as you can see, almost every other daw has it beaten. Logic, Ableton, pro tools, etc.
Cubase is full of noise and aliasing.
Considering how many sample rate conversions I need to do on an average day as it always seems audio I am importing needs to be converted to the project SR… i would like to think we would have a better algorithm.
Even Nuendo (that site sadly only has up to nuendo 5) is bad… They have up to Cubase 9. Check the graphs.
Nuendo is a 2K daw!
Why don’t Steinberg incorporate the crysal resampler from wavelab which is at least decent?
I think i’ll use an offline dedicated program to do conversion and then import from now on, but i shouldn’t even think to need to do this.
Are most SRCs really that bad?
No. If you look at the decibel scale to the right from the graphs, you can see that the range of these graphs is very wide: down to -180 dB. The distortions generated by most properly designed SRCs are below -100 dB and can hardly create audible artifacts. However SRCs differ in the transition band of the low-pass filter and in the amount of pre-/post-echo and aliasing. The bottom line is that most tested SRCs range from fairly good to excellent, but the graphs are very sensitive to emphasize the differences.
Cubase doesn’t have the best sample rate conversion around, but the quality is pretty good and introduces far less aliasing than using 48000Hz libraries in 44100hz projects with any sampler, which a lot of people do. It’s completely inaudible in 99% of the cases and a bit of low cut makes it almost completely disappear.
Wavelab now uses very high quality SoX resampling. The Crystal Resampler isn’t used anymore. As you may have guessed, this isn’t something that they can just copy and paste into Cubase, otherwise they would have already done it.
ok maybe not crystal resampler… but they can absolutely program sox into it and let the user choose if they want to use it… like they allow various algorithms to be chosen for stretch for example.
I am sorry i do NOT buy that they can’t code in other algorithms. Sox is open source isn’t it? Lol. Why can’t they add it exactly?
I have never ever imported files into a sampler that are the different SR from project… because i already know how bad most of them convert, I do the audio files first in something like r8brain on my windows laptop.
BTW, 99% of all my sample libraries and i have 2TB’s worth of just samples, are 44K… You are the first person to mention 48K sample libraries to be honest!
The issue happens to me when i am doing projects at rates other than 44K! If i worked only at 44K, i’d never have an issue LOL.
I said that changing the resampler isn’t a simple copy and paste job, I didn’t imply anywhere that they can’t change it.
48kHz libraries are far for uncommon, specially for orchestral stuff. AFAIK everything made by Spitfire is 48kHz, and even some of the stuff included with Komplete uses that sample rate.
Sorry the way i understood your post was that it was too difficult. My apologies.
regardless, i think it’ actually a very simple task according to a dsp programmer i have been talking to about the issue.
Arbitrary is what he called it, particularly in the case where the algorithm is already done… we are not talking about steinberg writing their own new SR algorithm… THAT’S what would be difficult and very time consuming (and pointless since perfect ones like izotope already exist).
PS i am happy to work with spitfire at 48KHZ but funny cause i often work at 88k so still, unnecessary upsampling then back…
still, since almost everything else i have at 44k i’d be converting everything… which is why i want a better algorithm.
The fact that one can humanly hear the quality it is concerning.
I think it’s a fair request as it’ stayed the same for a decade… apple thought it was important enough to drastically fix it themselves in newer OSX versions which is what logic uses.
In fact steinberg would do better just to use apple’s algorithm from OSX.
I’m sure it’s on the list.
I made the same request some 5 or 6 years ago
It would more than anything give some peace of mind, knowing that SRC is up to par.
Even if they used the open source SRC that Wavelab uses, would be an improvement.
Not that anyone would be able to tell the difference TBH.
