[JLM]

A friend asked me why I use ALAC.  As per JohnR's primer I choose ALAC.  His primer mentions both are lossless but ALAC takes up less disk space.  So if both are lossless sound quality shouldn't vary (famous last words of audio).  My friend says all his research indicates AIFF sounds better. 

So what's the story?

TIA

[srb]

The story is it depends who you ask!

I've had people tell me WAV sounds better than AIFF!  But wait ..... both of those formats are identical uncompressed linear PCM differing in only the file header (and AIFF supports metadata tags and WAV does not).

At least it makes some kind of sense that maybe ALAC (or compressed FLAC) could sound worse because even though they're lossless formats, they are compressed and need to be uncompressed on the fly, so there is some processing overhead.  But I suspect that might only be heard on devices with either/or marginal hardware processing and playback software in this day and age.

Still it's pretty easy to test on your own system.  Make a few playlists with just two versions of the same song in each - one ALAC and one AIFF and differentiate them with (ALAC) and (AIFF) appended to the end of the songname.  Then enable Random Play and Repeat for the playlist.  With multiple presses of Next you won't know which version you're listening to.

Your own ears will either hear a difference on your own system's unique combination of hardware and software or they won't - and that will be the only real meaningful answer.

Steve

[JLM]

Thanks for the info Steve.  And good suggestion.

But I'm such a computer dunce (and have had a rough year with El Capitan, Amarra, etc.) that I'm leary to try it for myself.

[dB Cooper]

Almost all "research" conducted by audiophiles wouldn't meet the loosest imaginable definition of the word "scientific" and consists of comparisons made when the researcher knew exactly which file/cable/component/etc/etc he/she was listening to. Expectation bias rules the day, and the "night and day" difference you hear when you insert that $1000 power cable into your system suddenly becomes impossible to identify when you don't know which is which. So my first question would be, What "research"? Who did it- and how?

I use both FLAC and Apple Lossless (mostly FLAC) and seriously doubt anybody could tell the difference in an honest (blind) comparison.

[JLM]

Yep.  I'm an engineer by training and 36 years of practice, so you can guess that my first questions were: "We who?  What gear? Blind A/B/X testing?"

And being a "speaker guy", don't lose sleep over such issues, but was curious.

[zoom25]

I've talked about this subject previously on both Audiocircle and Head-fi. I've done lossless format testing on both a Bryston BDP-1 and various Macbook Pros (w/SSD) and iMacs running Audirvana Plus and Amarra. I know that Audirvana and the Bryston load up the song into the buffer in advance, so it shouldn't matter what format the material is in, as long as its lossless

 I compared ALAC, FLAC, AIFF, and WAV...I didn't end up testing FLAC uncompressed however.

Few observations: ALAC requires the most amount of processing. Try loading an hour long lossless track into Audirvana Plus and you can  visually see the buffer line. ALAC takes the most time. FLAC is a bit better. AIFF and WAV are significantly better. WAV very slightly beats AIFF.

After testing and some assistance, I found ALAC to the be the worst, followed by FLAC, and very close between AIFF and WAV...Ultimately went with WAV. It works and sounds the best across every system from any year or platform.

Full discussion here: http://www.audiocircle.com/index.php?topic=139572.0

Have fun

[JLM]

What's the gap between best and worst? 

I can barely hear the difference between Amarra and iTunes on my $7000 system (but I'm a speaker guy so swapping out a cap on my ambience tweeter made way more difference to me).

[zoom25]

Just a bit of background:

I tested this on both my speakers (Amphion One15's (Amphion amp and cables) in a treated room) as well on my HD 800...I've found the differences to vary among DACs and connection (USB vs AES). My Amphion rig running Bryston BDP-1 and Dangerous Source with a linear power supply plus the HD 800 is about $8-9k...it's getting there in terms of resolution. I've also played around on the big full range ATCs.

Not to mention the room treatment REALLY REALLY HELPS. I have also started to isolate my devices from vibrations nowadays which can make a difference, even with solid state devices.

