AIFF Vs. WAV

[wgscott]

Well?

 

[Paul R]

I want to wait until Saturday, when I can relax and listen on two or three systems before making what I hear public.

 

-Paul

 

 

[twist222]

I have had a quick listen last night and have some initial thoughts but would like to spend a bit more time comparing and also seeing if the results are the same between the G5 powermac and my 2009 MacMini. I would also like to use a second person so I do not know what type of file is playing.

So far my comparisons have only been between to wave and aiff files. I use Amarra and it will not play CAF files.

 

Regards

 

Mark

 

[Paul R]

Miska kindly converted the WAV file to a Little Endian AIFF file, so I have replaced the CAF files in the download with AIFF files. All the files will now import and play in iTunes (Pure Music/Amarra, etc.)

 

You will need to re-download the ZIP file to get the new files and the correct afinfo.txt file. Click on the link below to download the files.

 

PLEASE BE AWARE

 

These files may not be redistributed.

They may be used only for the purposes of this test, on this forum.

Remember these are copyrighted materials!

 

Download Sample Files Here

 

 

 

Thanks for participating. It will be very interesting indeed to hear the files and see the results.

 

By the way, feel free to email me if you would like your results to be published anonymously. There is zero pressure here for anyone to defend the results - there is no right or wrong answer here!

 

 

 

 

 

 

 

 

 

[wgscott]

The old ones played in quick time, fwiw.

 

[twist222]

Firstly thanks to Miska for carrying out the conversion to provide us with a little endian AIFF file. Second thanks to Paul for posting the downloads and lastly thanks to Barry for allowing us to use one of his recordings to carry out the test. I appreciate Barrys support in allowing others to form there own opinion. Particularly when Barrys experience is that on his system he does not hear a difference.

 

Now let's just hope hope that all the people who posted so passionatly on this thread and the other silent majority who have been following the thread use these files, give it one more try and confirm yes or no if they hear an difference.

 

Whilst for most it will not be a proper double blind test it would be good to at least get the sample size up. It will be a real pity if only 2 or 3 people test and report back.

 

Now let me get down off my soap box and go and down load some new test files.

 

Happy testing

Mark

 

[driven -- by sound]

TGIF - O.k., will lend the group another set of ears later this evening...

 

[goldsdad]

Set Decibel to infinitely repeat the three files and took the dog out for about an hour. When I came back, the screen saver was active so I was almost literally listening blind. As interesting as the music is, after 9 iterations of the same 3.5 minutes, I needed a change.

 

I heard no reliable difference between the AIF-C (LE), AIFF (BE) and WAV.

 

Sometimes a new detail in the sound (either events, micro-detail or "quality", for lack of better words) of one file would seem to appear, but come the next files, I'd be able to here it in them because I was focused on hearing it. The same thing happens when listening to any single file of any recording a number of times; I can listen to a recording a dozen times with the same setup, then suddenly notice something that has always been there although I've been unaware of it, and from then onward I'll always hear it.

 

Obviously, my expectation bias toward hearing no difference may have prevented my hearing a genuine sonic variance due to file format.

 

I may have a tin ear or my speakers could be Bose under a cunning disguise .

 

Or, perhaps, after hearing a new aspect of sound in one file, when I was hearing it in subsequent files, that was my imagining it in these files. I could have imagined it in the first file. I could be imagining that I'm typing this. I could....

 

Or maybe the system was producing exactly (within reason) the same sound on every playback regardless of file format.

 

Sorry, my crude listening session cannot possibly help toward the drawing of conclusions in the WAV v AIFF conundrum, but I felt obliged to give it a go.

 

 

 

 

 

 

 

[Paul R]

I and I am sure everyone else appreciates the time and effort you put into that test.

 

Let me remind everyone, there is no need to defend your results, no matter what those results are.

 

Whether you hear a difference or not, everyone appreciates you taking the time to try it out and report your results.

 

Thanks!

-Paul

 

 

[bdiament]

I just want to say how glad I am about this thread and how proud I am to be a member here. Also, congratulations and big thanks to Chris for running such a great forum.

 

This thread seems to be such an extreme rarity among Internet audio threads. We have folks on both sides of an issue, some with strong feelings about their perspective. And yet, there is mutual respect and tolerance of opposing points of view.

 

How often does that happen on Internet audio fora? In my experience, pretty close to never. It always descends into folks making claims about other folks' experience (i.e. what they do or do not hear) -- always a sure sign, in my view, of a lack of confidence.

 

Here, we have folks talking about their own experience (the only one we can ever truly know and in my view, the only one we can defend logically). We have both ends of the spectrum and at the same time we have a friendly exchange. This might be a first in my experience of Internet audio fora.

 

What a great bunch of folks.

 

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com

 

 

[elcorso]

This was a very interesting research and experience for all of us.

 

But, Kote Moun Yo?

 

I guess listening to the same music, under the different file format!

 

My (very own) conclusions.

