joden
 
oky doky gotch Beep  Right well I do use MP3's in my live gigs as backing tracks...so my method is to use the Sonar for anything with fixed rate (128,256,320) and I use Audacity (free)  if I want to create VBR as the Sonartool for some reason gets hung up on those!
 
I create the project, mix it, and then master as well as I can, ensuring I get most of the level meters for all songs at approx the same level, as in my experience, the MP3 conversion can sometimes create louder songs than the one previous even if all things are (relatively) equal.
 
I convert at 320kbps (used to do 256, but one CAN tell the difference in this rate with some songs) and in blind tests with muso m8's and others, I have only ever had one person pick the difference - and even then I reckon they were guessing!
 
So if you get your tune mastered pretty well in wav format, then using a higher rate MP3 should result in a pretty reasonable result. Especially as the end user is listening on less than stellar gear!
 
Dunno if that helps???

Definitely does. Just hearing what others do and gathering up terms and concepts to research further is totally what I'm after.
 
I always poke around the intertoobeys for info on stuff like this but really all my buds here seem to have a way of making things easier to digest (compared to other forums, dry tech manuals or oversimplified bloggy type crud).
 
You tossed some stuff on the pile I would not have been able to look up before without prior knowledge.
 
Thanks.

Drewfx1 made a comment in a thread a while ago that got my attention, which was the algorithm in mp3s cannot be correlated apples-to-apples with PCM. Because of this I became very leery of LAME, since it seems to "dumb down" the audio to 16-bit prior to the mp3 conversion. Adobe Audition actually ignores any "bit-rate" settings you would use for an mp3 conversion since they do not apply, and I have since defaulted to using Audition for all mp3 conversions. The whole crux of that thread started when I found a comment saying that mp3s are actually closer to 24-bit (hence the question of why convert to 16-bit first then?).
 
I honestly have not tested this greatly, but did do some cursory tests last weekend because of the other thread. Null-testing inside Audition on a 32-bit wav gave a residual that sloped from -inf -> -60dB between 5K and 20K. I didn't check sample alignment in Audition because the residual was so low, but it may even be lower (not sure). I really should have checked that though.
 
Attempting the same test inside SONAR (same two files, mp3 converted with Audition), left a residual that made the track easily discernible. In SONAR I had to mate the tracks (they were obviously offset from each other), but the track was still highly recognizable with one phase-inverted. I have not gone back to figure out "why" yet, but was hoping it would be a slam-dunk verification of what Audition did.
 
Edit: Another side-point of note is that it is unclear what algorithm Audition uses to generate its mp3s.

I always thought this was a good way to go if I ever got really serious about mp3s:
 
http://www.sonnoxplugins.com/pub/plugins/products/pro-codec.htm 
 
Shucks, tho, ILok 2 required :(

dcumpian

irvin
I know this will sound like poor or casual advice, but there is actually not much behind creating an MP3 file from your project (in Sonar or any other DAW). If you have a good mix/master, your MP3 will be fine. If not, it will sound like crap - because it's coming from a crappy source.

So, I guess the real question behind all your questions is how to create a good mix - and that's a topic well beyond the scope or intention of this thread, so, I'll just leave at that. Good luck!

This really isn't completely true. Lame is updated fairly regularly and there are a lot of switches that affect the final conversion. However, you really want a good front-end for those switches to make it easier to see what you are doing and experiment with the different settings. Just Google "Lame front end" or use a client like Fre:ac and you'll be able to find the settings you need for various outputs.
 
Regards,
Dan

Yes, you can get as technical as you want, but - just look around this thread and you find some people praising lame, some others saying it sucks and everything in between. Very similar to the "which is the best DAW?" or "which is the best maximizer?" questions. My advice amounts to: anything you use will give you very good (as in "virtually un distinguishable from each other under normal listening conditions") results without much effort - it's far more productive to concentrate on getting a good mix, for example.

