Best MP3 Encoder - Outside of Cubase (Cubase's isn't the best)

Windows 10, Cubase Pro 12/14

Why even export as .mp3 these days?

My reason:

Mix project export as .wav >
mastering project >
mp3 for testing on smartphone through PowerAmp on Samsung Galaxy with earbuds (some dumbing down of mix too much low end of course) and wired AKG 240 Studio cans (better but a tad thin on low end) to see what is preserved, what is lost, etc.

Ultimate chain:
mastering project export as .wav >
Magix Vegas Pro 22 NLE (video editor) that exports my music video of the composition as .mp4 with audio as AAC 320, 48 Hz.

“Cubase’s (.mp3 export) isn’t the best”

My current understanding according to ChatGPT (but open to correction):

Cubase no longer includes the high-quality LAME engine it used to rely on in older versions. The built-in Steinberg encoder is functional, but not great — especially for dynamic material.

Below is a list of free alternatives but so far the simplest to use for me is Audacity.

But it would be nice to have something within Cubase instead (most likely unrealistic - export is coded into Cubase, not a plugin option as far as I know).

So if anyone has any better alternatives, feel free to comment.

ChatGPT info continued:


So the fix is straightforward:

:star: Do the MP3 conversion outside of Cubase using a proper LAME encoder.

This gives you MUCH better translation, especially for:

  • toms
  • snare body
  • bass warmth
  • reverb tails
  • stereo width
  • transient rounding

:star: A) WHAT TO GET (FREE, SAFE, CURRENT LAME OPTIONS)

:check_mark: 1. LAME + LAMEXP

This is the easiest, GUI-based, no-stress option.

LameXP

  • Includes the latest LAME encoder
  • Drag-and-drop your WAV
  • Choose MP3 320 kbps
  • Choose “Stereo” (NOT Joint Stereo)
  • Converts perfectly
  • Zero bloat, zero spyware

:check_mark: 2. “LAME for Audacity” standalone binaries

You don’t need Audacity — just the LAME DLL/exe.

LAME 3.100 (latest stable)
https://lame.buanzo.org/

This gives you the exact encoder the pros still use.

BUT — you’ll need a front-end or command line.


:check_mark: 3. Audacity (with LAME built in)

Audacity is:

  • Free
  • Light
  • Has LAME
  • Imports WAV instantly
  • Exports a perfect MP3 320 CBR STEREO

:check_mark: 4. ffmpeg (command line, highest control)

Example command:

This is for power users, but it gives the cleanest MP3 possible.
ffmpeg -i LadyInBlack.wav -c:a libmp3lame -b:a 320k -joint_stereo 0 LadyInBlack.mp3

Sure, I could do this but it’s convoluted when doing several tests in a row.

Personally, instead of relying on the opinion of a LLM, I’d rather do a proper ABX listening tests between the Cubase encoder and whatever else (well, that is if I cared what encoder produced the mp3 that I listen with earbuds on my phone to…)

4 Likes

You want the best mp3 encoder? Get iZotope RX. It can even prevent artifacts above 0dB that almost every other mp3 encoder generates (it uses post-limiter or attenuation).

But why do you want the best one? You said that you want to check how it sounds after compression. Listeners will often listen on 128kbps mp3 or sometimes even worse. If you really want to know how mp3 compression will affect your mix, use low quality encoders, low kbps.

1 Like

@battleangel - Very good point. Didn’t think of that.

I was just about to post the same.

OP, I know this is probably not what you want to hear :wink:

But especially at 320 kbps I would definitely try some ABX tests between LAME and the Cubase encode before digging deeper.

On Windows, you could try this or this for example.

But with a little known feature of cubase called “Audio Export Post Process Scripts”, one can actually call any encoder after doing the export. It might require a little scripting, but here e.g. is an old post of someone who wrote a multi format converter

1 Like

PS: If you’re really into testing deliverables, you could also look into something like Oxford Fraunhofer Pro-Codec.

From all your responses, looks like I may be overthinking this. The suggestion to use a mid/low 128 kp/s MP3 encoding instead even through Cubase – since the point is to see how crappy crappy can get – is probably the way to go. :upside_down_face:

Just in case anyone is interested, I wrote an “AEPP” once for myself which uses ffmpeg to convert the export audio to mp3. Not because I think it is the best, just because I immediately wanted to copy that file to my Dropbox folder afterwards:

(only for windows, and there is absolutely no documentation, so you have to ask your favourite LLM what to do with those files :wink: )

My ears sucks so bad, that even an MP3 at 64kbps|22050khz|16bits encoded with the worst encoder sounds perfect to me! :sweat_smile:

I record at 24 bit 48hz and mix down to mp3 at 16 bit 320kp/s for my phone. You can hear the track is slightly more compressed than original recording. Some individual tracks in a recording seem to gain or lose volume when written to MP3 but easy to make adjustments and overwrite to a new mp3.