Re: New Thread: Preferred Formats

Brent:

Nice call on the ABX. There are (in my opinion) not enough people that know about this methodology / device; this should get more 'play' in the audio community as I tyhink it's a pretty faiur way to assess to 'things' (sounds, cables, amplifiers, speakers etc). Your point is well-taken; in my line of work we do a lot of listening tests to differentiate the sound of products (although when we do so it is almost always in a paired-comparison method (sound A vs. Sound B) and all are wav), and when we do this it is done blind and random (i.e. using the Bradley-Terry model for instance). While I am used to doing this for differntiating product sounds (again, all sounds presented as wav with same sample rate and precision) I can see expanding this to test various formats (i.e. using parent-daughter files) to determine the 'just noticeable difference' between sounds and to put some statistical 
 'meat' behind the analysis.

There is a lot of controversy (better to say animosity) in the mega high dollar audiophile community regarding ABX - some people grow very angry / threatened when after taking an ABX-based test (to compare esoteric to budget cables for example) they find that their score is no better than chance. Not to start another war here...but I once suggested this methodology in an esoteric cable discussion forum and wow...did the flame war ever begin... So...you are right about people hearing (in some cases) what they want to hear - whether this is a comparison between file formats, cables, etc, but this is why controlled tests are so important. Still, in other cases, the differences between sounds are unmistakable - a comparison of a native WAV to a heavily compressed MP3 for example, but I doubt the difference would be as audible if the original WAV itself was of poor quality (i.e. think of a WA
 V file of a 78 rpm record for example that was then heavily compressed to MP3 - I doubt the
 audible differences would be that great).

Again, for me, my guideline has been to keep the WAVs (partly because I have paid for the music and have the disk space and thus, don't want to compress it unnecessarily) also for the reasons I cited, and the MP3s (at 192) as they are 'good enough' for most casual listening (in the car and on portable devices for certain). However, I think it is definitely reasonable to say that the difference has a lot to do with the type of music, its temporal and spectral characteristics and so on, as well as the amount of compression chosen.

So...192 seems to be a good garden variety rate that 'works' for me for all my 'non-critical' needs. Most of the time I can hear a difference between the files at 192 and the wav files (this I have actually verified using a 'blind' listening test) but my ability to tell the difference grows less as the amount of compression is reduced. This makes sense to me because with lower compresison rates more of the original file structure is preserved (albeit at the expense of disk space).

Interesting discussion thread...

Brent Aliverti <[email protected]> wrote:
I think the fact that you decided to "do it right and pull WAVs off the
source CDs" underscores a pre-existing internal belief that the WAVs would
sound better than the 320K MP3 files. This belief was validated when you
played back the CD. But it's still not clear whether it was psychological or
real. Only a blind test could tell you that with certainty.

As an aside, has anyone on this list actually done a blind or ABX test to
support their assertion that .WAV sounds better than (LAME 3.90.3 alt preset
extreme) MP3? Just curious...

Cheers!
Brent

----- Original Message -----
From: "Abraham, Howard E."
To:
Sent: Monday, January 03, 2005 9:18 AM
Subject: Re: [AT] New Thread: Preferred Formats


>I was making a mix CD. Normally, I only make mix CDs for the car and
>converting 320K back to WAV doesn't bother me. This time I was making a
>mix CD for someone else so I decided to do it right and pull WAVs off the
>source CDs. After I was done with the disc I listened to it to make sure
>there weren't any errors. It was then that I realized how good the music
>sounded compared to the MP3s I was used to hearing. This caught me
>off-guard. I didn't intend to make a comparison between WAV and 320K at
>all. I just assumed that 320K was indistinguishable from the source, so I
>never compared them before. Due to my assumptions, I don't think my brain
>was trying to convince me of anything.
>
> I used DBPowerAmp to rip my songs at 320K, CBR, stereo. I don't know
> which method that software employs.
>
> -----Original Message-----
> From: Turtle Beach Audiotron User Forum
> [mailto:[email protected]] On Behalf Of Brent Aliverti
> Sent: Monday, January 03, 2005 10:17 AM
> To: [email protected]
> Subject: Re: [AT] New Thread: Preferred Formats
>
>> Previously, I ripped everything to 320K MP3. However, after listening
>> to that for over a year I pulled out the original CD and was amazed at
>> the difference between the MP3 and the CD. I'm slowly converting to
>> WAV now and all new music gets ripped to WAV. I still keep the MP3
>> around for portable players.
>
> Hi Howard,
>
> When you compared 320K MP3 to WAV and were amazed by the difference, what
> was your methodology? And which MP3 encoder had you used initially? The
> reason I ask is, if you were comparing the two, and you knew which was the
> WAV and which was the MP3, it's entirely possible that your brain was
> subconsciously influencing your decision more than you realize and more
> than just what your ears told you. The mind is tricky that way. If you
> (like most
> people) are predisposed to think that non-compressed music is always
> better than compressed music, it's possible your mind tipped the scale
> that way to validate your mental model. I'm not saying that you didn't
> hear a real difference, but I would be interested to know if you see the
> same obvious difference with a true blind test (using ABX software or with
> a friend queuing up tracks and not telling you which is which until the
> test is over).
>
> Cheers!
> Brent
>
> --
> No virus found in this incoming message.
> Checked by AVG Anti-Virus.
> Version: 7.0.296 / Virus Database: 265.6.7 - Release Date: 12/30/2004
>
>
> --
> No virus found in this outgoing message.
> Checked by AVG Anti-Virus.
> Version: 7.0.296 / Virus Database: 265.6.7 - Release Date: 12/30/2004
>