Even with asynchronous DACs, I have found a spectrum of variance between how they handle different formats when being fed from different sources. In my testing I've used the digital out from these devices: Bryston BDP-1 (Linux), Apple TV, Marantz CD5004, Macbook Pro Mid 2010, Macbook Pro Retina Late 2012, iMac Late 2009, IPTV box (Android), Playstation 2, Playstation 3, cheap Toshiba DVD player.

For example, on my Emotiva DC-1, I have the option of choosing asynchronous and synchronous. The Dangerous Source is asynchronous. These are just some components I have at the house currently. In the past I have tried other DACs as well like Arcam D33, Bryston BDA-2, NAD M51, Naim DAC, some new Luxman item, all being through a decent server like Bryston BDP-2 or NAD M50, which I compared with my Macbook Pro feeding the same files in different formats. Most, if not all, of these are asynchronous DACs.

Results:

I found some DACs to be immune to changes like my Emotiva DC-1 to different digital devices. It doesn't matter if its being fed from a BDP-1 or a Macbook Pro. It sounds good regardless with all devices.

The Dangerous Source on the other hands sounds thin and weightless as if the timing is off, transients and phase aren't as tight when fed from cheap devices. The Amphions are excellent at showing phase problems. Many have found their imaging to better than ATCs. When fed from BDP-1's AES, the sound is at its absolute best.

Why this matters for formats?

With the Emotiva DC-1 as well as other DACs previously mentioned that don't really care what digital device its being fed with, also have a tendency to not care about whether the format of the file is FLAC, AIFF, WAV, or ALAC.

There seems to be a connection between these two things (responsiveness to file formats and digital input device).

MOST IMPORTANTLY - THE SOUND:

The absolute best sound and resolving details I've gotten from among high end DACs are the ones that have been sensitive to changes in file formats as well as the digital devices they are being fed from.

My Dangerous Source with the linear power supply destroys the DC-1 when both are being fed with the BDP-1. With cheap devices, I may at times prefer the DC-1's sound.

What does this mean?

Pick either one:

1) The asynchronous devices that do change in sound between different formats and digital devices are poorly designed

OR

2) These devices actually have the resolution capability to show the difference between file formats/digital devices that other non-sensitive DACs won't show.

It's one or the other.

My measurements, results, and advice on file formats:

On devices that are sensitive to changes between file formats: avoid ALAC. I'm running multiple Macs and both AIFF and ALAC are Apple formats, yet they aren't my favourite.

Compressed: FLAC every time over ALAC. ALAC takes the most time to buffer (in relative terms) and requires the most processing on both Macbooks and Bryston. I like the sound of FLAC more than ALAC. ALAC really sounds claustrophobic. My Macbook Pro's are no slouch either as they are all 16+ GB RAM, with SSDs, and i7 processors.

Uncompressed: Both AIFF and WAV are a bigger jump from ALAC and FLAC. Spend a day listening to WAVs and then go back to their counterparts FLAC and you can listen in long periods or quickly swap around between intervals of the same file. From an analytic point, its always the mids and vocals where you can most easily spot the difference. WAV sound more natural without a glare. FLAC almost colours the sound in a beautiful way but kind of puts some glare on it. WAV is just as is. No magic. On my systems, I gain this resolution in upper mids in female voices.

WAV and AIFF: I personally go with WAV. WAV works with every single device. I have never run into compatibility issues with WAV.

Methodology:

I have blind tested in the past and posted my methodology on Head-fi. If needed, I'll try to go dig it up if needed. SPL was matched.

Its also possible to run these tests independently by yourself. I do them from time to time on my Bryston BDP-1 through its software (Mpad and Manic Moose).

Use XLD to convert the file to different formats and dump them in a software where you can shuffle and it doesn't show you the file extension/format. So all you do is pick an interval of 10-15 secs in the song and change very quickly among all 4 files for that selected interval. Do that as many times as you need and then when you think you know which format is which, you can look at the properties of that file to see if its FLAC, AIFF, WAV, or ALAC.

[zoom25]

What's the gap between best and worst? 

To explicitly answer your question in numerical form. If WAV is considered 100%, then ALAC is in the 60-80% and FLAC is in the 80-85%.

