I normally use Reaper as my DAW. When I create mp3's at the highest standard 320 Kilobits/sec Constant Bit Rate. directly from Reaper using the LAME Encoder -I find that the mp3 versions do not sound identical to the rendered wav versions.

What would you recommend as the best approach to creating mp3's from my DAW, that are as indistinguishable from the 16 bit of 24 bit .WAV renders, as the lossy compression method of mp3 permits?

Thanks

Not sure to have correctly understood: you'll never get an mp3 sounding as a wav...the first is lossy and the second not, so no matter how hard you try, you will get better results avoiding ISP, for example, or using a better encoder (I personally do that wirh Sequoia), but mp3 even at 320 will aound worst than wav.

Thanks.

I'm not sure why, but in my testing there's something about the mp3 export in Reaper that's not quite right. I've gotten better results using a standalone lame interface (Razorlame) which allows deeper control of settings. MediaMonkey also seems to do a better job with the lame codec. Both are free.

Thank you. Response truly appreciated.

I'm relieved that I'm not the only one who hears a definite difference between Reaper's mp3 renders, puzzling enough to want to eliminate this as much as possible.

I assume you render in 24 bit .wav then convert to .mp3 in Razorlame.

One more question, which render method in Reaper do you consider gives you the highest quality .wav output, as I have experienced a difference between these renders on some plugins, so I thought it good to ask others?

1. Online render or

2. Offline render

Yep.

I haven't noticed a difference. I usually use offline for ITB only processing but always bring the rendered file back into the project and make sure it nulls with the track I rendered from (I can't remember it ever not nulling).

If you are aware of which plugins do this then you may just have to use the type of rendering that works best for those.

Cheers,

Dom

Thanks - I love your idea about bring the .wav track into Reaper, to ensure it nulls (I make the assumption that a phase reverse is applied to the rendered track after importing to achieve the null).

Thanks also for your pointer on Razorlame. I tried it out and am definitely happier with the ease with which I can choose conversion options.

All the options, tended to take away some of the bottom end, usually becoming overall brighter than the source wav file. I "developed" a few presets in Razorlame, which should cover most future needs.

I ended up preferring a set of options using Razorlame which did not take too much away from the low end, but maintained and did not overly exaggerate the brightness and dynamics.

Strangely - after a few listens I ended up preferring the MP3 version option I had settled on, via Razorlame, over the original wav file - that slight brightness was quite pleasing, and obviously some distortion - albeit pleasing had been added to the audio, giving a bit of what soldiers would call, from the terminology of polishing boots - a bit of spit and polish - a glossier version. Slightly compressed also. I dare to even use the W word - warmer - yes. Never knew MP3 conversion could add a bit of saturation! Even though I was well aware that this was not an authentic representation of the source..

I've always been suspect of the commercial mp3 equivalence with CD's and have never paid for an mp3 for this reason, never bought one - preferring to listen to the CD's. And I know a good many including seasoned musicians who happily buy the commercial mp3's exclusively, by default.

How the music industry ever got the entire world to buy into the lie that High Quality MP3's are equivalent to the CD, beats me. If I can easily hear the difference in my monitoring environment which costs less than $1000, how did most of the world buy this lie.

Thanks again.

I seriously doubt that any human being will tell the difference between .wav and proper mp3.
If anybody disagrees - try double-blind testing at least 4 out of 10, without guessing, and return to report how you failed. Difference is ****ing inaudible.

Partially agree with you, but it depends on how you're approaching: when you blind test on a final user environment it's true, but if we are talking about mastering (and due to the forum where the thread was posted I assumed that) the difference in a proper room with proper monitoring is clearly audible.

For the audio I converted to MP3, I could easily tell on my monitors which was which, it definitely sounded different, no matter which options I had chosen in the conversion to MP3 format.

As I said earlier, some even sounded "better" than the 24 bit wav version, which I interpret to most likely be - more distortion, more compression.

I however since learned from web sources that I need to give some headroom in the source wav file, and not attempt to convert a final master that had already been limited to -0.1 Db.......

It's like, MP3 needs its own separate mastering process, from the non-mastered audio file, so the same way I would approach a track and create a master for CD/WAV format, I also create a different master aimed at MP3.

Which makes sense.

In the old days, I can imagine that there were different masters created for Vinyl, reel to reel, cassette, and later on CD, which if you listened to them critically did not sound exactly the same....

I think I now understand the need for some of these Fraunhoffer plugin tools, which simulate the MP3 encoding process and playback process in real time, in your DAW so that you can adjust in mastering, to retain as much of the target sound, as possible, and have far less of a surprise after conversion. Will find time to try these out.

Something specific, and along the lines of what has been said here, for an MP3 to be rendered properly you cannot set the ceiling to -.1 db. You need more headroom than that. Export the file with a ceiling of -1 and see if that clears any distortion during playback.

A one-off encode to 320kb/s can sound ok, although it's source-dependent and encoder-dependent (and the difference signal will show you the artefacts and the effect of the low pass filtering, sometimes also high pass filtering, and the resultant clipping).

However, try encoding a wav to MP3 and then not applying any further DSP to it at all – no digital volume change, or SoundCheck/normalizing or EQ, and certainly no further lossy encoding or radio broadcasting. These are where the compound processing takes audio quality down hill faster than the steepness of said low pass filter.

Here's what I discovered, that has improved the quality of the MP3's.

Almost indistinguishable from the 24 bit version of the WAV render. Which is a great result. Compared to the 16 bit WAV version, there is unlikely to be a consistently audible difference!........, even in a high quality monitoring environment.

1. Rendering from the DAW to a WAV file, at a higher quality as best as you can in bit depth or sample rate.

2. Where possible using source audio WAV which peaks at no more than -0.2 Db.

3. Using razorlame transform the WAV to MP3, @320kb/s using VBR, rather than direct render to MP3 from my DAW.

Thanks to all who responded. Highly and eternally appreciated -this has been a life changer - learning how to do this properly, and delivering professional quality results, that made all the difference. Thanks indeed.

Did you mean 9 out of 10? I tend to agree with you about blind testing 320k MP3 vs WAV, but 4 out of 10 is less than coin toss 50/50 random chance. The following says 95% confidence is required, which would be 9 out of 10:

https://en.wikipedia.org/wiki/ABX_test#Confidence

Foobar 2000, Wavelab Pro, and Apple AURoundTripAAC have quite nice blind testers.

I'm on a Mac and been converting thru iTunes. Is this bad? Is there a good mp3 program for Mac?

It's very difficult to tell the difference between a proper 320kps mp3 and a wav in a blind test. This sounds like a reaper issue. Good luck.

You could try XLD, which uses LAME. Many, if not all of the online retailers or record labels use LAME to convert. But as long as you select 320kbps MP3 in iTunes, I think you'd have an impossible time telling the difference. iTunes 320k MP3 should be as good as anything else including the more expensive programs IMO.