You provide MP3 files, what is not really representative, because the compression could change the files quite significantly, especially, when we are hunting for such a details…
But what can I see in the SpectraLayers is:
Every single file is different. S1 and Ableton are not the same. So I would say, it’s really the picture of the Audio Engine.
Ableton and Studio One have hard cut at 20kHz. Cubase has cut around 21.5kHz.
Cubase seems to be the most “noisy”. But with this heavily distorted sound, the question is, if it’s not just more colourful in fact.
I’m still not really sure, how did you get the very source. Do you use a VSTi with the very same MIDI notes? Or do you have 1 WAV file and you just export it? Where does the WAV file come from? Or is the source an MP3 file?
From my point of view, the explanation is still the fact, these are 3 different Audio Engines, so every single sounds different. Now, you can decide, if you like one of them more.
I think, there is no more puzzle here.
Do you use the very same plug-in version in all 3 DAWs? Like VST3 version in all of them, or VST2 version in all of them?
Ok but on Studio One and Ableton i have the cut at 20kHz(MP344kHz) on Cubase i have the cut at 21.5kHz but on 44kHz. in 48kHz i have the same cut of Ableton And Studio One
Only VST2
In the last post i exported on Cubase the same vst and preset but in 48kHz. if you see i have the same cut of Studio One and Ableton… Why in 48kHz yes but on 44 no?
No…
In the end I tried an old projectI tried with a fairly complete old project and noticed that noise anyway without having the cut on the 20khz. Anyway from the photos I sent you notice a bang that Cubase has a strange noise that I can’t understand the reason…
No, i not use Modulation
I don’t use a midi monitor. I’m very sure midi data isn’t the problem since I keep it on other projects that I’ve worked on
The question that asks me why at 48khz has the same cut as Ableton and studio one that stand at 44 but Cubase at 44 cuts 21.5?
Ableton has the 20kHz cut if you export it at 48kHz or 44.1kHz Sample Rate.
Cubase has the 20kHz cut if you export 48kHz Sample Rate, but the cut is at 21.5kHz, if you export with 44.1kHz Sample Rate.
What Sample Rate was the source project you exported from (every single DAW)? To make the test properly, you would need to test all the combinations:
Project at 48kHz, export 48kHz
Project at 48kHz, export 44.1kHz
Project at 44.1kHz, export 48kHz
Project at 44.1kHz, export 44.1kHz
Do this for all 3 DAWs, so you get 12 results. Name the files properly to be know for sure, what export and source it is exactly. Analyse the results in trustful tool.
To have any meaningful comparison, the files should really be exported in wav or any other lossless format.
In mp3, we might just be comparing different mp3 versions/engines/settings.
To me, it looks like some samplerate mismatch, like if the Cubase version was recorded in 48k but the source was 44.1k. Please double check that Audio interface, project, and export samplerate are identical.
Also using Win10 and having system sound shared with ASIO has for me produced the same kind of noise.
Assigning the win10 sound to a different audio interface, and using my main ASIO soundcard just for ASIO solved that. But never checked if that had any influence on rendered vsti, on audio recording the cumulative noise was horrible. But then again, it would happen to every DAW and not just Cubase.
I’ve made several attempts in the last few days and apparently it seems that my Cubase with vst exports me strange high frequencies. Mostly it makes a big difference from as I said above to ableton and studio one
Made some tests myself, but only difference I can find is in the MP3 encoding. Lossless outputs between Cubase and REAPER seem somewhat identical (same noise patch from FabFilter One).
Top track is Cubase MP3, and the second is REAPER MP3. Both at maximum quality. REAPER shows steeper cut at 20kHz, and Cubase at around 21.5kHz. I can’t tell if this will result in any audible difference, even at studio/reference conditions. Both bandwidths seem somewhat silent before Nyquist limit and possibility of aliasing. Of course there can be other differences between the MP3 encodings.
Lower tracks are Cubase and REAPER lossless pcm 32-bit floats, both running at 32-bit float engines, and they seem nearly identical, except for the randomness of the noise itself.
What makes you think it is Cubase?
Maybe it’s your instruments preset that generates very high content and the engines in StudioOne and Live filter it out… but the Cubase engine doesn’t.
So each daw has a different encoding in mp3? but I don’t understand why in a 48khz export it has the same cut as reaper … In the end in wav they should all be almost the same the problem was more about mp3 encoding
Yes, I think the difference is in the MP3 encoding. DAWs can have different MP3 encoder libraries, but I think most of them use either Fraunhofer or LAME. If two different DAWs are using the same library, they can still have different parameters ‘under the hood’, other than the user defined bit and sample rate settings, ending up with slighlty different results. But I can’t say if this has any real world significance, even if the difference can be technically measured.
Again, can’t be sure, but High Quality Mode seems to force the MP3 output to 48 kHz sample rate, no matter what the project sample rate is. It is possible that the encoding library has some sort of ‘48 kHz sample rate only’ special algorithm. E.g. more gentle highcut filter slope or something.
Other real life practical reason for outputting 48k MP3s could be, that most if not all todays consumer digital media devices run by default (or even fixed) at 48k. Most MP3 exports usually get sent as email or IM attachments, or as DropBox links etc, directly to clients for daily auditioning or other drafting / demoing purposes, and are most likely listened with smartphones etc. In case of a sample rate mismatch, to my understanding, most modern playback device software do on-the-fly sample rate conversion, to match the device internal audio clock. For example, listening older CD based releases (44.1k) from Spotify on a 48k device. Realtime conversion can have audible degrading effect on the audio quality depending on the device/software, so doing an offline conversion to 48k during the export can result on better quality, when played at 48k device.
Yes. In any case, sample rate conversion will technically degrade quality, and a simple upsampling process won’t improve quality. But, because the conversion from 44.1k to 48k can’t be avoided while most of the the playback devices run at 48k, the conversion will most likely yield better results as an offline process (when done as a part of export in Cubase) compared to leaving the conversion to be handled in realtime during playback on the playback device itself.
For example when exporting direct-to-client MP3s from 44.1k project at 48k, you have with great statistical probability delivered highest possible quality, as you have with great statistical probability eliminated any further need for sample rate conversions done in the client’s perhaps-not-so-high-quality playback system. Assuming that the client playback system runs at 48k, which is highly probable. And assuming that the DAW’s offline sample rate conversion function can deliver higher quality than the client playback system integrated realtime conversion, which again is highly probable. So, delivering lossy encoded files at 48k isn’t exactly a Hifi myth, but more like ‘damage control’ to minimize the possibilty of lower quality conversion happening in the client end.