We live in very good times and the emphasis should be on making good music - most of the technical decisions have been made for us, and that's a great thing for the overwhelming majority of users. That was my point and the OP may or may not find it valuable. That's ok, too.

With the current trend of actually selling MP3 files. I guess it's good to understand as much of this as possible.

Sadly, I have not gone into it that deeply though I've been getting very good results within the medium's limitations...

I export my mixes as 24/48 and open them into a new Sonar project setup for mastering..

When I'm happy with my mastering I export as 16/44.1 wav as well as (using Sonar's MP3 dialog) a 320 stereo, high quality MP3....

I don't expect them to sound the same, but as long as it seems to resemble the mastered wav file effectively, I'm satisfied. It seems to me that the majority of MP3 files sound less quality due to heavier (lower) bit rates...

I would rather people buy the wav files. I played with modified mixes specifically for MP3 conversion and could force more elegant files, but then they stick out awkwardly in most MP3 collections. So I stopped.

Dropping the gain does give a touch more clarity, but again the offset here of volume wars makes it a moot point knowing that especially with MP3 collections, the listener is not putting on your album and setting playback. The listener is setting playback and hearing playlists.... All need to conform somewhat.

So I've been making my mp3's at -0.3 and continually get praise for my MP3 files sounding good...

If the MP3 sounded as good, they'd gave less reason to buy the album for higher quality, eh? ;-)

They're a hardly medium for passing around work versions of songs and the likes, but though I alliw sale of them I would much prefer to have listeners buying higher quality...

Again not to minimize the need/desire for quality, but don't overthink this.  Spend time on mixing and mastering and any decent (256 or 320) mp3 will sound fine to everyone.
 
Going down the rabbit hole of different encoders, conversion algorithms and all the blahdeblah is just taking time from where change and quality actual matter:  the song, the mix and the master.
 
 
 
 

sven450
Again not to minimize the need/desire for quality, but don't overthink this.  Spend time on mixing and mastering and any decent (256 or 320) mp3 will sound fine to everyone.
 
Going down the rabbit hole of different encoders, conversion algorithms and all the blahdeblah is just taking time from where change and quality actual matter:  the song, the mix and the master.
 

That was my point, exactly. Well said.

I just setup Lame today and after a little fussing I got it to work. I'm currently going to export 320k bit rate. I have not figured out how to manage the peaks etc.
 
What I have been doing before was exporting a wav and then converting it using Nero Wave editor. I also normalize the file before the conversion.
 
I will be watching this thread to learn more as well.
 
 

fif4lifefif
Welp I must admit I skimmed through your post a bit as I have a meeting to get to, but this is a post that I found helpful when looking to maximize the quality of my lossy format work. 

Nowadays, I've gotten great results by working with my mastering engineer to get a "Soundcloud Master," which has a very gentle lowpass rolloff (CRINGE, I know, but it works), and a peak of -1.0 dBFS. 

 
Are you rolling off the lows or Highs?  As I understand a lowpass will roll off the highs. I'm only checking in to make sure I understand what you are doing.

The point is actually on-topic as I see it.  The particular bit rate for the avenue one is distributing through must be taken into consideration for the output file.  Otherwise that distributer will then be converting the conversion.  
 
Consumers have a choice, I've never bought an MP3, but I have thousands of CDs/FLACs/WAVs and a rather extensive music collection.  They only lack the choice if they choose to limit where they acquire their music from.  (i.e. iTunes).  The trend you note isn't anywhere close to what it used to be when CDs were popular, not even in the same ballpark.  So I have trouble viewing it as "growth"
 
As for the OPs question.  My method is to Mix and Master for the WAV.  Take that Track and encode it at the exact bit rate the particular distribution method would use.  The free Program Traders Little Helper works fine.  The quality differences at a particular specified bit rate are generally pretty negligible and my experience suggests those consuming MP3s truly are not really going to care about such subtle differences anyway.  
 
As for all the discussion of you need to master at sub 3 db, etc.  Well if you rip a CD that is fully limited and then convert it to MP3, it generally sounds like you would expect it to, right?