FLAC vs. AIFF: Any Difference?

I know they shouldn't sound different and yet they do on my equipment. I have compared numerous songs and every time the FLAC sounds choked. Too bad really, would be nice to save space.

 

What equipment and what songs? Can you give samples of the songs you are comparing so we can compare on our equipment? Thanks.

 

JRiver WASAPI to Stellar Gain Cell DAC with Sony MDR-7506s. In my experience the song doesn't matter, just pick anything you want to compare.

 

I just find it extremely difficult to believe a difference can be heard between FLAC and AIFF when the files contain the exact same data. I really doubt modern players or a computer are not up to the task of decoding the files without any jitter.

 

AIFF has always sounded better than ALAC to me. Just throwing that in the pot.

 

Details? The GPL literally says, as you mentioned, that anyone can use it at any time for any purpose. What happened with Samba?

In any case, FLAC is so widely-used and firmly-established that I wouldn't worry. If, for some reason, some sort of licensing issue arose, I'm pretty confident that someone would come up with a compatible replacement pretty quickly.

 

The issue with the GPL is that if you modify the source code in any way to use in your commercial project, you have to publish those modifications. This is not usually a problem, as most systems just use the whole library as is, with in integrated into how the underlying operating system handles shared libraries. However with embedded systems, there's just a single chunk of code, and technically the modified system is the entire code base.

Hence using FLAC in embedded music players could be a problem - hence Apple never used FLAC in their iPod range.

 

Are you possibly doing some form of DLNA or client server with your JRiver setup? If you're doing something like that it is possible to have some file formats transcode. So the WAV files could be getting transcoded in a situation like that.

 

I have posted this here severaltimes before. I took Titanic's 5.1 DTS-CD and ripped it to FLAC with dBpoweramp, all tracks ripped accurately. I ripped Independence Day original CD from 1996 HDCD encoded with all it features enabled.
I then played the FLAC rips by an Oppo BDP-93 that can play HDCD. The Titanic 5.1 DTS-CD ripped to FLAC decoded perfectly to DTS and I listening to the whole soundtrack ripped to check for any glitches, there were none. If you tamper in any way with a DTS data stream it doesn't decode at all.
I played the whole HDCD encoded Independence Day rip with the Oppo BDP-93 that can decode HDCD, the HDCD logo showed on display and the recorded sounded just like the actual CD HDCD decoded by the Oppo BDP-93 without a single glitch. Again,if you tamper with an HDCD data stream when playing it doesn't decoded as HDCD.
FLAC was able to encoded and decode both formats accurately and without a single glitch. What does this mean?

 

As a rock solid, but open minded technical person, I do not believe most claims about the difference between the sound of .wav vs. .flac, except for these causes:
1) MAYBE differences in bit depth: flac encoded at 16 bits vs. .wav at 24 bits or floating point (differences either way.) -- should be inaudible
2) Differences in noise generated by CPU during the decoding process -- normal signed 16/24/32 .wav or fp32/64 .wav vs. greater work (noise) decoding .flac

Do I strongly believe this? no. However, a few very strong advocates that I know ascribe to the above possibilities.

There might be other reasons, but again -- I am skeptical because proper dithered conversion between 16 bits vs 24 bits or floating point should be audible. If there is an audible difference in noise between decoding the flac vs. transparent conversion from .wav, then there is a noise problem in the system. Differences in noise caused by CPU, memory, etc should be inaudible (but possible.) When a CPU has current glitches in the 100amp range, anything is possible .

One personal anecdote -- once in a while my system does seem to produce a kind of transient distortion, definitely a glitch somewhere... Could that kind of noise be a reason for 'item 2' above? I wont' make claims either way, but I am skeptical of any differences in sound other than some strange glitch or excessive system noise. Given my own case of tinnitus, I doubt that I could detect a difference in system noise at the levels caused by '2' unless my system was really borked.

