Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: How does Lossless compression to audio really work? (Read 10937 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

How does Lossless compression to audio really work?

I am a University of Plymouth Music Technology student,
I wish to create and test a hypothesis comparing lossless to lossy audio compression against uncompressed files.

I am currently questioning how lossless compression really works?
But any advice would be appreciated

How does Lossless compression to audio really work?

Reply #1 –
Good luck comparing lossless files to uncompressed files. There's a reason why the term "lossless" is used. If you search the site you will find threads where the audibility of lossless compressed music has been shot to death. Lossless compression can be compared to simple zip-archiving. You compress the bits that can be optimized as long as you don't throw anything away. Therefore you can't specify a bitrate you want the encoder to use as opposed to lossy compression.

Lossy compression however is an estimate of the audio that serves to compress the audio as good as possible within a given size (bitrate).

So I guess your hypothesis should regard lossy compression compared to lossless compression (since obvoisly lossless and uncompressed are exactly the same).
Can't wait for a HD-AAC encoder :P

How does Lossless compression to audio really work?

Reply #2 –
cheers for the reply,
I know it claims to be 'lossless' but are we talking to exact 100% replica or within a margin or error?
on a set of headphones, the difference may be negligible nay transparent,

but tested over a live rig, would the 'margin of error' become more apparent?

How does Lossless compression to audio really work?

Reply #3 –
Exact 100% replica. No errors. No margins.

How does Lossless compression to audio really work?

Reply #4 –
Use the Search on this forum: lossless has been proven to be bit perfect to the original WAV file.  Because they are identical, from an audibility standpoint they are 100% the same, hence "no loss" or "lossless," no "margin of error" and even "no error."  You can verify this yourself by checking the MD5's, line up the samples visually, run a null test, run an ABX listening test...its all been done.  Your hypothesis is not very useful because the lossless = the original.  You can compare the lossy to the original and try to find some new news there, but I would scrap your original hypothesis as not worth the time.

I will never understand why the concept of lossless compression is so mysterious to people...
Was that a 1 or a 0?

How does Lossless compression to audio really work?

Reply #5 –
cool, cheers

So my hypothesis should conclude lossless having no difference to uncompressed; with quality,
stereo spread, dynamic range etc. across all of my different delivery systems?

How does Lossless compression to audio really work?

Reply #6 –
Hi
just do a bit-for-bit comparison.
You will see no further comparison (listening test) will be necessary ;-)

How does Lossless compression to audio really work?

Reply #7 –
Quote from: cosmo25 on
cool, cheers

So my hypothesis should conclude lossless having no difference to uncompressed; with quality,
stereo spread, dynamic range etc. across all of my different delivery systems?

Yes.

How does Lossless compression to audio really work?

Reply #8 –
If you are interested in a demonstration in the mathematical sense of the term, you can start from the point where lossless is made of mathematical formulas where a file made of many elements A, can be converted to a file of elements B with the property that using the decoder,
those B elements used as input generate the original elements A. You either will reach the point that either the encoder/decoder is not lossless, or the hypothesis is true.

If you want to demonstrate it in a sonically way, you cannot demonstrate it with an ABX, since you cannot demonstrate a non-difference with this tool. The other suggestions are just alternate ways of saying the same than the mathematical explanation above.

There is no point in checking differences in
Quote
quality, stereo spread, dynamic range etc.
since a non-difference cannot generate differences.

How does Lossless compression to audio really work?

Reply #9 –
HA Wiki - Lossless

C.
PC = TAK + LossyWAV  ::  Portable = Opus (130)

How does Lossless compression to audio really work?

Reply #10 –
Imagine a zip-archive modifying your files for the sake of compressability...
Can't wait for a HD-AAC encoder :P

How does Lossless compression to audio really work?

Reply #11 –
University student, eh? Better learn how to research, then. (For the most basic details, not to mention technicalities.)

How does Lossless compression to audio really work?

Reply #12 –
Quote from: cosmo25 on
I know it claims to be 'lossless' but are we talking to exact 100% replica or within a margin or error?
on a set of headphones, the difference may be negligible nay transparent,


*Another* one of these?

Do they grow these people in a field or something?

How does Lossless compression to audio really work?

Reply #13 –
Quote from: ech3 on
Do they grow these people in a field or something?

Kinda makes a university education seem a great deal less impressive, doesn't it? [/font]

How does Lossless compression to audio really work?

Reply #14 –
Im sorry, thought the idea of people helping people was a good thing..

I started this post as a way to obtain ideas about lossless compression,
while I was researching at the same time..
I do now feel it was naive of me posting before I had started my research
for that I apologies.

but in my research; including other forums, people do claim to hear a difference
between AIFF and Lossless files, Lossless being Quote: 'sound dead and flat.. a
bit lifeless'

This could be just a placebo effect but I feel further testing could be done in testing
lossy, lossless and AIFF? or am I barking up the wrong tree?


