KVR Audio

ChamomileShark · Post by **ChamomileShark** » Thu Nov 30, 2017 3:23 pm

hi, some of my work needs to be at 48k and some at 44.1k so I have some projects set at 48k and others at 44.1k.

Sometimes I need to shuffle audio between the two sets of projects - I've been using the "import from tracks" and letting Cubase reset the sample rates.

Is this the right way of doing it? As far as I can tell it works both ways (ie 44.1k to 48k and 48k to 44.1k) and to me it sounds fine but I just wanted to check.

lfm · Post by **lfm** » Sat Dec 02, 2017 10:54 am

I tested sample rate conversion with C9 Pro i RMAA(this was 96k->48k).

Cubase is worst in class(green in graph). But since introducing new 64-bit audio engine in 9.5 maybe this is fixed now, haven't checked yet.

rmaa=original pattern
snr= an old Sonar 8
cp9 - cubase
r8 - r8brain.

Also protools, reaper and others are hard to distinguish from original pattern.

Depending on how much processing is made after conversion I would do conversion externally in a tool like voxengo r8brain or similar. That is free.

You can also compare certain conversions quality here:
http://src.infinitewave.ca/

ChamomileShark · Post by **ChamomileShark** » Sat Dec 02, 2017 6:34 pm

thanks for this and the website. Looks like I need to avoid doing this as much as possible - although when doing the 44.1k to 48k or the other way I wasn't hearing any difference. Probably need to also have a look with a scope though.

lfm · Post by **lfm** » Sun Dec 03, 2017 2:10 pm

No, you would not hear immediately - it's at low levels. Just a comparison of distortion created.

If you mix and modulate and process in various ways it can show up there as something you don't want and try to get rid of. A resonant filter can boost certain frequencies up into audible range.

It just hurts to know you use a tool that create worse result than others - so you are aware.
Like when selecting hifi amps - you at least glance at THD to compare apart from listening.

And every daw of main daws did constant improvements on this, which Steinberg did not do for 10 years - and I hope Steinberg also did something about it as they introduced 64-bit audio engine in 9.5. I mean even their own product WaveLab had better conversions - in house - that is odd. Cubase was overlooked, kind of.

I also wrote wrongly 96k->48k, my rmaa tests was like infinitewave site 96k->44k.

ChamomileShark · Post by **ChamomileShark** » Sun Dec 03, 2017 3:51 pm

hmm, ok that's interesting - I have Wavelab Essentials so I might use that if I notice any problems

SJ_Digriz · Post by **SJ_Digriz** » Mon Dec 04, 2017 5:40 pm

yeah, down SRC in Cubase is the suck from detailed measurement standpoint. Use Wavelab or something else. SRC up should be identical, at least that's been my experience. But, to be honest although statistically the Cubase SRC doesn't measure well against some other tools, if you are doing a single file export I don't notice an audible difference.

Romantique Tp · Post by **Romantique Tp** » Mon Dec 04, 2017 7:16 pm

The SRC tests you'll see online are intentionally exaggerated because the actual difference in quality between DAWs isn't very big.

Playing a 48kHz sample library while working on a 44.1kHz project will introduce more aliasing than Cubase's SRC, even if your sampler is set to its highest quality setting.

SJ_Digriz · Post by **SJ_Digriz** » Mon Dec 04, 2017 7:19 pm

Romantique Tp wrote: Playing a 48kHz sample library while working on a 44.1kHz project will introduce more aliasing than Cubase's SRC, even if your sampler is set to its highest quality setting.

Actually, Halions internal resampling produces pretty much identical results to Cubase. Falcons is different, I don't know about better but the noise and aliasing are in a different place. Kontakts is crap. And again, as long as you aren't doing batch conversion with summing, you won't be able to hear it.

Romantique Tp · Post by **Romantique Tp** » Mon Dec 04, 2017 7:26 pm

They were all worse than Cubase's SRC when I tested using the test signals from Infinitewave. I checked the results side by side using Wavelab, and the Cubase result matched the Infinitewave SRC test.

SJ_Digriz · Post by **SJ_Digriz** » Mon Dec 04, 2017 7:30 pm

Romantique Tp wrote:They were all worse than Cubase's SRC when I tested using the test signals from Infinitewave. I checked the results side by side using Wavelab, and the Cubase result matched the Infinitewave SRC test.

Did you do Halion 6 and Falcon? I didn't go back to Halion 5 to do the tests. But the conversion was dang near identical doing that same test. Falcons was odd, as I expected the aliasing to be similar. But, it had its own flavor. But, it was all so far below the noise floor that I've never been interested in figuring out the various reasons for the differences.

Changing sample rates in Cubase 9