I have Cubase Artist. I’m trying to get a high quality MP3 audio mixdown file of the INDIVIDUAL channels in my Cubase project. The channels are all VST instruments. Is there a way to export these channels so they are not the same size as the whole project which is 10 Mb? In my mind the individual channel MP3 should be a fraction of that size. Here’s what I’ve tried so far and still the individual channel MP3 download is the same as the full project:
Disabling ALL the other channels apart from the one I’m trying to export
Muting ALL the other channels
In the Export Audio Mixdown dialogue box, selecting the one channel I want to export
Combination of all above
Surely, one channel audio mixdown export does not have to be same size as the full project? Help much appreciated.
No. Although it might seem counterintuitive at first sight, even when you export to lossy formats (MP3), the individual stereo stems are roughly of the same size as the whole mix, unless there are significantly long portions of silence in the individual tracks. And if you export to WAV, then the individual stereo stems are always of exactly the same size as the whole mix.
I think FLAC is a format, where file sizes become much smaller when there’s lot’s of silence in them. And it’s non-lossy, so I tend to use that format when I want to save tracks while saving disk space.
If that format doesn’t create much smaller files from tracks with lots of silence could imply that maybe those tracks don’t actually contain true silence, but potentially some very quiet (sub audible) audio, maybe created by some insert fx (including channel strip items) on those tracks.
Many thanks for explaining - makes sense. Yes, on some of the tracks, it’s 95% silence, but that hasn’t made a difference. All the tracks need to be the same length, so looks like I’m going to be stuck with the same file sizes. At least I know - thanks.
I don’t know if you misunderstand how file size of audio happens or if you have certain expectations at the MP3 encoder.
Lossless audio: file size = sample rate * bit rate * channels * time [s]
MP3 encoder: CBR (constant bit rate) * time [s]
MP3 encoder: VBR (variable bit rate) * time [s] ----> this will yield a varying file size depending on the content; however, it will not be a lot of variation
I leave out FLAC as I have never dived deep enough into the encoding procedure.
Thanks Nico5. I would try FLAC, but format needs to be MP3 or WAV. And yes there’s a lot of silence in the tracks. Thanks for the heads up about the insert fx and strip - I’ll check them.
If you are doing this to sync up later then absolutely stay away from mp3. I did a project years ago where different parts of a track were exported as individual mp3 files and all were then looped and the user got to mix between them. But apparently mp3 files don’t export sample-accurately so there was crazy drift between the different files (“Stems”) as they looped.
So if you care about timing stay away from mp3. There’s basically close to zero reason to use that format for these sorts of things.
Since no one has mentioned this so far. The Cubase .cpr file size only reflects the size of the Project file & excludes its Audio which is stored separately in the Audio sub-folder. You could for example have a 20MB project file which is using 12GB of Audio.
Also the channels are vst instruments which means they are not audio files. These are midi which then triggers the vst instrument engine. The actual audio is in the vst instrument or where it stores it and as mentioned is not part of the project. In fact it’s not part of the whole project folder. It only becomes audio from when a vsti when rendered.
You have to remember midi us not audio and is just messages and takes up a tiny amount of space compared to audio. Even low bitrate compressed audio.
I think it was about the actual file size of renders, I mean the mixdown of the main output vs separate channels. The main out is a 10 MB file (which the user calls the whole project), and they were wondering why do individual channels aren’t just a fraction of that size, for example if there are five channels in the project they would be 2 MB each… Answered in post 2 by @valsolim
How much silence exactly, tends to depend on the mp3 encoder used to create the mp3 file.
So when exporting tracks as mp3 files from the same Cubase project using the same encoding parameters, they should all have the same small silent moment at the beginning (easy to check via re-importing the mp3 files into a new project) and zooming into very fine detail at the beginning of the imported tracks.
If it’s just a number of tracks all starting at the same time, it’s rather trivial to adjust for that little gap by moving all of the audio events after import by the same small amount. (Select all imported tracks and perform one single move.)
However, this typical mp3 silence is more painful when dealing with dozens of small mp3 files – especially if they are intended to be looped, because then you need to make this timing adjustment for every little imported file.
I’ve read that some apps adjust for that silence automatically when importing mp3 files, but I haven’t tested that myself.
So in some use cases, mp3 files are really quite fine and if the requestor of the files says they’re ok with mp3, I’d take them at their word.
If you are sending the resulting files via ZIP archives, the audio files with long silence should compress much smaller than the audio files without silence.
It’s been a while since I tested all of that, but I remember it like this: Just create 2 ZIP files from identically encoded audio files of the same audio duration, but one file has no silence, and the other has lots of silent stretches. The resulting ZIP files should be much smaller for the tracks with silence.
I did this all out of one DAW, Pro Tools, several years ago. They didn’t loop correctly. Whether it was the top or end I don’t really know, all I know is that there was a difference.
Regardless, I can’t really see why anyone would want to do this. .Wav is surely better and disk space is cheap.
haha - very funny - but not a good analogy for encoded data streams, which can create widely different sizes of files, depending on the compression and encoding algorithms in use.
The closest analogy in the world of food I can think of is food that’s been dried or freeze-dried.
But in reality, all digital music is already an encoding of instructions how to move some air back and forth.
So you could make the argument that digital music files are more akin to recipes that can be fed into a machine that can create food from scratch - Star Trek style.
2 wav files of the same length are the same size regardless of content. 2 mp3 files encoded at the same settings will be approx the same size, unless there is silence. This is known info.
You’re just upset cos I made a joke that pretty much sums up hundreds of words worth of technical posturing. And now you resurrected my joke. Thanks.
But there’s always one who will try to prove you wrong on these forums isn’t there. Even when you’re just talking about a sandwich. I accept that and wish you a great day ahead.
Meanwhile… I’ve got 2 sandwiches to eat that I didn’t even really want.
Well, I was deeply into a remixing community some years ago and therefore bumped into this issue quite deeply in the context of many projects (not just one) - so that may explain, why my take is more nuanced than yours.
Sure, WAV is better, but …
I’m still seeing lots of forum posts here where people make music with a lot more modest budgets and setups than you and I may have.
Or maybe just lousy or expensive Internet.
Not everyone has the same budget and access to technology.
