Interesting Article on MP3's

[zotjen]

I don't believe this was posted yet:

http://www.sfgate.com/cgi-bin/articl...DDEJR7KN11.DTL

 

[mwallace573]

I'm guessing the guy that wrote that article was basing all his observations on 128kbps iTunes files. I personally cannot tell the difference between a v0 LAME encoded MP3 and FLAC. I mainly use FLAC at home because drive space and battery life are not factors as they are with my portable setup.

 

[goldenratiophi]

"FLAC: This codec, favored by Grateful Dead tape traders, stands for Free Lossless Audio Code. It reduces storage space by 30 to 50 percent, but without compression."

O RLY?

I wonder how they do it then!

 

[oicdn]

^^ haha, I was thinking the same thing.

In anycase, I'm wondering what his observations were based on, and howcredible those people really are. Just because you're a producer doesn't mean you have a good set of ears, you just know how to string together music....there's a difference. Just look at some of the producers today...lol.

 

[Redo]

This sums it up to me right here:

But Sweetow also notes that music with lyrics may act entirely differently on a cerebral level than instrumental music. "The words trigger the emotion," he says. "But those words aren't necessarily affected by fidelity."

I've had this idea, and have held on to it for a long time. When people focus on the words/lyrics, the fidelity doesn't matter. But when you listen to the instruments, ambiance, and even the tone of the vocals, you introduce yourself into an entirely different world.

 

[YamiTenshi]

Sorry, I'm a bit late to the thread.

It could be that MP3s actually reach the receptors in our brains in entirely different ways than analog phonograph records. The difference could be as fundamental as which brain hemisphere the music engages.

"Poorer-fidelity music stimulates the brain in different ways," says Dr. Robert Sweetow, head of UCSF audiology department. "With different neurons, perhaps lesser neurons, stimulated, there are fewer cortical neurons connected back to the limbic system, where the emotions are stored."

So here's what I'm wondering, even if we can't perceive the difference on a conscious level (i.e. 320 v. wav), doesn't our brain still perceive a difference (i.e. emotional response)?

 

[majkel]

This article is unprofessional in many terms and shouldn't be linked anywhere as a reference. Some of the misleadings you have already pointed out.

1. We know nothing about the mp3 codec/settings used.

2. No compression in FLAC?! It IS compression but lossless, works similar to zip, rar, etc. The same lossless algorythms are involved in the mp3 encoding but they do just a small part of the job to shrink the bitrate.

3. Less than 1/10 of music information remains?! Guys, please give me the equation and data to calculate this. LOL!

4. It is true, there are many informations lost when encoded to mp3. Limiting to the LAME 3.97 Final, here is what I have observed:
- CBR files sound the way I cannot resolve the soundstage and the distance from the sound sources
- VBR files give overall better SQ/bitrate ratio and do have audible soundstage/source localization
- 320kb/s CBR Joint Stereo is the only bitrate giving the full color of the midrange. Everything below this bitrate is more or less "grayed out"
- no mp3 is transparent when listened from a decent rig. It's mostly the players cannot deliver the advantage of listening to 320kb/s mp3 or lossless files. I realised it when playing with many different mp3 players.

 

[Kilane]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif 3. Less than 1/10 of music information remains?! Guys, please give me the equation and data to calculate this. LOL!

Very interested to see how they did their math.

The mp3 debate is one in which my audiophilia and b.s. meter are constantly at war. To say that music impacts the brain in ways when it's subjected to a full range of the sound spectrum is one thing, to say that we may not be as moved by music because of what we can't hear is... quite another. I'm willing to admit compression is bad, but saying it's robbing your music as much as poor source mastering seems off to me.

Fight the battles in the loudness war first and the rest of the compression stuff later.

 

[Redo]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif 3. Less than 1/10 of music information remains?! Guys, please give me the equation and data to calculate this. LOL!

CD Audio is 1411kbps PCM. MP3's/AAC/whatever aims for 128kbps as their optimal bitrate (well not really, but most people accept 128kbps as near perfect). That's less than 10% of the original file size.


However, the CD Audio is uncompressed. MP3/AAC/Whatever uses both lossy and lossless compression, so it's more like 15-20% of the original on 128kbps. Turn it up to -V 2, -V 0, 256kbps, or 320kbps, and the result is more like 30-60% of the original (depending).

This is just in regards to filesize though. The actual sound quality is much different.

Let me throw together quickly what I think lossy encoding SQ represents (lame 3.97 mind you):

128kbps - 85% (can be annoying at times, OGG is better @ this bitrate)
V 4 - 92%
V 2 - 95%
V 0 - 97%
320cbr - 99%

That is on my current setup though. I'm sure with more upgrades those percentages may dive down a bit.