[kgturner]

So what's the story?

Disk space is cheap. I chose WAV. Sure FLAC, ALAC, and WAV *should* be the same, but with WAV I don't have to wonder if I'm missing anything.

Kevin T

[Starchild]

Disk space is cheap. I chose WAV. Sure FLAC, ALAC, and WAV *should* be the same, but with WAV I don't have to wonder if I'm missing anything.

Kevin T

I agree with your sentiments even though I am a Flac user.  I use JR Media Center and I don't think I'm missing anything.  The deal killer for me with Wav files is not sonic but file management.  The inability to retain meta data makes it impossible to organize a large library.  For this reason, if I were going to go uncompressed, I'd use an AIFF file format.

[Jonathon Janusz]

I spent a little time looking in to metadata and WAV files. and am confused.  Some information online says that WAV does support metadata, others say it doesn't.  Some also say that cover art can also be embedded, some not.  Is a WAV music file just a file name and the PCM data and nothing else?

zoom25, as someone who has decided on using WAV, how do you organize your music library if WAV doesn't support any metadata?  Do you just - not - organize it, or do you use some other software or service (iTunes, ROON or something like it) to analyze your untagged music files and manage them outside of any embedded metadata?

This conversation is timely because at some point soon I really need to move my existing music collection off my current iTunes based system and on to a Salk Streamplayer.  Thanks for bringing it up!

[srb]

Is a WAV music file just a file name and the PCM data and nothing else?

Yes - for the most part.  There is an extended WAV format that does support some metadata tags, but player support for it is few and far between.

If you are going to stick with one audio player such as iTunes or JRiver for life, then the program can store metadata for WAV in its internal database.  But if you change players or want to use an additional program to play the folders of WAV files, the metadata won't be with the files.

I've mentioned this in other previous posts but it bears mentioning again - If you use AIFF, MP3, AAC or ALAC in iTunes - IF the album artwork is (a) downloaded automatically when ripping or (b) you use the manual Get Album Artwork command, the artwork is stored in the iTunes database even though these file formats do support embedded artwork.

For that reason, if I get artwork in iTunes by either of those methods, I will then manually copy the artwork and paste in back into the Artwork Tab, in which case it will be embedded within the file.  That way if I point JRiver to the iTunes folder, artwork will be seen because it is embedded in the file.

Steve

[zoom25]

When I convert my FLAC files or whatever else to WAV, I use XLD and write tags for it. This consists of "INFO chunk" and "id3 chunk" and encoding is UTF-8. It's very, very straight forward in XLD. Under "Batch" section, I also put "0" under subdirectory search depth, which makes batch conversion a piece of cake. You can have folders inside another folder which is inside another folder and so on...simply drop the containing folder into XLD and everything inside it will be converted with the same name and all folders will be in the same place. It's the best thing I've used in the past 15 years.

Also, with this, Audirvana automatically gets all it needs. I keep all my music on an external drive and connect it to the Bryston BDP-1. The drive gets shared on the network and use that to also sync with Audirvana. Everything is preserved. For artwork, I'm very particular and so are some softwares on naming and capitalizations. I keep only one artwork image inside each folder and call it "Cover.jpg"....the "C" is capitalized, and its "jpg" not jpEg." These things matter! Pick one and be consistent. It really goes a long way in future proofing and saving time.

Also, Manic Moose on Bryston automatically looks it up. So it doesn't end up mattering.

I could have gone with AIFF to save headache, but decided against it for two reasons:

1) Compatibility: Occasionally I will run into some devices that don't support AIFF. WAV is unparalleled in both consumer and pro world.

2) Sound: I'll take both AIFF and WAV over FLAC and ALAC. Although comparing AIFF and WAV was very, very hard. There might as be no difference...but in long term listening I found WAV slightly more dynamic. Everything I've said previously in this thread, I 100% stand behind.

However, sound between AIFF and WAV is still out for me and DEFINITELY requires controlled testing. Although, I fear the approach of most ABX intervals might not be temporally appropriate to the subtle differences between the two. I advice to listen for longer period. This is the one topic where I will hold my tongue.

