Please find out which WAV file is which - and which WAV file does sound best:
 
CLICK HERE   

MP3      : A
Original : B
Amiga    : C

but I have a doubt for the last two because the listening was done via a laptop.

However, we have another reason why we think your audio samples give away the answer - not because of audio quality. It's because only Donau_C.wav will play with Quicktime Player 10 [QT ver. 10.0 (118)] for MacOS Snow Leopard 10.5.6 whereas samples A & B will not play under the QuickTime 10 player. In fact, loading samples A & B shows zero play time available for samples A & B, although both files are 17MB. Sample C shows 00:01:36 minutes of play time.

However, all three samples will play under QuickTime Player 7. Why that is I cannot say. All three samples will also play under RealPlayer version 12.0.0 (1444).

The movie inspector for QuickTime Player 10 shows all three as being "Linear PCM, 16 bit little-endian signed integer, 2 channels, 44100 Hz"

-Ernest

Excellent guessing!
Judging from the player incompatibility was very clever!

I've now listened carefully to the samples using audiophile headphones. My conclusion is as follows:

Sample C sounds a bit clearer and more detailed than the other two. I assume this is the original.

Sample A seems to have less details than sample B. It also seems it has a poorer stereo perspective than the other two (but not much).

Sample B sounds better than sample A, but has lost some of its dynamics compared to sample C.

I would rate them:

1. C. I believe this is the original because it simply sounds most musical.

2. B. I assume this is the MP3 version since most of the higher frequencies are still present.

3. A. This is probably the Amiga sound. It doesn't sound much worse than sample B, but since it seems to be less detailed I believe this is the converted 8bit sample (unless you used a very low bitrate for the MP3).

Audio jitter is smaller on sample A, and i have seen some Alias errors this one has passed through several filters.(less tremble, sounds quirky when played next to another)
And also Sample A is the longest of the three time wise.

Sample B actually leads when placing C next to it it causes a echo effect, on some parts of the music.
This is the shortest sample of the three time wise.
There are still some alias errors but this is on the low end of the volume where it should be on a good recording.

Sample C the only one of the three that would actually sync on frequency, Date of creation is the oldest of the three.(7:19:56)

My conclusion is that C is the original.
however which is the Mp3 wave converted is a mystery since it is unclear what MP3 rate it was converted to.
Less the 64kb/s i would say it is sample A if it is higher then 128kb/s it's sample B.

Anyhow the second one created was B 7:46:00
The third one was A 8:01:22

They all date to 5-1-11
they have no further ID tag AFAICS but then i didn't really checked for it.(call me sloppy)

if anyone has questions about the methods i used i am in the chat PM me there.

I think another test is necessary, as well as, maybe a modern song too.... from the 80s or 90s...

Second all forms of ID-tags should be removed from those files and Any participant shouldn't be able to see another participant's Review.

If those things are met then can we call this a blind test.

Totally agree, we Amigans really are sleuths, aren't we?....

However, this test still doesn't disprove Gunnar's contention that it could serve us very well in the scope of the  purpose it is intended for. Conversion, and back, does degrade the music/audio that is listened too. Do it with any lossy image compression format, the result will be the same, you will have a lower quality picture image later. Look at HAM images, they don't work across the board, but the results CAN be astonishing, especially in 1985!!!!!

Crackling/hiss may only be because of multiple (de)compressions, and probably could be ironed out by just as Basil said, "it's completely different, it's a semitone higher!"

Use a different volume or note as a hack and patch dirty fix....

So, it's not for "conventional full music output"; in that case use a real program for singing and acoustics then!

For what we get, if the occasional blur can be isolated and rectified to an alternate satisfaction, it's pure gold!!!!

NatAmi........ really Think Different!!

Yes, (C) is the original.

(A) is the MP3 file.
(A) was created by converting the original PCM16
into 128KBit MP3 and straight converting back.

(B) is the AMIGA format file.

That so many people find that our AMIGA audio floating format file "AAFF" does sound better than MP3 makes me happy - as it shows that we have in the AMIGA a good file format for audio.

Thierry Atheist wrote:

I think another test is necessary, as well as, maybe a modern song too.... from the 80s or 90s...

This is a very good idea!

I've created a small collection of popmusic converted to A8 format.
The files are stored in stereo. The archive includes the A8 player for Windows. Here is the archive CLICK HERE (130MB)

My problem is that I only have MP3 for the pop music to start with.
This means all the files were not created from a PCM16 CD rip but from a MP3. Therefore all AMIGA files include the MP3 errors to start with. Nevertheless the pop music files all sound quite OK IMHO.

If someone can provide me with a compilation of pop, rock or electro sample sound for testing in proper PCM16 format  (not MP3) then we can make a real comparison with them also.

Cheers

I didn't know the following till today, interesting read:
EXTERNAL LINK 
Soooo, IFF is big endian just like our CPU!!!

What's interesting is this passage, to me!

"An IFF file is built up from chunks. Each chunk begins with what the spec calls a "Type ID" (what the Macintosh called an OSType and Windows developers might call a FourCC)."

You see it clearly there in black and white, microsoft (M$) administers "FourCC" or "Force" or they "Fourk" you, while we AAFFF or lAAFFF at them and their yokes!!! HA HA HA HA!!!! :-DDD

The comparisons are to do with classical music because the recording of an instrument is what is closest to its original, and the sound is not colored by commercial considerations.
 
But the best comparisons should be done with the documentary records (eg wildlife).
 
For pop, rock, etc., and based on my experience, the "Sound Studio" records are preferable because closer to the reality of the studio (eg, Chaka Khan "I Feel for You", Paul McCartney "Spies Like Us", Turbulent Blue "Fou, Fou, C'est Fou !"). It can be noticed that the "Sound Studio" is not a musical style.

I hope that this can be of any help.
 
Back on the initial test and compare the spectral analysis of the MP3 with the other two, it distinguished himself as much as listening.

Who would have thought this?

Its problem is that it introduces a new degradation (often discernable by ear!) to a source that is already degraded. And this degradation is different from a simple reduction of the precision of the digital signal.

Regarding the sound quality of the Amiga, it is relative to the quality of the sample.

If it has very low frequencies and very high frequency, it would be better to spread them in two samples.