24/96 Files? - Page 4

I suspect it may have something to do with the LP mastering.

here is a view of a needle drop that I'm doing right this minute, this is from Audio Fidelity's new release of Sade's Diamond Life.  Audio Fidelity makes a point of trying to get the original analog masters for their releases.

look at the difference:

In general, high-resolution audio isn't upsampled from 16/44.1, so I don't understand your logic, stv014. The point is to see whether there's a difference between these versions. If there's an audible difference due to audio equipment, then so be it. No factor should be ignored, because these tests are only meaningful to the individual in the first place.

Quote:

Originally Posted by spaark 

In general, high-resolution audio isn't upsampled from 16/44.1, so I don't understand your logic, stv014. The point is to see whether there's a difference between these versions. If there's an audible difference due to audio equipment, then so be it. No factor should be ignored, because these tests are only meaningful to the individual in the first place.

The purpose of the test is to find out if there is a difference that is inherent to the 44.1/16 format. Without the conversion back to 96/24, if someone posts a positive result, there is always the question if it was only an artifact of the equipment used. By the way, if someone has equipment that performs poorly at 44.1 kHz, but better at 96 kHz (e.g. the headphone output of the Xonar Essence sound cards can be audibly noisy at 44.1 kHz with very sensitive headphones), it is easy to resample in real time in software. But, as I already mentioned, unwanted sample rate related differences that can result in a false positive ABX are not necessarily ones that affect sound quality, it can be simple things like the amount of delay before the playback starts, the presence or absence of a click at the beginning, etc.

Ah sorry, I misunderstood you. I thought you meant the 24/96 files you were to compare would be upsampled from 16/44.1, rather than the 16/44.1 files being upsampled but still considered "low resolution".

Quote:

Originally Posted by spaark 

Ah sorry, I misunderstood you. I thought you meant the 24/96 files you were to compare would be upsampled from 16/44.1, rather than the 16/44.1 files being upsampled but still considered "low resolution".

I would still prefer that the low res files be at redbook. Period.

The limitations of redbook, all of the limitations of redbook, should be experienced just as the benefits of high res should be experienced during any comparison IMHO.

In my mind that's what we're testing: can you hear an improvement, or even a difference, with different sample rates and bit depths.

Just my $0.02.

This was an interesting read.  I have lurked for awhile but I thought I would chime in with my first post.

I have been a vinyl/analog guy for a long time.  I recently decided to start transferring my music to a music server.  After experimenting I decided to use JRiver Jukebox for playback.   I tried Foobar2000, Winamp, and JRiver and I felt that JRiver had the best SQ.

I bought an ADL Esprit so I could use the DAC for CD redbook, stream music from a music server via USB, and record some needle drops.  Since the Esprit had a headphone amp, I dusted off my Sennheiser HD600s and I was pleasantly surprised how good they sounded.  

I decided to use Audacity for recording my need drops and save the files as 24/96 WAV files.  

When I converted the 24/96 WAV file to FLAC, the file size dropped and so did the SQ.  I used the FLAC converter built into JRiver.  For fun and games I converted the WAV to MP3 using Koyote Free MP3 Converter to see how that would impact the SQ.

I consider my system to be transparent and high resolution.  I asked my wife to listen to the same song (WAV, FLAC, MP3).  She consistently picked the 24/96 WAV as the best sounding.  That matched my impressions as well.  I would rate the results as WAV > FLAC > MP3.  I let a co-worker with good (and considerably younger) ears listen to the same songs in the different formats, and he said the 24/96 WAV sounded the best.  Although I like the sound of 24/96 WAV files, I dislike the lack of standards for tag support.

The audible improvements are minute and subtle.  The WAV sounded slightly better than FLAC, and notably better MP3.  The differences are as the OP stated, more air, better separation, increased dynamics, and a sense of you are there.   

A lot of people focus on analytics and measurements as a means of discerning audible differences.  A perfect example is USB cables (and digital cables) that supposedly carry 0/1s; however, my experience is they do have an audible difference.  For example, I found that Wywires USB sounds the best in my system, and that is supported by more than just my ears. 

This seems like a great community.  For an analog dinosaur like me I appreciate the wealth of knowledge.  I even took the plunge and bought a Clip + for my wife and for when we travel.  

Welcome to Head-Fi, sorry about your wallet!    

Quote:

Originally Posted by nelamvr6 

Welcome to Head-Fi, sorry about your wallet!    

Thanks, I have spent a bunch of money and time putting together my 2 channel/HT system and I decided for convenience sake to put everything on a music server.  I have ripped some of my favorite vinyl and I will continue to do so when I have some free time.  

For fun and games, I took one 24/96 WAV track (ripped from vinyl) and I converted it to FLAC and 320kbs MP3.  Here is a summary of my results (with the assumption the WAV sounded best and  serves as my reference): 

The difference in sound quality is subtle and seems proportionate to the resolution of your playback system.  I could easily live with any of the formats, especially for portable playback.

