tl;dr first and foremost before I prattle on like a bumbleclown.
Let's ignore the hubbub going on in that other thread (just bad timing that got zombied right as I decided to get boned up on all this). I want to be able to INTELLIGENTLY prepare my tunes for proper, high quality, best fidelity mp3 exports using Sonar and all its tools (and learn about useful 3rd party stuff that might help but would prefer to remain in Sonar) AND learn about the fundamentals of the actual mp3 export process.
Based on what I know so far this is two completely separate steps. 1) Getting the audio output ready for the mp3 format and 2) the actual conversion.
This is one of my "totally open" threads where I'm just looking for key topics to research, descriptions of how you guys do it, general info on WTF the conversion is doing to the audio and any other bloody thing that comes along (without getting into some huge debate about the format itself or whatever).
So far what I "know for sure" is I can integrate the LAME encoder into Sonar or buy their paid mp3 thingie or use a third party program. I do "know" quite a bit more than that but it's all sketchy and as recently proven apparently a very contentious and subjective issue... which is likely why I'm still totally fuxxored on the subject (no one seems to agree on anything).
So all I need is to know what I gotta do in Sonar/to the wave so the conversion goes well (no clipping/distortion) and best practices/tools for doing the actual conversion. Yes this is likely more of a Techniques tab question but since I'm doing this from Sonar (at least the prep work) I'm posting it here.
Have at 'er or read on (blathering incoming, and no it has not starting yet if you can believe that... lulz)....
So here's the trip. mp3 is another one of those topics I just always figured would start making sense the more I studied. Most things do... but this hasn't. The more I read about it the more I get confused. I have a HUGE book I found AGES ago on the topic that looks like Tolstoy and Mitchum artificially inseminated a pulp mill. It's really old and honestly I think it's more about coding or something. Based on my frustrated scans of it it sure isn't helpful for just wanting to output good quality files.
What I have gleaned from my recent searches is that I need to reduce my Master level much more than I would for a wave export. The number that keeps coming up is -3db (I usually export my waves to -0.1db using a limiter). That number though apparently seems to be some kind of ultimate buffer/safeguard and not REALLY necessary in all cases. Like it's overkill just to make sure dumblef*cks like me don't distort their files.
I am also gleaning that the reason for this need for reduction in output volume is that the way mp3 reduces the file size is by removing "quieter" samples from the audio file... thus only leaving the peaks/louder samples of the file which raises the average volume of the entire file (or the RMS which is another subject that confuses the bloody heck out of me).
To add to that I also read that sometimes the peaks might increase as well due to two or more "louder" samples (like peaks) getting stacked or jammed too close together which "sums" them and creates a much louder "peak" than the original peaks which then can create momentary clipping/distortion. This concept seemed almost like speculation and depended on the converter/material AND that the distortion/clipping (since only occurring for a single sample or so) isn't really noticeable most times BUT maybe if consistent enough WILL be noticeable.
So that's a big point of confusion. How does one decipher how much to reduce to avoid that "summing/peak sample bunching"? Is -3db going to always do the trick and should just be done ALL the time or is it the type of thing that causes other stupidity so should be approached on a case by case basis (I'm assuming it should and the type of thing a mastering house takes into consideration).
Then there are different qualities. I'd like to understand a little better what the standards are. I will generally try to use the best quality conversion (which I'm assuming results in larger files due to less orignal data being yanked out). Like does more care need to be taken when prepping the files for mp3 for lower res files? Is there a fixed equation for drive space/per minute of audio like there is for wave files (I'm assuming no and that it is case by case depending on the audio)? What export qualities are standard for various purposes/common applications (for example I think sites like Soundcloud actually convert to mp3 so is there a "standard" to shoot for when uploading to places like that or does it vary)?
Then there are the "CODECS"? I think that's what they called the various programs/instructions that do the conversion/compression. I am getting the impression this is likely of HUGE importance so what's the deal with LAME? Is it able to do the best conversions? To specific programs do the job better? Does the Cakewalk encoder work better than LAME (or does it use LAME)? How about PYRO (which I own) or Audacity or Reaper, etc?
Well... you get the idea. It's just such a large subject for something I always thought was super simple.
All I want to do is export to mp3 in a way that will best represent the original files whether it's my own stuff or if I'm ripping my music collection or whatever. Apparently though to do this PROPERLY one needs to actually know a little bit more than how to set up the encoders and choose the mp3 export option.
As I said... just looking for facts and anecdotes so I can understand it all better and it could (hopefully) be a good thread to nail down some of this stuff for anyone else getting confused by all this like I am.
Cheers, thanks and I hope everyone is having an awesome day.
:-)