FYI: if your just here to mock and not really help then shame on you..
.. hope your attitude has got you far in life.. x

How does Lossless compression to audio really work?

Reply #15 –
You may want to read at least this thread.
Can't wait for a HD-AAC encoder :P

How does Lossless compression to audio really work?

Reply #16 –
Lossless means lossless. The decoded stream it bit-identical to the original stream.

Have you ever encountered a zip or rar archive? Did you notice how the zip file is smaller then the original files.
If you mistrust the definition of loss-less compression, write a small text file, compress it with 7zip, decompress it and then look whether the text looks "dead and flat.. a bit lifeless".

How does Lossless compression to audio really work?

Reply #17 –
These seem interresting as well: here, here and here
Can't wait for a HD-AAC encoder :P

How does Lossless compression to audio really work?

Reply #18 –
Quote from: odyssey on
Imagine a zip-archive modifying your files for the sake of compressability...


LossyWav + zipped WAVs ain't too shabby.
Copy Restriction, Annulment, & Protection = C.R.A.P. -Supacon

How does Lossless compression to audio really work?

Reply #19 –
cheers everyone

especially Odyssey the links are very helpful, and appreciated.

I thought I was on my own and had posted a highly childish question

I do see my previous angle was off and I will reanalyse my testing

As much as it will pain people though, it will take a test (on not just myself)
to confirm it before I install i-tunes and start listen to MP4(/AAC) over my CD's.



on returning true from my hypothesis I WILL buy an i-pod. and drop my argument 

How does Lossless compression to audio really work?

Reply #20 –
Quote
I thought I was on my own and had posted a highly childish question

Correct.

Quote
As much as it will pain people though, it will take a test (on not just myself)
to confirm it before I install i-tunes and start listen to MP4(/ACC) over my CD's.
on returning true from my hypothesis I WILL buy an i-pod. and drop my argument

Test for what? Encode a lossless file, then bit-compare its output to the original source (either directly, or by decompressing it back to WAV/whatever and comparing that to the original). There's no need to involve anyone else in it, either; mathematics doesn't act differently depending on who's observing.

And what on Earth is all this talk of buying an iPod about?

How does Lossless compression to audio really work?

Reply #21 –
I also wonder why you mention mp4/cd tests. When you have accepted that lossless are just as good as uncompressed audio, you can move on to comparing them using ABX-tools. You may want to search for listening tests at this site, since most lossy codecs do such a good job that noone can tell the difference in a blind test.

Maybe you should do some research on placebo now you're at it
Can't wait for a HD-AAC encoder :P
 

How does Lossless compression to audio really work?

Reply #22 –
Just to clarify further confuse the issue.

There is lossless = mathematically lossless, bit-exact, bit-perfect, exactly reversible, just like ZIP-archiving. This is what the word lossless means on this forum.

And there is "lossless" = high quality lossy = "perceptually lossless" = transparent = audibly indistinguishable from original. This is how the word "lossless" is used (misused) in marketing of lossy encoders sometimes.

How does Lossless compression to audio really work?

Reply #23 –
Quote from: cosmo25 on
but in my research; including other forums, people do claim to hear a difference
between AIFF and Lossless files, Lossless being Quote: 'sound dead and flat.. a
bit lifeless'

This could be just a placebo effect but I feel further testing could be done in testing
lossy, lossless and AIFF? or am I barking up the wrong tree?

Most lossless codecs include ReplayGain information in the resulting file tags and while WAV and Audio CDs contain no such information and lots of hardware and software players support ReplayGain adjustment that might be one of the reasons why the two sound differently, also maybe just the implemented decoder these people use is faulty.

The best way to test whether lossless compression is truly lossless is to compress a WAV file and then decompress it, then run a bit-comparison analysis between the original file and the file you got from decompression. You'll see that there are no differences between the two.

How does Lossless compression to audio really work?

Reply #24 –
Quote from: kdo on
Just to clarify further confuse the issue.

There is lossless = mathematically lossless, bit-exact, bit-perfect, exactly reversible, just like ZIP-archiving. This is what the word lossless means on this forum.

And there is "lossless" = high quality lossy = "perceptually lossless" = transparent = audibly indistinguishable from original. This is how the word "lossless" is used (misused) in marketing of lossy encoders sometimes.
The former is not just what the word lossless means on this forum, but what it means everywhere. That it might be misused (do you have an example/source?) is unfortunate, but does not change its meaning, and is no more worthy of consideration than false or disingenuous advertising of any other type.