Quote:

Originally Posted by GregC 

 

 

 

 

Thanks, I have spent a bunch of money and time putting together my 2 channel/HT system and I decided for convenience sake to put everything on a music server.  I have ripped some of my favorite vinyl and I will continue to do so when I have some free time.  

 

For fun and games, I took one 24/96 WAV track (ripped from vinyl) and I converted it to FLAC and 320kbs MP3.  Here is a summary of my results (with the assumption the WAV sounded best and  serves as my reference): 

 

Format	File Size	Bit Rate	Percent SQ
WAV 24/96	98.6 MB	3270	100
FLAC	43.3 MB	1350	98
MP3	10.2 MB	320	94

 

The difference in sound quality is subtle and seems proportionate to the resolution of your playback system.  I could easily live with any of the formats, especially for portable playback.

FLAC and WAV should have identical playback. And I'm not just talking about in terms of human perception, it is literally the same.

Lossless compression is actually a relatively simple concept, and is similar to putting a file in a zip archive and removing it. Once the decompression is done, a FLAC file will contain every bit of PCM data that a WAV file has. 

You might be wondering how it could possibly have all the same data when it is half the file size, but just think of it as being made of shorthand code. It can be translated back to the original code, and the only drawback is that this takes some additional processing(which isn't relevant in regards to sound quality).

Quote:

Originally Posted by chewy4 

 

FLAC and WAV should have identical playback. And I'm not just talking about in terms of human perception, it is literally the same.

 

Lossless compression is actually a relatively simple concept, and is similar to putting a file in a zip archive and removing it. Once the decompression is done, a FLAC file will contain every bit of PCM data that a WAV file has. 

 

You might be wondering how it could possibly have all the same data when it is half the file size, but just think of it as being made of shorthand code. It can be translated back to the original code, and the only drawback is that this takes some additional processing(which isn't relevant in regards to sound quality).

I am not an audio engineer.  I have read FLAC is lossless, so it must have some compression algorithm to make file sizes smaller.  Could the process of expanding the file back to its original format and size impact the sound in any manner?  

Does FLAC have any type of bit rates or does it retain the original format?  I am curious because I want to encode a 24/96 WAV into FLAC and have it be playable on most devices.  For example, the Colorfly C3 says it supports FLAC (32-48 KHz 16-bit) and my encoder in JRiver does not give me an option to choose bit rates.

I did some research on FLAC and I see that the only options are compression levels 1 to 8.  If I compress a file above level 5 does the hardware requirement to decode it cause any degradation to sound?  I am curious if there is a way to see what compression level a FLAC files is saved at and convert it to level 5 when it is 6 or higher.

Quote:

Originally Posted by GregC 

I am not an audio engineer.  I have read FLAC is lossless, so it must have some compression algorithm to make file sizes smaller.  Could the process of expanding the file back to its original format and size impact the sound in any manner?  

 

Does FLAC have any type of bit rates or does it retain the original format?  I am curious because I want to encode a 24/96 WAV into FLAC and have it be playable on most devices.  For example, the Colorfly C3 says it supports FLAC (32-48 KHz 16-bit) and my encoder in JRiver does not give me an option to choose bit rates.

Nope, the process of expanding the file back to its original format and size wouldn't effect the sound quality at all. 

 

Lossless compression is just pattern analysis. It finds patterns in the data, and writes them in a shorter way. Then it has a program that knows what all these shorthanded expressions are, and expands them to their original data. For example a$10 could be expanded to aaaaaaaaaa. The way patterns and such are found in audio files is obviously a lot more complex than this, but that's just a simple of example of how the basis of it works.

 

I assume you mean bit depth(like 16 or 24 bit) and not bit rate(1411kbps etc) for your second question though? I'm not familiar with JRiver, but stv014 mentioned the SoX plugin for foobar which can change that. 

 

 

Quote:

Originally Posted by GregC 

I did some research on FLAC and I see that the only options are compression levels 1 to 8.  If I compress a file above level 5 does the hardware requirement to decode it cause any degradation to sound?  I am curious if there is a way to see what compression level a FLAC files is saved at and convert it to level 5 when it is 6 or higher.

 

Sound quality shouldn't be effected, no. It would just take longer to encode and decode. 

Quote:

Originally Posted by chewy4 

Nope, the process of expanding the file back to its original format and size wouldn't effect the sound quality at all. 

 

Lossless compression is just pattern analysis. It finds patterns in the data, and writes them in a shorter way. Then it has a program that knows what all these shorthanded expressions are, and expands them to their original data. For example a$10 could be expanded to aaaaaaaaaa. The way patterns and such are found in audio files is obviously a lot more complex than this, but that's just a simple of example of how the basis of it works.

