AIFF Vs. WAV

[Paul R]

Hey Barry -

 

Yes, which as I am sure a half dozen folks have proclaimed by now, is AIFF-C/SOWT.

 

However, non-Apple applications, and even some Apple sound editing apps still create standard AIFFs, which are big endian.

 

Moreover, I think (not positive) that the header (header fields in the chunks) still must be in network byte order, which is big endian.

 

I wrote a couple little sample programs to test out if the Apple Core Audio code still works with big endian AIFF files, and amazingly enough, it does- transparently. Not only that, an AIFF file in big endian format sounds different to me than one in little endian format, when played through Core Audio or Quicktime.

 

None of which is conclusive, but it would be interesting to find out if this is a general thing that most people can and do hear, or if it is just something localized to the systems here.

 

Even more interesting, if what I think might be happening is really happening, it should only happen on Core 2 Duo, and possibly Solo machines. It should not happen on i3/i5 or Pentium machines. (Might happen on Celeron or Atoms, don't know.

 

-Paul

 

 

[goldsdad]

"an AIFF file in big endian format sounds different to me than one in little endian format, when played through Core Audio or Quicktime"

 

That is interesting, Paul. With all due respect, expectation bias is one possible valid explanation. I realise that there are other reasonable explanations.

 

 

 

[goldsdad]

Perhaps you can just ask Damien?"

 

For sure, I wish I had used the words "identical sequence of bytes" rather than "identical sequence of words" when I previously asked him about the buffer.

 

Anyway, yes, asking Damien would help... to some extent.

 

If he says that I am wrong in my assumption of Audirvana's track buffer contents having invariant endianness with all source file formats then that will be an end to the matter as far as I'm concerned. There would be a rational explanation other than expectation bias or placebo for Audirvana sounding different to some people when comparing AIFF to WAV.

 

For the sake of argument, let's say he says that my belief is correct; i.e. the program is designed to fill the track buffer with one particular sequence of bytes for a given set of audio samples, regardless of the source file format. I have no doubt that lack of empirical proof of the buffer having the expected contents would be raised as an objection to accepting that playback with Audirvana results in objectively identical sound when playing AIFF and WAV files which contain identical audio samples.

 

 

 

[Teresa]

"We cannot measure all this because these are not objects of real world, but exist only in our minds."

 

Not true check out the Chesky Records Jazz Sampler & Audiophile Test CD (JD37) LEDR tests with "up and over" in which the sound actually raises up in an ark starting with the left speaker, slowing raising until it reaches about four feet above my speakers at center position and starts to descend as it goes over to the right speaker until it is as low as my speaker. This no illusion but very, very real which you can PROVE in your very own system!

 

Thus until we are able to measure the MOST important things with sound reproduction we will never understand it completely. Dynamic range only tells us how low and soft the music can be without excessive distortion and frequency response only tells us how accurate we can reproduce the frequencies in the original sound wave and how extended that reproduction is.

 

We need to learn how to measure soundstage, image width, image height, air between instruments, ambiance, smoothness or roughness of string tone or timbre accuracy, just to name a few unmeasurable parameters of sound.

 

GB said "The last measurable parameters in playing sound chain are air pressure at the listener's place and vibration of his chair."

We also need to learn how to measure those parameters as well, I'm not sure they will be the last but they are important as well.

 

[Miska]

Apple uses this new little-endian AIFF type as its standard on Mac OS X. When a file is imported to or exported from iTunes in "AIFF" format, it is actually AIFF-C/sowt that is being used.

 

No, it's not. It's the old style big-endian AIFF format. I found this out when I was adding AIFF support to my player software...

 

Here's hexdump of a header of an AIFF CD-rip made with iTunes (using what ever was latest version at May 26th 2011):

00000000 46 4f 52 4d 04 9f b7 0c 41 49 46 46 43 4f 4d 4d |FORM....AIFFCOMM|

00000010 00 00 00 12 00 02 01 27 eb 88 00 10 40 0e ac 44 |.......'[email protected]|

00000020 00 00 00 00 00 00 53 53 4e 44 04 9f ae 28 00 00 |......SSND...(..|

00000030 00 00 00 00 00 00 d0 07 da 46 cf fc da 0f d0 03 |.........F......|

 

[thesurfingalien]

Teresa,

 

May I comment on something?

 

Not true check out the Chesky Records Jazz Sampler & Audiophile Test CD (JD37) LEDR tests with "up and over" in which the sound actually raises up in an ark starting with the left speaker, slowing raising until it reaches about four feet above my speakers at center position and starts to descend as it goes over to the right speaker until it is as low as my speaker. This no illusion but very, very real which you can PROVE in your very own system!