John

 

Probably people are just seeing the lower bitrate number displayed in their media player (of FLAC vs. uncompressed) and their mind automatically associates that with lower audio quality.

 

I think you're right, I've since changed my opinion after doing blind tests.

 

I've changed my opinion of blind tests after doing some blind tests where there is a difference and not being able to pass the blind test while being able to recognize those differences in non-blind testing. The brain listens and processes differently during a blind test than a regular listening comparison. Call it test anxiety.

 

Equivalent:
"When I know the source, I am able to identify a difference, but when I don't know the source, I am unable to identify a difference. Therefore blind testing is invalid."

 

I think my tests with a DTS-CD and an HDCD CD ripped to FLAC and then paying back as intended says it all, same in, same out.

 

My experience listening to headphone setups by people who believe too strongly in blind testing is that those setups never sound great or better than just average. Too much reliance on blind testing to choose gear and pair gear yields poor results.

 

I just to compared the same song a/b and I couldn’t spot a difference between AIFF and FLAC. I use a Bluesound Node 2i with a dac. On my old Squeezebox Touch AIFF sounded a little better, but now there is no difference. Perhaps the added CPU power of the Bluesound makes the difference.

 

I just did a new direct comparision here via my Bluesound Node 2i to my external Audio Alchemy Dac and to my ears this time I could maybe hear a touch more "air" and "transparency “ with voices on the AIFF file. It is very slight however.
I also did a comparison with Deezer HiFi FLAC and to my ears that sounded the worst of the three.
So it goes like this: AIFF, FLAC, FLAC via Deezer.

 

This hasn't stopped Sansa, Fiio and all of the other DAP producers from supporting FLAC. Apple has a severe case of Not Invented Here syndrome.

 

Using my digital device via the digital internet posting to this digital online forum, here is the digital typographical expression of the format FLAC:

FLAC

Using my digital device via the digital internet posting to this digital online forum, here is the digital typographical expression of the format AIFF:

AIFF

Using my digital device via the digital internet posting to this digital online forum, here is the digital typographical expression of the format ALAC:

ALAC

The differences are subtle, but note the additional clarity and fullness of “AIFF,” the slightly lean glare of “FLAC,” and the very minor but noticeably unpleasant (using a truly revealing and resolving display) congestion of “ALAC.”

 

Thankfully I am not concerned about believing any of these lossless file types sound different from one another. But I use all four file types as a Mac user:

• WAV - rip CDs with pre-emphasis to apply correction EQ using the XAct app. Convert to ALAC to import into iTunes for database management.
• AIFF - import purchased files from Bandcamp to confirm I capture any hi res files that may not be in ALAC which I download at the same time from Bandcamp. If identical, I discard the AIFF file and import the ALAC file into iTunes for database management.
• FLAC - rip multichannel audio files using DVDAE as FLAC and remain as FLAC since I do not use iTunes to manage this database. It is a file in the same folder as iTunes Music database folder, so Roon just reads the FLAC/ALAC files from all folders contained in the "Music" folder.
• ALAC - import all rebook CDs using iTunes (if batch ripping) or dBpoweramp (if only a few CDs) for iTunes database management.

At the end of the day, I mainly use ALAC because I like iTunes to manage my database (I am a column browser addict) and keep multichannel files as FLAC so that I can make sure that hardware will play nicely with them, as all the metadata is manipulated by me using DBPA or XAct.

 

FYI, if you rip a CD using iTunes, it does de-emphasis automatically.

 

I was determined to hear a difference in a FLAC and an AIFF file, spent all afternoon listening to all sorts of music. I could never hear any difference at all, and I really tried, as I had already made the statement the AIFF file had to be better, it just had to be, but after hours of listening, I could not hear that difference, FLAC, in my opinion is equal in every way to an AIFF file, just much smaller!

 

Not true

 

It is true in my tests. I burned a test CD with FLAGS PRE added to the cue sheet and iTunes applied de-emphasis when ripping it.