FLAC files from 16/44 CD (WAV) files are smaller than MP3s.

[esldude]

Why would I lie to you.  Don't you think this is getting ridiculous now.  Jeez!

 

Here is what I used. 

https://www.amazon.com/J-S-Bach-French-Overture/dp/B0000028NL

 

They list it as a 1995 CD.  The CD says it is a 1972 recording. 

 

 

[Kal Rubinson]

I am running out of popcorn.

[mansr]

3 minutes ago, dalethorn said:

No it is not the answer.  The answer is when you have something that can be verified by everyone.  A CD that I can buy and test myself.

You already provided files that can be verified by anyone. What exactly do you think is wrong with them?

[dalethorn]

2 minutes ago, esldude said:

Why would I lie to you.  Don't you think this is getting ridiculous now.  Jeez!

 

Here is what I used. 

https://www.amazon.com/J-S-Bach-French-Overture/dp/B0000028NL

 

They list it as a 1995 CD.  The CD says it is a 1972 recording. 

 

 

 

I will buy the CD as soon as I find one in stock from someone who can ship overnight.  This supplier has only a 2 week option.

[dalethorn]

Just now, mansr said:

You already provided files that can be verified by anyone. What exactly do you think is wrong with them?

 

That is what I asked here, to be investigated.  So far all I see are excuses that do not address the FLAC anomaly.  Now if you have an additional CD to recommend that compresses to 25-35 percent in FLAC, please tell us.

[dalethorn]

8 minutes ago, Kal Rubinson said:

I am running out of popcorn.

 

That's what happens when everyone ducks the issue.

[dalethorn]

9 minutes ago, esldude said:

Why would I lie to you.  Don't you think this is getting ridiculous now.  Jeez!

 

Here is what I used. 

https://www.amazon.com/J-S-Bach-French-Overture/dp/B0000028NL

 

They list it as a 1995 CD.  The CD says it is a 1972 recording. 

 

 

 

I bought it from a Amazon provider who promises to deliver Jan 31st.  This should be good, unless someone here closes the topic.

[esldude]

6 minutes ago, dalethorn said:

 

I bought it from a Amazon provider who promises to deliver Jan 31st.  This should be good, unless someone here closes the topic.

I could have sent you a couple tracks from it.  You seem to wish to have your own copy which is okay.  

[dalethorn]

1 minute ago, esldude said:

I could have sent you a couple tracks from it.  You seem to wish to have your own copy which is okay.  

 

Sending tracks would not be verifying.  But when I rip the tracks and convert to FLAC, there may be some differences from yours, which will be interesting to compare.

[dalethorn]

6 minutes ago, esldude said:

I could have sent you a couple tracks from it.  You seem to wish to have your own copy which is okay.  

 

BTW, any investigation of the tracks I posted on Dropbox is very much appreciated, but, since I have no idea what JRiver did to make that WAV file, any conclusions from those investigations are still tentative until I find how that WAV can convert to FLAC at 23 percent size while no CD rip ever converted that small.

[esldude]

https://www.amazon.com/Glenn-Gould-plays-Bach-Variations/dp/B0085MK2M6/ref=sr_1_3?ie=UTF8&qid=1517009355&sr=8-3&keywords=glenn+gould+goldberg+variations

 

Here is one you can get overnight. 

 

I did track 16.  As a wav it is 53.1 meg and as a Flac 16.4 meg.  I didn't do any filtering to it.  I had it ripped from CD as a FLAC on my hard drive.  Opened it in Audacity.  Saved it as a 16 bit wav.  Saved it again as a level 5 16 bit FLAC.  31% size as a FLAC.    I used a 3rd order rolloff at 5 khz.  Saved as a FLAC and this time 23% of the original size. 

[dalethorn]

2 hours ago, esldude said:

https://www.amazon.com/Glenn-Gould-plays-Bach-Variations/dp/B0085MK2M6/ref=sr_1_3?ie=UTF8&qid=1517009355&sr=8-3&keywords=glenn+gould+goldberg+variations

 

Here is one you can get overnight. 

