What is Cubase 12 Sound Engine

I am considering using cubase 12 for a simple high quality DJ manager/controller.

I am obviously not speaking of control features
but of sound quality.

I have heard that cubase has a “sound engine”,
and I don’t understand what it means
since DA converters are, one main factor in audio quality.

This has kept me from wanting to use real DJ software since I understand
they must have a “sound engine” as well,
but probably not as HQ as cubase,

Can someone help me understand what steinberg means by “sound engine”.

Riverside CA

This is the part of the software that is working directly on producing the audio real-time. Since this occurs ‘under the hood’ the engine metaphor got used. But it’s not something you will generally need to think about or interact with. Kind of how you don’t need to think about your car’s engine when running errands.

If your only concern is sound quality, then you can close this thread.
Digital DAWs don’t have a sound (except of Harrison Mixbuss, which introduces non linearities).

If you’re looking for live use and not mastering/mixing, forget about the quality of the AD/DA converters as well. Today even the entry level interfaces of Focusrite, Steinberg, etc. have good enough AD/DA for that.


That’s right - any professional piece of software will sound as good as any other, so choose the software that has the functionality and stability you need.

1 Like


Thank you for your time raino!

Thank you for you time Richard

Thank you for your time tj99

It must have being
steinberg was talking about
better functionallity
when they wrote on an email to me sometime back
that some version, I believe version 10 or 11 " had a better sound engine".

Well they do change it over time as new technology & capabilities become available. Kind of like saying this years Honda has better acceleration than last years.

thank ou raino

An optional 64-bit floating-point mixing engine was added in Cubase 9.5. PG, the main developer of Wavelab, listed the pros and cons in this post:

The effect of this is extremely subtle, so it’s mostly for peace of mind and for a possible small performance increase.

Cubase 10 introduced the ability to handle and record audio files in 32-bit integer and 64-bit float. This is useful if you’re using an interface that supports 32-bit audio such as Steinberg’s UR24C.

There have been other changes that aren’t directly part of the “engine”, if you want to be technical, but may affect your end results. For example, Cubase now has some of the best sample rate conversion in the DAW market (great if you’re using imported samples and loops), and the Sampler Track now lets you increase the quality of its pitch shifting algorithm. Furthermore, the AudioWarp tools and VariAudio are constantly being improved upon.

Thats great to hear!
I stopped using Melodyne long time ago, because VariAudio just sounded better, let alone the much better integration.

You think so? I am still experimenting with melodyne but I find it really good soundwise, especially when changing drastically some notes, something that variAudio didn’t responded as good.

Also I use a lot the “pitch drift” tool, which is something like how symetrically will stay a vibrato related to the center pitch. And I can not find (is there one?) in VariAudio. I can see the “straighten pitch curve” option in VariAudio which is equivalent to “pitch modulation” tool in Melodyne but unfortunately not the drift.

Also why am I forced to use the standard solo algorithm when editing in variaudio? Is it the best for every situation? When I change it to elastique it changes back to standard and prompts that the standard is required for variaudio editing.
I can read on manual:

Standard – Solo

This mode preserves the timbre of the audio.

élastique Pro

For best audio quality

So I would like to have the option to try the elastique and deside for myself. When I compared the algorithms in the past the elastique sounded better on time stretch, so there is a chance that will sound better but I guess I will never find out :joy:

Now you may say Melodyne has also one algorithm for monophonic material and you 'll be right…

Yeah absolutely. Somehow variaudio sounds more natural to me and the detection of tunable notes and plosives/sibilants is usually also better in my cases.
But I rarely have to correct performances where the singer is off more than a semitone.

I remember one time where I was using melodyne for a performance which was off by way more and everything sounded weird to me compared to the same performance corrected with variaudio afterwards. YMMV.

Also there is no way I’ll go back to the melodyne workflow… :thinking: :joy:

I will directly compare both for sure, I am very interest in this because of

the whole transfer thing and then what you see is not what you get so you have to render if you need direct offline processing if i.e. realtime rx9 mouth declick is not removing some specific clicks, this whole thing is close to nightmare…

I 'll request the drift option for vari-audio :upside_down_face:

1 Like

Cubase 9.5 may have added 64 bit processing, but it’s in Cubase 10 where the actual audio engine has received a massive overhaul.
The following image shows the result of a 48 kHz sine sweep after conversion from 96 to 44.1 kHz.
This is just self explanatory. Top is 9.5, bottom is 10.

Website is http://src.infinitewave.ca/
You can also compare other results like THD, and hundreds of programs are available to choose from.
Not to mention that Cubase 10 does incredibly better than some other popular DAWs.

Just for the fun, here’s Ableton Live 7 for comparison :

ableton live 7 src

What the hell. That are some obvious differences. If this isn’t a reason to jump ship if you’re using Ableton, then I don’t know what is.

But if thats really the case, why isn’t that discussed everywhere? I am active on a lot of audio nerd forums, but never came across a topic which discusses this. :thinking:

Logic, Reaper and Cubase all seem to be in the same ballpark. Protools on the other hand introduces artifacts. I already hear the Protools fanboys: “thats what makes Protools sound so good and PRO” :joy::joy:

Yeah lol :joy:
But in reality, the back bounce after it reaches 22 kHz that you see on Pro Tools (2021) isn’t really important, since you will most likely set a low-pass filter at around 20 kHz just before the limiter, so this “artifact” won’t be part of the final mixdown.

I’d be much more concerned about the artifacts that occur in the 0 - 22 kHz range.

Howerver, still on the Pro Tools sweep, and other DAWs too, we can see that there is some noise across the full range, which I hope, is just dithering noise.
In the case that it would not be caused by the dithering, but by the resampling filters, then this is a real problem. If you look at Pro Tools 2019, the noise is much lower, so yeah : What the hell.

Hi, thanks for that @Romantique_Tp !

It was a few versions ago i looked into this closely, and Cubase’s sample rate conversion (SRC) was not one of the best at that time, it actually was measured as one of the worse ones.

Looking now i see there was a big jump in SRC quality between Cubase 9.5 and Cubase 10, the latter "as good as"R8 brain, all according to these plots:


Thanks for that heads up, i wasn’t aware!

(PS: Anyone know any further details of what Cubase did to offer us a better SRC?)