Myspace Mp3 format

Whenever I load a song onto myspace, it sound crappy, worse than others. My original mp3 doesnt sound that bad. Im using adobe audition and working with a wav file, and then just saving it as an mp3. Am I doing something wrong. Here is the song I am talking about

http://profile.myspace.com/index.cfm?fuseaction=user.viewprofile&friendID=201175145

Thants the link to the page where the song is. Any suggestions??

Try mine out. They sound as crappy as anyone elses...

http://myspace.com/ognisonlyremix

Definitely you want to use a variable bit rate. It doesn't make any sense to do otherwise for best quality for the compression level. I kind assumed that was the case. Does MySpace have size/rate limits on the uploads?

Box.net has something called a 'widget' that they built to work on myspace. That would be a way to play high quality mp3s (say 320kbs) or even wave files on myspace. Box.net allows you to store up to 1 gig for free.

I think they might convert everything, but then again, how do I know. I wonder if there is info on this anywhere? I haven't found any yet. You think that as popular as this site is, they would offer better audio!!

I thought they had a proprietary algorithm to convert into their format, does anybody know?

The files are MP3s, so at most it could only be some MP3 encoder that they've tweaked for their own purposes or something. But I'd imagine it's just more that they are using some standard MP3 encoder that they set for quite high compression for minimum file sizes. This all assumes that they do in fact re-encode everything, even if you send in a well compressed MP3 already.

Variable Bit Rate sounds better than Constant Bit Rate?

Variable bit rate doesn't sound better, it just allows you to get the same quality (other things being equal) into a smaller file. With constant bit rate, you'll spit out as much data for a section of silence as for a section of "wall o'sound". With variable bit rate, that section of silence will spit out very little data because that's required to represent something that simple. Same with something like a long, dry synth pad. That will compress very well and use many fewer bits that a finger picked accoustic guitar part with reverb, because the later involves much more complex and subtle sounds. But at constant bit rate, you'd get the same amount of data for a second of each.

But, if file size triggers a re-encoding, using variable would be a lot more likely to get you in under the line, and the higher the bit rate, the more variable will help you on file size.

And variable vs. constant shouldn't sound any different either way, even at low bit rates. The variable bit rate just doesn't waste space in the output file by encoding content that wouldn't be heard. A 256K/sec file at variable or constant rate should basically sound the same since the same level of agression is being used in compression, it's just that variable doesn't waste space because it only uses as many bits as required, while constant rate always spits out XK/sec, even if it doesn't actually need to. I don't believe that it lowers the level of agression in the compression as it goes, because it wouldn't know that it needed to or not until it had already done the compression, and it could lead to strange side effects if the level of aggression was constantly changing over the course of the song.

That's my understanding of it anyway.

A variable bitrate file of the same sound quality as a constant bitrate file will be larger in size.

If there are differences, they would be quite subtle, even when the variable rate file is much smaller. The variable bit rate file will use up to the maximum set bit rate if it needs to, so it only backs off where it's not required while the constant bit rate one will encode things like silence at full bit rate. So if you encode a very complex file at variable bit rate, it'll be close in size to the constant bit rate (for the same quality level), because the variable rate encoder cannot find much in the way of savings. Where there are savings to be had, it'll take them, but the variable rate doesn't compromise the sound quality at a given compression/quality level relative to the constant bit rate at the same compression/quality level. They both max out at the same level, one just is more efficient than the other by not encoding things that it doesn't need to.

I.e. it's the compression/quality level that controls the quality of the encoding, not the variable/constant rate. The generated quality is the same in both cases, since they are both limited by the same maximum bit rate.

I'm pretty sure they re-encode everything to either 128k or 112k, so if you send them an mp3 that 128, they will re-encode it to whatever they do. Even if you take a plain 128kbps mp3 and re-encode that same file to another 128kbps on your own computer it will sound worse than the original 128kbps. I upload my files as a 192kbs or 224kbps (depending on the size) and it doesn't sound too bad, a lot better than others I've heard, and variable bitrate doesn't make it better, just a smaller file size. And I think they now have a 10MB file size limit. And you can set up a variable bitrate mp3 to have a maximum of whatever you want, like 160, it doesn't always have to be the maximum of the highest, plus variable bitrate is pointless unless you have a song with lots of dynamics and softer parts with less going on.

Well the thing is that the compression quality level varies on a variable bitrate file. For instance it'll be from 100 - 125, or 105 - 140, 110 - 150 and so on. But even if you settle on the 100 - 125 bitrate average it'll still produce a file that's as much as 25% bigger than a constant 128k bitrate file and will sound about the same (at least in my opinion, although Fraunhofer considers the 100 - 125 range variable bitrate to be of low quality).

It would have to be a pretty pathological file for the variable rate one to come out larger than the constant. In any sort of normal content it will never be bigger, else the whole point of variable bit rate would be useless. I thought that most encoders worked on a maximum bit rate (the one you indicated) in either mode. In constant bit rate mode, it'd always use as many bits per second as the maximum allows for the quality level you've selected. In variable mode, it'd use up to the maximum bit rate for the quality level you've chosen, but wouldn't exceed that bit rate, it'd only do less when it can, and it always can in any real world audio content at reasonable bit rates.

But, maybe MP3 encoders don't use a maximum bit rate for a given encoding quality and would allow it to go over in more complex sections when on the lowest quality settings. And that would be a good thing I guess, because it's struggling so much for enough bits to represent the more complex parts of the content at the very low rates. But at more normally used bit rates, I don't think you'd be likely to see a variable rate file coming out larger than the constant rate, because there are enough bits in the budget to do a good job and it's more likely to use less than that maximum than to go over the maximum.

So maybe I assumed that they never exceed the user indicated maximum bit rate, because I'd never tried a very low quality compression at variable vs. constant bit rate, and therefore only ever saw it come out considerably smaller.

Flash does streaming, but up until 7 it most certainly did not like VBR mp3 - that used to be the cause of the chipmunk sound problems people had on myspace etc. Dunno how 8 fares, but I do read ads for codecs *adding* that function to 8, so that's not very hopeful.

As I was saying above, it's probably because of the use of the quite low bit rates. At that bit rate, it's going to need every bit it can get to get a decent sound, so if the content isn't not highly compressable, it may run over that bit rate a good bit of the time. But at the more commonly used bit rates, it shouldn't ever be bigger and in fact the variable should be much smaller for most content because there are enough opportunities to use fewer bits. And it may also be due to MP3s compression technology, which is pretty old school by now. Something like WMA, which is what I use and therefore I'm basing my obvservations mostly on that, is a much more modern codec and I always find that the variable bit rate file for the kinds of quality levels I use (80% to 95%) get a big side reduction advantage being variable rate.