 

I did track 16.  As a wav it is 53.1 meg and as a Flac 16.4 meg.  I didn't do any filtering to it.  I had it ripped from CD as a FLAC on my hard drive.  Opened it in Audacity.  Saved it as a 16 bit wav.  Saved it again as a level 5 16 bit FLAC.  31% size as a FLAC.    I used a 3rd order rolloff at 5 khz.  Saved as a FLAC and this time 23% of the original size. 

 

I already bought it.  I'll be doing a Level 5 FLAC conversion in Foobar2000.  I'll be genuinely shocked if the FLAC is below 40 percent of the WAV size.

 

I have always assumed that all CD WAV rips are the same - i.e. just what's on the CD, but who knows?  I use verify on my rips to make sure they're bit perfect.

 

If by some chance my FLACs aren't as small as you expect, then that will be one more thing to investigate.  Eventually I hope to be able to estimate the quality of downloaded files, DSD or otherwise, since I think we all know that recorded quality varies a lot, and it might not be obvious whether the download provider or their suppliers have "slimmed down" a recording before putting it out for download.

 

I hope too that this topic stays focused on compression issues and not go off into download quality etc.  I've seen plenty of that, but I might have to revisit some of that eventually.

[dalethorn]

2 hours ago, esldude said:

https://www.amazon.com/Glenn-Gould-plays-Bach-Variations/dp/B0085MK2M6/ref=sr_1_3?ie=UTF8&qid=1517009355&sr=8-3&keywords=glenn+gould+goldberg+variations

 

Here is one you can get overnight. 

 

I did track 16.  As a wav it is 53.1 meg and as a Flac 16.4 meg.  I didn't do any filtering to it.  I had it ripped from CD as a FLAC on my hard drive.  Opened it in Audacity.  Saved it as a 16 bit wav.  Saved it again as a level 5 16 bit FLAC.  31% size as a FLAC.    I used a 3rd order rolloff at 5 khz.  Saved as a FLAC and this time 23% of the original size. 

 

Oops - couldn't resist -bought that one too.  This will be twice as interesting.  Or maybe 123 percent as interesting.

[mansr]

17 hours ago, dalethorn said:

BTW, name one CD of solo piano music that will compress to 20+ percent of WAV size to FLAC.  I will buy it now and test it immediately when it arrives.

Here you go: https://www.amazon.com/Debussy-C-Complete-Preludes-Illumines/dp/B002XC7FQA/

 

Compresses to about 24% of original size.

[dalethorn]

42 minutes ago, mansr said:

Here you go: https://www.amazon.com/Debussy-C-Complete-Preludes-Illumines/dp/B002XC7FQA/

 

Compresses to about 24% of original size.

 

Ordered.  That is now the 3rd CD I've ordered on these recommendations.  3 should be enough, yes?  Some will arrive within 2 days, and I'll do the bit-perfect rips and use Foobar2000 with the v1.3.1 FLAC converter (c.2014) at Level 5 to get a representative compression.

 

If nothing else, and they compress close to your expectations, it will be worth the effort.

[dalethorn]

Last night I managed to get FLAC Level 5 compressions to mid-upper 30-percent of WAV size for two piano CDs I dug out of storage (still waiting on the new CDs for Monday delivery).  These were a Rachmaninov BBC recording (Barry Douglas) from 2010, and a James Boyk "Tonalities of Emotion" SACD where I ripped the regular 16/44 CD layer.

 

While these compressions surprised me, and I don't see anything odd about the recordings (they were done with quality in mind apparently), it will be far more surprising if a quality recording on the new CDs I ordered can compress (as above) to less than 25 percent.

[CatManDo]

Sorry, I didn't read the entire thread, and I'm not an encoding expert. Just an experienced user.

 

Could this have to do with the treatment of mono recordings (the 1955 Goldberg Variations are mono)?

 

Every CD has two channels. In case of a mono source, the L and R channels are identical. FLAC recognizes this, therefore FLAC encodings of mono recordings are very small.

 

Some MP3 encoders (depending on the settings) treat CDs from mono and stereo sources the same way, as an input signal with two different channels. Therefore the MP3 output file can be bigger than necessary.

 

Maybe this can explain why for certain CDs, FLAC is smaller than MP3. Normally, FLAC is approximately twice the size of a 320kbs MP3.

