we make the assumption that were are talking about the bits
in the (linear PCM) samples, then the term is ambiguous. Since
bits are used in other ways in audio formats, it's not clear that
"bit reduction" should mean reducing the number of bits per sample.
It just means reducing the number of bits in some unspecified way,
and there's an infinite number of ways to do that.

IMHO, some term like "sample size reduction" would be much clearer.

- Logan