Mp3 this, Mp3 that. Is it really that bad?

[@?,*???&?]

I just had my ear training class wade through this exercise. Listening to the difference between 16-bit .wav file, 128Kbps, 160Kbps, 192Kbps, 320Kpbs and a 128Kpbs .mp3 converted to a .wav and then back to a 128 Kbps .mp3.

60% of them could not tell the difference between mp3s. 100% could hear the difference between 16-bit .wav and .mp3. 100% of them could easily pick out the mp3 of the mp3 file when compared to everything else.

Indeed, mp3s are much ado about nothing when compared to themselves it seems. Resolution? Who cares? How much storage space do you have?!?!

lol

[@?,*???&?]

Until they can get MP3s to stop sounding like music being played through a ceiling fan, I'll continue to assert that they sound terrible.

But Chris, the average listener can't tell...

[fossiltooth]

I wonder if they could be improved further by mastering to that format. I know nobody would want to, but considering the master is reduced to 16/44.1 and then that gets reduced to mp3, seems like a better mp3 might be possible by skipping the middle step.

I just dumped a bunch of mixes to MP3 straight off the 32-bit mixes and they sounded a whole world better... there's definately something to that "skipping the middle step"

MEAT

Are many mastering studios offering this service yet?

It seems like it could be a good source of additional revenue.

If they have the means, most people will gladly pay for something that should sound better, whether or not they can easily hear the difference themselves. Just look at the hobbyists on Gearslutz.

[rwc]

Ok. Here is my take on it based on a lot of ABX testing.

MP3s are fine.

If you encode them inproperly, with 128k, with a bad encoder like xing, or even worse iTunes, they will sound bad. Check hydrogenaudio.org for extensive blind testing that has been done to show which encoders go over better with general music.

However, if you encode properly - say, with LAME, and --alt-preset-standard, it'll be transparent. Not on all test clips, but on 99.999% of music.

I know because I've given the ABX test 1000 times already. ABX is just put the 16/44.1 wav as A, the MP3 --alt-preset standard as B, and X you have to guess which it is. There's 0 delay when switching, you hit the button on the keyboard and can switch back and forth as fast as you want. You have to get 12/16 right to have heard a difference.

NO ONE HAS EVER GOTTEN PAST SIX! and the ones that did got something like 3/16 or 2/16 on the next test.

I think there are better formats out there. I think the fact that Ogg Vorbis, Musepack, are behind inferior formats like AAC and MP3 sucks. I think the fact that people still use CBR in this day and age instead of VBR sucks. I think the fact that formats that are not natively gapless get used sucks. Above all, I think the fact that 24/96 and 24/192 recording hasn't caught on as a consumer format sucks. I think 24/96 vs MP3 would be an easy ABX test.

But, MP3 is not only tolerable to me, it is transparent. If you do a true blind test with properly encoded material, you'll see what I mean.

I think a lot of people have associated MP3 with youtube or myspace or 64k stuff or badly encoded crap on kazaa. But that's too general. That's like saying "CDs sound bad" after hearing the latest Ashlee Simpson album with an RMS of -4 or so. It doesn't mean CDs are bad in general.

I am confident that when compared to the 16/44.1 CD(not the 24/96 or 24/192 that, for 95% of all music, is never released), a well encoded MP3, in a true blind test, will fool everyone, and be 100% transparent.

[@?,*???&?]

It's all about the perceptual coding and how convincing that is. It's really up to the softeware engineer writing the program.

What we as engineers should focus on is which companies have the best sounding perceptual coding process.

No?

[fossiltooth]

I think a lot of people have associated MP3 with youtube or myspace or 64k stuff or badly encoded crap on kazaa. But that's too general.

I think that's an accurate observation.

There was a day when pretty much all mp3's generally sounded like garbage, but it's been years since this was true.

On the other hand, I haven't heard any good-sounding streaming content on the web yet. I assume paid streaming subscription services can't possibly sound 1/10th as bad as myspace or youtube, but I've never checked any of them out.

[palinilap]

Exact Audio Copy (EAC) with the LAME codec is my favorite encoder. It saves a text file log when it's complete, so you know if their were any errors. Used by the tweakiest bittorrent folks.

[dynomike]

192kbps in LAME always sounded fine to me. But I've switched to 256kbps because I don't trust myself and other people say they can hear the difference. Stupid!

[rwc]

Exact Audio Copy (EAC) with the LAME codec is my favorite encoder. It saves a text file log when it's complete, so you know if their were any errors. Used by the tweakiest bittorrent folks.

I've been using that since before bittorrent existed.

It is amazing. Put in CD that's barely playable, get a song that has one short skip. Amazing.

A lot of, even portable players, are beginning to support better formats now. Use the latest build of Vorbis at -q6(192k vbr). I have to say it's pretty transparent on everything. It was a huge improvement over MP3 when it came out 6 years ago, and has developed very well since.

[mvollrath]

lame.exe --vbr-new -V0 -b128 -h

The only way to encode mp3's. Hard disks keep getting cheaper, so why not?

[operator_tape]

yes they suck! Nothing like a real CD or Vinyl !!!!

[fossiltooth]

yes they suck! Nothing like a real CD or Vinyl !!!!

OK.

Please post a link to a modern high-quality mp3 file and the AIFF or WAV it came from and explain how it "sucks". That's a pretty strong word for what has become a fairly subtle difference.

If you send a link to a myspace page, you've clearly missed the point.

Once you pick out the difference in an ABX test, please explain how it sounds worse than the first listening format and system you were familiar with as an adolescent. That was the original point of this thread.

I don't know, maybe you listened to everything on 1/2" 2-track through mcintosh power amps as a teenager. Did you?

[KilledByAlbany]

word.
Man, I have a tape that I dubbed of the Germs in 8th grade. I have listen to that tape waaay more times than I can count, mostly on this shitty yellow tape player that only had one speaker!

Then the $100,000 question really is: When you listened to "Forming", did the one speaker play the side that is all the music, or the side that is only the vocal track?

[fossiltooth]

word.
Man, I have a tape that I dubbed of the Germs in 8th grade. I have listen to that tape waaay more times than I can count, mostly on this shitty yellow tape player that only had one speaker!

Then the $100,000 question really is: When you listened to "Forming", did the one speaker play the side that is all the music, or the side that is only the vocal track?

I'm guessing it was summed to mono.

[adam]

My only problem with mp3 - and digital in general - is the perception by so many that it is actually superior.

That's definitely something to be bothered by.