[CatManDo]

12 hours ago, dalethorn said:

Last night I managed to get FLAC Level 5 compressions to mid-upper 30-percent of WAV size for two piano CDs I dug out of storage (still waiting on the new CDs for Monday delivery).  These were a Rachmaninov BBC recording (Barry Douglas) from 2010, and a James Boyk "Tonalities of Emotion" SACD where I ripped the regular 16/44 CD layer.

 

While these compressions surprised me, and I don't see anything odd about the recordings (they were done with quality in mind apparently), it will be far more surprising if a quality recording on the new CDs I ordered can compress (as above) to less than 25 percent.

 

Piano recordings can be compressed a lot with FLAC. I have many solo piano albums with around 40min that are just 100-150MB.

 

On the other extreme, CDs with a lot of dynamic compression (i.e. the majority of current pop releases) result in FLAC sizes of 450MB and more.

[mansr]

1 hour ago, CatManDo said:

Could this have to do with the treatment of mono recordings (the 1955 Goldberg Variations are mono)?

The files dalethorn posted are not mono, nor is the Debussy album I suggested. However, piano recordings still have strongly correlated channels, and this helps compression.

[esldude]

3 hours ago, CatManDo said:

Sorry, I didn't read the entire thread, and I'm not an encoding expert. Just an experienced user.

 

Could this have to do with the treatment of mono recordings (the 1955 Goldberg Variations are mono)?

snip...

The Goldberg Variations disc I suggested is a two CD set.  One of the 1955 sessions and one from the 1981 sessions (which were digitally recorded).  I used the stereo track from the latter.

 

Digital tracks will compress better (on a limited range instrument like piano) because there is less high frequency low level noise from tape or noisy analog electronics (like when tubes were used).  Noise doesn't compress much at all.  

 

[dalethorn]

3 hours ago, CatManDo said:

 

Piano recordings can be compressed a lot with FLAC. I have many solo piano albums with around 40min that are just 100-150MB.

 

On the other extreme, CDs with a lot of dynamic compression (i.e. the majority of current pop releases) result in FLAC sizes of 450MB and more.

 

It all gets much more interesting when FLAC sizes are as small as or smaller than MP3s.  When I mentioned than I converted an MP3 to WAV and then converted the WAV to FLAC, and the FLAC was 2.4 times the MP3 size, one user suggested that the conversion from MP3 to WAV generated some anomolous artifacts which could not be so compressed.

 

With that in mind, and noting that my DSD piano tracks aren't exactly "simple or low complexity" - rather, the background has been made extremely quiet through some time-intensive physical preparation, and allegedly the compression is smart enough that the foreground (the piano) compresses exceptionally well.

 

So I got to thinking "Why shouldn't MP3s be high resolution, since they are sometimes bigger than corresponding FLAC files - all we need do is teach the compressors to ignore background noise and those pesky digital 'artifacts' as noted above?"  But I could extend this question to say "Why isn't the compressor capturing more of the sound in its 23 percent compression, so that the 16/44 FLAC's content is more comparable to the source?"  EDIT: The 16/44 WAV probably not the FLAC.

[dalethorn]

My wording above is gnarly - in better words, if a high-res recording is so low in complexity that a down-rezd 16/44 WAV file is extremely compressible, why then can't we put more info into the 16/44 WAV file.?

[mansr]

51 minutes ago, dalethorn said:

It all gets much more interesting when FLAC sizes are as small as or smaller than MP3s.

Not if you understand how those compression methods work.

[esldude]

Or to look at it another way Dale, I took the Goldberg Glenn Gould track.  Put a brickwall filter at 6 khz.  There is nothing other than low level noise above that.  Noise at -100 db.  The RMS level of the track was -27 db before I did it. And the same after the brick wall filter.  All that is up there is noise or harmonics buried in the noise.  One could have recorded this with a 14 khz sampling rate and it may not have been audibly different.  14 khz vs 44 khz is right in that low 30% range it compressed into.  

[dalethorn]

2 hours ago, mansr said:

Not if you understand how those compression methods work.

 

I'm not so much interested in "how they work", I'm more interested in why - if the 16/44 FLAC compression is going to be so extreme, why more resolution can't be stuffed into the process that the FLAC is generated from.