Printable Version of Topic

Posted by: ItsMe Dec 14 2009, 09:44 AM

Hi there,

recently uploaded my latest EP of acoustic songs onto Myspace ( http://www.myspace.com/aloneinboston ). My audio masters have good audio quality but my mp3 files (which sound good on computer or on different stereo systems) sound just crappy when uploaded to Myspace. What is the reason for this and how can you change files so that they don't sound extremly compressed ?

Thanks

Posted by: Emir Hot Dec 14 2009, 09:49 AM

I think they use their own compression on top. That's why. I've never heard good quality track on Myspace.

Posted by: ItsMe Dec 14 2009, 10:02 AM

Yes I heard that they compress everything you upload but still some things sound better than others.

Posted by: Staffy Dec 14 2009, 10:19 AM

Try to lower the output on the MP3's with some 3 or 6 db so they aren't "maxed out" before You upload them. Maybe its gonna make a difference since as Emir says they probably compression it to get max headroom whatever the volume of the original file.... and then You loose the dynamics.... This is of course most obvious on acostic music, if its metal or music with heavy drums You want notice it so well, anyway by converting a good sounding file to MP3 already spoils the sound imo.....

//Staffay

Posted by: ItsMe Dec 14 2009, 10:21 AM

Hey thanks for the tip. I already lowered it -3db. Which didn't change anything. I will try more though.

Posted by: Staffy Dec 14 2009, 10:29 AM

No probs... Another version would be to make a MP3 from the final mix without any mastering and then upload it, since if the mastering uses some compression its gonna be like a compressed compression... I think You got the idea....


//Staffay

Posted by: Emir Hot Dec 14 2009, 10:55 AM

Be careful when exporting mp3. Some programs use "normalization" so you end up with the same thing. Make sure that you turn off all processings except raw mp3 export.

Posted by: tonymiro Dec 14 2009, 11:40 AM

As Emir said their codec (and Youtube is as bad) adds compression so Staffy is right any compression or limiting that you do on the main stereo bus will end up being compressed again. The best solution here is to drop your output gain level so that your crest is down to about -14 or lower. Also note that if you are using the LAME codec that LAME often has a gain increase of up to 6dB and in that situation you may well have to aim for a crest of -20.

Also as Emir says normalisation is bad here (it is generally bad anyway).

Posted by: ItsMe Dec 14 2009, 11:47 AM

Hmm. I usually use the mp3 converter in Sonar7 producer edition. I don't normalize and usually use a Power3 dithering algorythm. These mp3 files sound good by themself but sound crappy when uploaded. When I export files as wave and use external converters the quality gets a little better after upload even though the mp3 file by itself sound definetly more crappy than the ones made with the Sonar converter.
I don't max them out as i said before (about-3db) so that shouldn't be a problem.





Hey thanks,

great advice.


edited for language - tony

Posted by: tonymiro Dec 14 2009, 12:03 PM

MP3 is a lossy format and is compressed whereas pcm wav isn't. So just rendering to an mp3 means you've compressed the original file.

Posted by: Staffy Dec 14 2009, 12:11 PM

This is an interesting topic. I had very much the same problems with both audio and video at Youtube before I finally found an optimal setting in Adobe Premiere...... I dont really spot a bigger difference on my files on MySpace, but mine isn't of course acoustic and there's a lot of noise everywhere as well.... It would been nice if You uploaded or linked the files so we can hear the difference, maybe something else is going wrong????


//Staffay

Posted by: tonymiro Dec 14 2009, 12:30 PM

Depends how far in to this you want to go guys. There is an issue here with format - using mp3; another with the dithering used (PWR3 etc) and how many times dithering occurs, how it is handling the word length reduction and whether it is adding high frequency artifacts; another with the SRC and how much and where.

To some extent all of that starts to go (a long way) beyond itsme's thread. If he wants we can discuss here but it may be better just to put up a new thread to keep this one OT

Posted by: ItsMe Dec 14 2009, 01:25 PM

Yes, I can definetly tell if I use dithering or not. Without, the files sound less good but after uploading with additional compression you get some hiss. I know the theory behind dithering and compression but the funny thing is that you still have to do some trial an error kind of thing because every codec behaves differntly.
Anyway we could dicuss that here but it would be great if we could start some more in depth threads to adress these issues. Like this we would attract more people since the title of the threat doesn't refer to the theoretical background

Posted by: tonymiro Dec 14 2009, 01:42 PM

Yes it can require some listening and testing as noise shaping is a bit subjective. Some like , or just get used to, shaped dither like PWR3, others prefer unshaped like TPDF (triangular probability density function), some prefer not to dither at all but just truncate. I tend to use TPDF - but that's a personal preference and it depends on the type of music.



I'm happy if you do.