 

I assume you mean bit depth(like 16 or 24 bit) and not bit rate(1411kbps etc) for your second question though? I'm not familiar with JRiver, but stv014 mentioned the SoX plugin for foobar which can change that. 

 

 

 

Sound quality shouldn't be effected, no. It would just take longer to encode and decode. 

Thanks for your replies Chewy.  It may be placebo but it seemed like I heard a slight difference between FLAC compression 8 and the original 24/96 WAV file.  There was not enough difference to make me not use FLAC for reduced file size and increased portability.  I am definitely a noobie and realize I have a lot to learn.  

I do know how pattern recognition and compression algorithms work because of my background is as a software developer. 

I was speaking of depth when I mentioned 16-bit versus 24-bit,  I just want to be informed if my original source is 24/96 WAV and I encode it as FLAC, would that prevent a player like the Colorfly C3 from playing my FLAC file since I read it only supports 32-48KHz 16-bit FLAC.

Quote:

Originally Posted by GregC 

Thanks for your replies Chewy.  It may be placebo but it seemed like I heard a slight difference between FLAC compression 8 and the original 24/96 WAV file.  There was not enough difference to make me not use FLAC for reduced file size and increased portability.  I am definitely a noobie and realize I have a lot to learn.  

 

I do know how pattern recognition and compression algorithms work because of my background is as a software developer. 

 

I was speaking of depth when I mentioned 16-bit versus 24-bit,  I just want to be informed if my original source is 24/96 WAV and I encode it as FLAC, would that prevent a player like the Colorfly C3 from playing my FLAC file since I read it only supports 32-48KHz 16-bit FLAC.

Glad to be of help, nice to see another dev on here. 

But yeah you would need to re-sample those FLAC files as 16/48(or 16/44.1) for them to work. 

Quote:

Originally Posted by GregC 

This was an interesting read.  I have lurked for awhile but I thought I would chime in with my first post.

 

I have been a vinyl/analog guy for a long time.  I recently decided to start transferring my music to a music server.  After experimenting I decided to use JRiver Jukebox for playback.   I tried Foobar2000, Winamp, and JRiver and I felt that JRiver had the best SQ.

 

I bought an ADL Esprit so I could use the DAC for CD redbook, stream music from a music server via USB, and record some needle drops.  Since the Esprit had a headphone amp, I dusted off my Sennheiser HD600s and I was pleasantly surprised how good they sounded.  

 

I decided to use Audacity for recording my need drops and save the files as 24/96 WAV files.  

 

When I converted the 24/96 WAV file to FLAC, the file size dropped and so did the SQ.  I used the FLAC converter built into JRiver.  For fun and games I converted the WAV to MP3 using Koyote Free MP3 Converter to see how that would impact the SQ.

 

I consider my system to be transparent and high resolution.  I asked my wife to listen to the same song (WAV, FLAC, MP3).  She consistently picked the 24/96 WAV as the best sounding.  That matched my impressions as well.  I would rate the results as WAV > FLAC > MP3.  I let a co-worker with good (and considerably younger) ears listen to the same songs in the different formats, and he said the 24/96 WAV sounded the best.  Although I like the sound of 24/96 WAV files, I dislike the lack of standards for tag support.

 

The audible improvements are minute and subtle.  The WAV sounded slightly better than FLAC, and notably better MP3.  The differences are as the OP stated, more air, better separation, increased dynamics, and a sense of you are there.   

 

A lot of people focus on analytics and measurements as a means of discerning audible differences.  A perfect example is USB cables (and digital cables) that supposedly carry 0/1s; however, my experience is they do have an audible difference.  For example, I found that Wywires USB sounds the best in my system, and that is supported by more than just my ears. 

 

This seems like a great community.  For an analog dinosaur like me I appreciate the wealth of knowledge.  I even took the plunge and bought a Clip + for my wife and for when we travel.  

please watch this : http://www.youtube.com/watch?v=BYTlN6wjcvQ

You're more likely to hear a difference because you assume there is a difference between a flac file and a wav file, we are more influenced by what we see than what we ear when testing gear or listening to music.

Our expectation is our greatest ennemy for audio, I mean saying there is an audible between wav and flac is just wrong, there is no difference, the same applies to playback programs like foobar or jriver, if you have wasapi on both, they sound the same, it's just a fact.

Analytics and measurements help us to discern what part of our experience is biased or not, that's just it. There have been people believing putting rocks on cables would increase the SQ just because they expected it would sound better. Science help us to confront fact to subjective opinions. Saying that a usb cable sound different from another just shows how biased your opinions are, audio is no magic.

It is difficult to admit it, but ABX tests can prove, even to the most hardcore subjectivists, that they paid for something they didn't need to improve SQ (cables, esoteric objects, magicians, etc)