 

This is absolutely true, as it is based on phase-shifting between tweeter and woofer frequencies (or any loudspeaker in between depending on what #-way speaker one has). However, with a MTM or WMTMW based loudspeaker this effect can sound rather strange :-) Another factor is the actual phase-behavior of the individual loudspeaker-units used that can enhance the effect. Listening-position and room-acoustics are other factors.

 

However, this is not so much a measurable phenomenon, and as said, is highly influenced by loudspeaker configuration.

 

 

For the rest, I must concur with GB's statement in this, and like I said before, I believe that most can already be measured by slew-rate...

 

But I have to ask you... Can you describe to me what you mean by air between instruments? Is that how well one can position individual instruments / vocals in the sound 'image'?

 

Regards,

Peter

 

[Paul R]

Expectation bias.

 

It is possible of course, Owen. But I think it unlikely. If anything I was expecting to NOT hear a difference. I spent a lot kore time convincing myself I actually did here a difference than was reasonable, including getting everyone within reach in tbe neighborhood to confirm the results.

 

Until some folks here volunteer to duplicate the results, the results are certainly tenative.

 

[wgscott]

How do we duplicate what you did? I'm happy to give it a try?

 

[Teresa]

"Can you describe to me what you mean by air between instruments? Is that how well one can position individual instruments / vocals in the sound 'image'?"

Yes the position of the instruments in the soundstage and the spaces between them. Analog recordings are more revealing of the ambiance (air) in the spaces between the instruments than digital, although high resolution digital can often come very close.

 

In addition some speakers are more revealing of the empty spaces in the soundstage than others as is tube amplification over transistor. If it was someday discovered how analog, tubes and certain speakers fill in the empty spaces with ambiance "air" we will have more realistic digital recordings, digital equipment and transistor equipment. But before we do that we have to learn how to measure the phenomenon first.

 

[sandyk]

"In addition some speakers are more revealing of the empty spaces in the soundstage than others as is tube amplification over transistor. If it was someday discovered how analog, tubes and certain speakers fill in the empty spaces with ambiance "air" we will have more realistic digital recordings, digital equipment and transistor equipment."

Teresa

That is a misleading statement about "Hollow State."

A very high quality Solid State amplifier (especially Class A) can outperform the best of the affordable Valve amplifiers in every commonly measured parameter, as well as give a greater sense of height, width and depth, along with more air between instruments and performers along with pinpoint localisation.

However, this thread is not about "Hollow State" vs. Solid State, so let's not go further off topic .

Kind Regards

Alex

 

[goldsdad]

Yeah, more people with various hardware testing your software could generate interesting results. Anyway, disregarding bias, an objective difference in sound output when playing little-endian vs. big-endian data does have a rational explanation in the execution of different instructions. That gets us no closer to explaining why some people hear file format differences with Audirvana unless I'm mistaken about its track buffer always having the same endianness.

 

 

 

 

 

[goldsdad]

Teresa, if you listen to a pair of speakers and a non-existent voice/instrument can be heard in a location between (or above or beyond, etc.) the speakers, then the apparent presence and location of the voice/instrument is either an illusion or a delusion.

 

illusion |ɪË?l(j)uːÊ'(É?)n|

noun

a deceptive appearance or impression

 

delusion |dɪË?l(j)uːÊ'(É?)n|

noun

an idiosyncratic belief or impression that is firmly maintained despite being contradicted by what is generally accepted as reality or rational argument, typically a symptom of mental disorder

 

[thesurfingalien]

goldsdad,

 

I am not sure what triggered your post.

 

... But in general one can say that any form of sound-stage can be regarded as being an illusion simply because the performers of the music are not present at all

 

Regards,

Peter

 

[goldsdad]

"I am not sure what triggered your post."

 

Yes, I completely agree with you; all stereo image is an illusion. I was simplifying it to the case of one voice or instrument.

 

My post was in reply to Teresa's post a short time ago. She said:

 

"check out the Chesky Records Jazz Sampler & Audiophile Test CD (JD37) LEDR tests with "up and over" in which the sound actually raises up in an ark starting with the left speaker, slowing raising until it reaches about four feet above my speakers at center position and starts to descend as it goes over to the right speaker until it is as low as my speaker. This no illusion but very, very real which you can PROVE in your very own system!"

 

[The bold is my addition.]

 

 

 

[Gregor Samsa]

My favorite equipment manufacturer name has always been Audible Illusions.

 

In my profession, delusion is a term of art and it has no place in an audio forum.

 

[Paul R]

If you feel like it I would be very much interested to read you thoughts about this in a dedicated thread. I certainly don't know enough to speak intelligently about it, and would love to soak up some knowledge on the subject. -Paul

 

 

[thesurfingalien]

I see your point :-)

 

However, I do not believe that Teresa was trying to communicate this in that respect...

 

But perhaps we should open another topic for this?

 

Peter

 

[elcorso]

"Art has a double face, of expression and illusion, just like science has a double face: the reality of error and the phantom of truth".