 

On the listening test, music player was Audirvana on Integer Mode, no upsampling, 2011 Mac Mini under SL 10.6.8, on a quite night after a tremendous rain and thunderstorm on the middle of the Rainforest, no hallucinating drugs, but some stimulation by boutique coffe and some tobacco cigarettes, after a normal nice day.

 

 

The music files were compared against Barry Diament 24/192 WAV "Kote Moun Yo?" from "Equinox". One instrument against real life play, Drums (Tuba on my country).

 

Third Price: AIF-C LE, noisy highs from the Iron Strikes (Ogan), no bass depth from the Drums (Tuba).

 

Second Price: AIFF-BE, less noisy highs from the Ogan, some more bass depth from the Drums. But very close to the:

 

First Price: WAV: Almost no distortion from the highs, but not all the extension it should have, the ultra highs notes remains reverberating in the air, and you can even listening the decay, but never like the 24/192 version. From the bass, the depthness is improved, and the hands hitting the leather in the Drums is closer to live listening to this instrument, and of course at some distance from the 24/192 WAV.

 

Since every file won a price I hope nobody will argue nothing, and this is not the meaning of my reply. My only mean is to contribute a little to the effort from Barry, Miska and Paul.

 

I guess every one has his own winner, and they will remains on his preferred format, as it should be.

 

Congratulations and thanks again,

 

Roch

 

PS/ After all this congratulations, I like to blame iTunes for his nasty encoding, Little Endian AIF-C.

 

[Paul R]

Part of the choice of the music is simple- it will be slightly unfamiliar to almost everyone, and that can enhance critical listening.

 

We are really interested in the differences some people can hear, such as you enumerated for us.

 

Thanks again - Paul

 

 

[jhwalker]

Players used: iTunes, Audirvana, Fidelia

2010 MacBook Pro, 256M SSD, 8G RAM, no other programs running

HiFiMan RE-272 IEMs plugged directly into MacBook Pro headphone out

 

Comments: Very cool track and full of audio cues that should have revealed any differences if they were there, but I didn't hear any. No difference between any of the files with any of these three players.

 

Analysis: On my system, with this playback chain, I heard no difference. I am *not* saying no one else can possibly hear a difference or that a different playback chain might not turn up some anomaly in processing that might reveal a difference. But I guess I'm lucky that either my system or my hearing is relatively insensitive to those types of changes

 

 

[Paul R]

The results are a mixed bag, because Karen can sometimes hear a difference I cannot. We spent a good deal of time testing these tracks on three different systems today.

 

The tests consisted of one play through with us knowing which track was which, and one play through NOT knowing which track was which. This was done by putting the tracks into a repeating playlist, putting the mouse on the next track button, and shutting the lid of the laptop till we could not see it. Then hitting the button randomly for 60 seconds, and listening to all three tracks. We took notes.

 

It was a little more difficult with J.R. Media Center, but we duplicated the tests the same way.

 

I also played then with hardware on the third system a bit, just for fun. The tests with changed hardware were not as extensive, often I would only listen to half the track, just to stave off ear fatigue.

 

Our results indicate that there is a audible difference, though it is at best, minuscule in our systems. We would not have heard it had we not been intensely listening for it.

 

For your pleasure, here you go:

 

System one:

WAV vs. AIFFBE - Difference

WAV vs AIFFLE - No diference

AIFFBE/AIFFLE - Difference

 

Comments: The differences were small, and mainly noticeable on the drums when introduced at just over 1 minute. The drums had a different decay with the AIFFBE file.

 

System ID: Macbook, Lion, Amarra Mini, Wavelength Proton, NADT747 (Analog ports, no processing), Nordost Flatline Gold MKII, Magenpan 1.7s.

 

System two:

WAV vs. AIFFBE - No difference

WAV vs. AIFFLE - No difference

AIFFBE/AIFFLE - No difference

 

Comments: We thought we heard a difference in soundstage, but this turned out to not be the case when we ran the second set of tests "blind".

 

System ID: PowerPC Mac Mini, OSX 10.5, 1gb ram, iTunes, Monster mini to RCA Audio cable, NAD T715 receiver, AntiCables, Cambridge Soundworks Model 6 speakers on bookshelves.

 

 

System three (normal):

WAV vs. AIFFBE - Difference

WAV vs. AIFFLE - No difference

AIFFBE/AIFFLE - Difference

 

Comments: As with the main system, we heard small differences. In this case, the differences were with imaging and soundstage. Instruments and performers moved around unexpectedly.

 

System ID (normal):

HP Quad Core PC, J. River Media Center, Music Hall Dac 25.3, NAD BEE326, Nordost Flatline MK II cables, PSB Imagine B Speakers on Target speaker stands.

 

----------------------------

With System three, I also tried several different hunks of equipment and software. In all cases, the amp, speaker cables, speakers and stands remained unchanged.

 

Replacing the Music Hall 25.3 DAC with an old Beresford 7510 MK II DAC minimized the differences to the point I am unsure they were still there.

 

Adding a V-Link USB->S/PDIF converter brought the differences back, even with the Beresford DAC. More so with the Music Hall.

 

No differences between optical and coax inputs.