So just to get the bug out of my system, I went with WAV. I have never run into dissatisfaction problems with WAV.

I know how seriously sketchy all of this sounds, but this is my honest opinion after months or relentless testing under many circumstances and with bunch of gear of all price ranges. You name it and I have done it.

[kgturner]

The inability to retain meta data makes it impossible to organize a large library.  For this reason, if I were going to go uncompressed, I'd use an AIFF file format.

I use a program called Tag&Rename which allows me to enter metadata for my WAV files. It works fine in JRiver and Squeezecenter. It won't work in iTunes or Foobar though so there are trade offs depending on which playback program you intend to use.

Kevin T

[Jonathon Janusz]

Thanks for the replies.  My tentative plan when moving to the Salk streamer is to move my library (all AIFF, ripped with iTunes), entirely over to the local storage on the Salk and play it from there.  I don't have anywhere near the library a lot of people here do, so putting everything on a single local drive isn't and won't be a problem for a long time (if ever).

My questions regarding WAV came in that reading about zoom25's preference (even slightly) for it for sound quality reasons has me considering converting my AIFF over to WAV.  I have no interest in the rabbit hole of multimedia immersion software like ROON is making popular right now, so I don't need crazy amount of tagging info, but keeping basic album/artist/track name/cover art info (basically giving me functionality comparable to a plain old ipod) is enough to make me happy.

When the time comes to move the music library, it will be a good excuse to clean up the mess itunes has made of the tags that are already in place, and to verify all the rips to fix any errors, as I really haven't done any even basic care and feeding on my music library drive; I've just been feeding discs into a mac mini and letting a very old version of itunes do its thing.  I just got to thinking that if I was going to the trouble to take care of the data, if I could pick up a sound quality improvement via conversion it would be a good time to do it if I could have it without making a mess of the library itself in the process.

Last, I've always found artwork on my own and added it manually to files right after they are ripped, so I know all of my art is embedded with the files right now.  So, just to make sure I have this right, if I were to convert to WAV, all that would get chopped off and I would have to make sure I stored the art files separately in respective album folders to keep functionality?

[zoom25]

The last time I brought up file formats, I had a guy with Naim setup (including player) do these formats and not only did he prefer it in the same order as me, but also by the same amount. ALAC dead last. FLAC better. Both AIFF and WAV better. WAV slightly wins. The analog format of the digital world.

I will also add that this uncompressed vs compressed lossless discussion has been repeated frequently in various forums and some have also found uncompressed better. A subset from that group may have even found WAV to the better counterpart against AIFF....However, I have never come across anyone preferring something over WAV when it related to sound quality. ZERO.

Also, about album art. I think its a good practice to have an album art file inside each folder regardless of WAV or AIFF or FLAC. Even with my FLAC library for the past 15+ years, I've always maintained this practice and its always paid off. Some software need a discrete image to load up album art even if the FLAC image has artwork. It takes a little bit more time, but its really worth the effort if you want to maintain universality for
your work for many years to come.

[testsound]

I've talked about this subject previously on both Audiocircle and Head-fi. I've done lossless format testing on both a Bryston BDP-1 and various Macbook Pros (w/SSD) and iMacs running Audirvana Plus and Amarra. I know that Audirvana and the Bryston load up the song into the buffer in advance, so it shouldn't matter what format the material is in, as long as its lossless

 I compared ALAC, FLAC, AIFF, and WAV...I didn't end up testing FLAC uncompressed however.

Few observations: ALAC requires the most amount of processing. Try loading an hour long lossless track into Audirvana Plus and you can  visually see the buffer line. ALAC takes the most time. FLAC is a bit better. AIFF and WAV are significantly better. WAV very slightly beats AIFF.

After testing and some assistance, I found ALA

C to the be the worst, followed by FLAC, and very close between AIFF and WAV...Ultimately went with WAV. It works and sounds the best across every system from any year or platform.

Full discussion here: http://www.audiocircle.com/index.php?topic=139572.0

Have fun

Have you got a chance to test uncompressed FLAC as well?