 

Publilius Syrus

 

Biography

 

Author Profession: Writer

Nationality: Roman

Born: 85 BC

Died: 43 BC

 

Regards,

 

Roch (the philosopher)

 

[goldsdad]

Whatever was meant, yes, it was a distraction from the topic of this thread and should be continued elsewhere, if at all.

 

 

 

 

 

[goldsdad]

delusion

 

I thought it was a term of psychiatry.

 

 

[The Computer Audiophile]

This is a great read or listen if anyone is interested.

 

 

 

This Is Your Brain on Music: The Science of a Human Obsession

ABRIDGED

by Daniel J. Levitin

Narrated by Edward Herrman

 

 

http://www.audible.com/pd/ref=sr_1_1?asin=B002V0189Y&qid=1313371257&sr=1-1

 

 

Publisher's Summary

 

In this groundbreaking union of art and science, rocker-turned-neuroscientist Daniel J. Levitin explores the connection between music - its performance, its composition, how we listen to it, why we enjoy it - and the human brain. Drawing on the latest research and on musical examples ranging from Mozart to Duke Ellington to Van Halen, Levitin reveals:

How composers produce some of the most pleasurable effects of listening to music by exploiting the way our brains make sense of the world

Why we are so emotionally attached to the music we listened to as teenagers, whether it was Fleetwood Mac, U2, or Dr. Dre

That practice, rather than talent, is the driving force behind musical expertise

How those insidious little jingles (called earworms) get stuck in our heads

And, taking on prominent thinkers who argue that music is nothing more than an evolutionary accident, Levitin argues that music is fundamental to our species, perhaps even more so than language. This Is Your Brain on Music is an unprecedented, eye-opening investigation into an obsession at the heart of human nature.

©2007 Daniel J. Levitin; (P)2007 Penguin Audio, a member of Penguin Group (USA), Inc. and Books on Tape. All rights reserved

 

 

 

 

 

[Paul R]

Had to take a few minutes to think about this, as I had a friend at UT create the files for me. Actually, I walked over to his house with a couple cold beers, and this is his suggestion.

 

(1) Start with an ALAC file that was ripped from a CD at full resolution. We used

"I Want A Little Sugar In My Bowl" as sung by Nina Simone on "Bittersweet: The Very Best of Nina Simone". (I would provide these files or at least the ALAC file but copyright issues are involved.)

 

(2) Convert it to a WAV file with iTunes. We named the ones here with "(WAV)"

 

(3) Convert it to a AIFF file with iTunes. "(AIFF)"

 

(4) Use dbPowerAmp to convert the ALAC file to Big Endian PCM data.

 

(5) Use dbPowerAmp to convert the PCM file to a BIG Endian AIFF File

 

(6) Use dbPowerAmp to convert the PCM file to a LITTLE Endian AIFF file

 

(7) Use "hexdump -Vc [filename] > [filename].hexdump " on each file

 

(8) "diff -y" each file against the PCM data. Diff AIFF files against each other.

 

The next step requires a bit of assistance. Have someone put all the files into a playlist in some order other than the order they show up in the import. Have them start the playlist.

 

Take darn good notes.

 

Compare notes with playlist order. Compare individual files again as often as you wish. Note that if you try to compare them too quickly, any differences in sound will blend together and you won't be able to hear them.

 

Do this when you have the time to listen to each track, even perhaps taking a break between tracks.

 

When you A/B compare them, don't switch tracks back and forth quickly too often, to avoid ear fatigue.

 

-Paul

 

 

 

 

[elcorso]

'Delusion' is very different to 'illusion' in English

 

Is from the Spanish Psychiatry (some kind of illusion, but more an hallucination) but from an Schizophrenic woman, men never got this since we are only nuts.

 

Also, I must clarify, than we WAV lovers (no matter the sex) are nuts, but never Schizophrenic.

 

Regards,

 

Roch

 

PS/ Thanks Miska for clarify AIF-C on Apple/iTunes format, I remember your post from another thread, but I can't found it.

Also, if you use iTunes for ripping or XLD you finish with an old big-endian AIFF format, on 'Max' you can change the settings in order not to follow iTunes compatibility mode. On WAV is another history.

 

[soooowhat]

 

"Also, if you use iTunes for ripping or XLD you finish with an old big-endian AIFF format"

 

I've only ever used XLD or iTunes for ripping my files.

 

Can't help but wonder if the folks who're able to tell the difference have used dbPoweramp for ripping and inadvertently created the conditions for hearing a difference.

 

another variable to consider.

 

curiouser and curiouser

 

 

PS. this is consistent with what Miska reported earlier - despite what Wikipedia says about little-endian, big-endian seems to be the current Apple standard for endianness.

 

[Paul R]

To add one more factor to this, there is a standard for the byte order of a USB transmission, which I think network order (big endian). Can someone verify or correct that please?

 

-Paul