 

Running iTunes on the PC instead of J.R. Media Center removed all differences with any hardware. Running JRMC brought back all differences. (JRMC configured with WASPI Event and no oversampling, no effects, etc.)

 

Putting the Wavelength Proton in place of the V-LInk/DAC, besides giving better sound, resulted in much more subtle differences, closer to what we heard in the main system. The soundstage stabilized and changes were again, mostly audible in the drums.

 

 

-Paul

 

 

[wgscott]

I could not detect any differences on 3 systems and a wide variety of players.

 

[nigel]

All three files sounded identical to me, both via Airport Express and via local Mac Mini and Musical Fidelity V-Link.

 

But the NAD M2 is very good at eliminating the effects of jitter as well as noise and distortion, whether EMI, RFI or power supply sourced, so identical input bits (which we know these three files contain, albeit in different formats) always sound the same on my system.

 

[Paul R]

If I had not written down what I was hearing, I would have probably said no differences at all myself. The differences here were subtle, but we were able to pick them out, reliably. At least according to our notes.

 

-Paul

 

 

[One and a half]

all files sounded the same to me.

 

For fun, I converted the WAV file to FLAC, still no difference.

Can someone explain the file size differences between the formats, is it due to header information?

 

The FLAC is 18306kB, pretty close to 50% the size of the WAV FWIW.

 

Excellent recording BTW, can all recordings made be like this? I can dream on.

 

[twist222]

I would still like to try this agin when I have a bit more time but based on todays listen I can not reliably hear any difference A few nights ago when I was using a different piece of music I had a quick comparison between wav and AIFF and decided wav did sound better. However on that night I did know which file was wav and which aiff during playback.

 

Tonight I put the 3 different files into a single playlist and set it to repeat. The monitor is lying flat on top of my music rack and I cannot see it from my listening position

 

After I used the IR remote to quickly click forward numerous times so I was not sure which of the 3 files I was starting with I was not able to clearly identify one of the tracks being clearly better or worse than the others.

 

It was not an ideal listening session as I have numerous other things happening today and I would like to do it again some time but on my tests today you can put me in with the group who can her no difference.

 

For today's session I was using a 2009 Mac Mini with a Paul Hynes linear power supply, SSD, 8Gb ram, Amarra, feeding a Meic Halo ULN8 via FireWire. Music is on a bus powered Oyen mini pro via USB to the mini.

 

Happy testing to everyone else

Mark

 

[wgscott]

"Can someone explain the file size differences"

 

FLAC and ALAC use lossless compression. The wav and aiff files we are dealing with do not compress the information at all.

 

[ted_b]

to report back.

 

First, I had one of my sons name the files randomly, then I put them in a playlist and played them in no particular order. I used PM 1.8a integer mode with my 2009 Mac Mini (8gb, SSD) and Antelope Gold/Voltikus combo. The songs were resident on my Synology NAS.

 

I was very familiar with this song, as I have Barry's 24/192 Equinox and love it, know it, etc. So, I was aware that any of these renderings would be sub-standard to my normal Equinox listening.

 

I detected basically no differences whatsoever between the WAV and AIF-C (LE) files, but the classic (for me) but subtle AIFF softening in the AIFF (BE) file. All the leading edges were slightly dulled in the AIFF file (again, not a big difference and not unlistenable by any stretch, assuming I hadn't heard the other files), along with softer dynamics and slightly less micro-details and deep soundstage ambient cues. The differences in LE and BE , to me, were identical to the differences in WAV and BE.

 

Does this mean I need to have my BE engine fixed (whatever that means)?? Dunno. Does this mean I can now convert to AIF-C for everything and not have to deal with metadata issues? Maybe. I will see if i can create AIF-C hirez comparisons for myself and understand if some of my fave 24/192 and 24/176 songs sound identical in AIF-C. Hmmmm...

 

Thanks again to Paul, Barry, and all involved.

 

[twist222]

Miska or anyone one else,

 

If you are still following this thread could you tell me what program you used to create the LE AIFF files ?

 

If it is easy to create LE AIFF files it might be a good option over the standard BE versions. Although for me on my listening test so far I could not reliably tell any difference.

 

Regards

Mark

 

[One and a half]

Sorry to lead astray there, I meant the difference in file sizes for the AIFF's and the WAV. For the same song, the differences are small but they aren't identical.

 

[Miska]

what program you used to create the LE AIFF files

 

I had to write a small command line utility (35 lines, for Linux) to do it, since there were no nice ready made and reliable tools available.

 

If it is easy to create LE AIFF files it might be a good option over the standard BE versions.

 

Making a suitably flexible command line utility for OS X or Linux wouldn't be hard and it could be used from a script to batch convert large amount of files if needed.

 

Current one is very simplified because I made it just for this particular purpose. Nice one would still probably be

[Miska]

Sorry to lead astray there, I meant the difference in file sizes for the AIFF's and the WAV. For the same song, the differences are small but they aren't identical.

 

There are differences in format header sizes and amount/type of metadata included. Audio data is same size (9261000 frames) in all files.