Why MP3 encoding restricted to 16 bit ?
 

Hi,
I am increasingly relying on the Wav 32 bit floating point as a main format to transfer down the line of the workflow (mix->mastering, mastering treatments -> finalizing/brickwall limiting)
And I am discovering some quite surprising stuff on the way : I recently received to be mastered (yes I know crazy things happen sometimes) an mp3 file with a very hot level, audio in it that far exceeds 0dB when opened and decoded in Wavelab 9.5 ! And no clipping !
To such extent that by attenuating by -10 dB I still had peaks approaching 0dB !!! Obviously not just inter-sample peaks.

Suddenly I realize there are plenty of software out there that output 32 bit float signals, and this includes Pro Tools 2019 ! If you bounce stuff *really* hot into a 32bit float file, open it in Wavelab 9.5 (my goto two track treatment and mastering software), attenuate it, no clipping ! Magic ! (opening it in Pro Tools poses another problem, that I will discuss in another thread)
What this demonstrates is the fact that mp3 encoding is by design a 32 bit float resolution process ! Originally, since most encoding is made extracting CD no wonder the signal of most files circulating is dithered 16 bit just under 0 dB FS
Now I am getting quite pissed off at companies like Avid, Steinberg, Apple etc for not (or little) documenting questions about how resolution is handled throughout their platforms and the output busses and hardware.
For example, I tested the AES/EBU signal on the back of my HD IO, and since this format is 24 bit fixed point, there is a very small and subtle amount of dithering, which actually is a good thing. I used to mix through this AES/EBU output, but now I prefer to bounce my mixes in 32Bit float to really get the max resolution out of PT without any extra signal added.

And now to the subject of this post, I want to point to the way Pro Tools exports mp3 : I see the Fraunhofer logo so obviously it is the official codec.
But it is very annoying that the output is *forced* to 16 bit without any explanation or option !
I perfectly understand the reason, preventing people from creating "ceilingless" overly hot mp3 files that could distort on "normal" players (that where meant to read audio normalled or brickwall limited to 0dBFS or just slightly under)
[Edit to make myself really clear, obviously if you bounce to 16 bit wav or mp3 with a very hot signal, your file will be heavily clipped. The problem with a "hot" MP3 file encoded (not on Pro Tools) from a 32bitf signal exceeding 0dBFS, is that it will play ok on certain players (VLC for example) but will distort like hell on others]

But folks, this is *professional* software, even Logic Pro leaves you the choice of what the input resolution into the mp3 encoder you want. If you choose 32 bit float, no dither is needed and you can happily exceed 0dB FS on your mp3 file ! I not sure this is a good idea, but that, IMHO, is for the user to decide

And by the way, I tested this and I confirm that Pro Tools mp3 encoding is 16 bit as announced and WITHOUT DITHER, so if you don't want to end up with sub standard audio, YOU HAVE TO insert a 16 bit dither plugin on the master fader BEFORE bouncing.
The problem is that Pro Tools offers you the possibility to bounce let's say in 24 bit, AND make an mp3 at the same time if you check the "Add MP3" box. So you are left with this dilemma: Either you dither you 24 or 32 file to 16 bit which is absurd, or you make your mp3 out of truncated 16 bit, which is bad.

Oh and just for the fun of it, if you read a extra-boosted over the top mp3 on VLC, you lower the volume on VLC *OR ON THE MAC* --> no distortion :
Which means that the whole audio chain in a the mac is in fact 32bit float.
At this stage it seems this is not the case with iTunes.

This all for now :-), thanks for reading !

Re: Why MP3 encoding restricted to 16 bit ?
 

I'm not sure you understand bit depth. The number of bits does not increase or decrease levels. it only increases signal detail and raises the S/N ration. You can overload a signal just as easily in 16-bit as higher bit depths.

If a player distorts with a 32-bit version of an MP3, and does not distort with the 16-bit version of that same file, it is broken.

Re: Why MP3 encoding restricted to 16 bit ?
 

The OPs post makes sense to me. Fixed point audio, regardless of whether it is 16,24 or 32 bit will adhere to a limit of 0dbfs where it inevitably clips and information is lost forever. A 32 bit floating point audio file (like used in PT and most modern DAWs) is different. It has a 24bit mantissa that essentially carries the audio sample information, and an 8 bit exponent that effectively scales the range of audio up and down across a huge dynamic range.

If you create a 32bit floating point file with peaks well beyond 0dbFS, you can still play it through a 16bit fixed point converter without clipping. All you need to do is turn down the gain digitally so that the peaks are below 0dbfs. According to the OP, this is possible even with common apps like VLC on OSX.

I never considered 32bit floating point as a delivery standard. It's an interesting thought. How do you compress and brick wall limit to 0dbFS if there effectively is t one?

Re: Why MP3 encoding restricted to 16 bit ?
 

I am sorry if you understood my post like this, I happen to teach and write for technical magazines, so I think I do understand bit depth quite well. I always try not to be too long, which on such subjects is not an easy task... As most of my business is done in french maybe something wasn't clear enough I will read this now and see what can be corrected !
I just edited my port to make things a little clearer, I hope :-)

Re: Why MP3 encoding restricted to 16 bit ?
 

I use third-party software (MP3 Encoder from the App Store), it has batch converting, high quality encoding, etc..

Re: Why MP3 encoding restricted to 16 bit ?
 

we all still use mp3 today but ...

"On April 23, 2017, Technicolor's mp3 licensing program for certain mp3 related patents and software of Technicolor and Fraunhofer IIS has been terminated.
We thank all of our licensees for their great support in making mp3 the defacto audio codec in the world, during the past two decades."

Re: Why MP3 encoding restricted to 16 bit ?
 

Very interesting, I will investigate this, thank you for bringing this up !
Mark

Re: Why MP3 encoding restricted to 16 bit ?
 

you're welcome,
another one

https://www.cnbc.com/2017/05/15/mp3-...licensing.html