NO rant , just facts Cubase 13 Audio engine Spikes compared to Cubase 12

Ummm , i think am @Norbury_Brook Markus was sort of trying to point out , they are very system specific , my whole last build was geared round Acustica , but by G’s own admission , the new Hyper tech is geared round the new processors so , i’m an advocate for keeping old tech alive especially if it’s the top of the chain tech so , between Steinberg having some sort of Over Sampling leak and AA having Naffed up over sampling on it’s last 4 releases ive decided to go back hardware based , lighter load on the Cpu .
Everything was fine till their Ash Ultra release and then iyt completely Naffed up my system , unrepairable but unlike Steinberg , who jumped on the Situation and acknowledged there was an issue , AA tried to keep everything closed , you were not allowed to discus with others to see if they had had the issue and more importantly a resolution , the post was deleted and i received a warning for posting the Vid showing the issue . THAT was the limit , over 2k spent on their plugins and you get treated like your nothing . Steinberg on the other hand showed willing for resolve which is the most important thing to me . 5 days after my machine broke and receiving 7 messages from AA users that were to SCARED to speak up in public AA released fixing for 4/5 of the latest plugins for guess what ’ broken Over sampling recall issues ’ .
I will not even entertain loading their software now , i want total stability , not plugins from a company that release everything at an intro price because you are beta testers .Also bare in mind AA are meant to be a Steinberg Partner which makes it a little bit more of an issue .
As you can see im still not over the heartache caused but goods came out of it , this DAW is pro setup , All crash dumps from C12 and windows dumps were checked for the issues while the new install of windows was installed . It’s flies like never before .

Ive discovered the new black Valve compressor can do most of AA stuff if you add the stereo enhancer just before :wink:


getting my 7950x made a huge difference to how many AA plugins I could use . Having the AVX instructions makes a big difference as AA seem to make good use of it.
Did i read that the current gen of INTEL CPU’s has dropped support for AVX? that might be a reason they’re not performing as well…not sure about apple silicone.


1 Like

me too

Been using 13 for 1 week now. Im still using 11 (never liked 12) and the difference on the same machine just proves that the issues are with 13. Stop thinking your daw is doing something weird or your cpu isn’t good enough. Its all rubbish.

C11 runs just fine with now sudden CPU spikes. In 13, all i have to do is load up ZebraHz which is almost what, 10 years old? ) and the meter goes up all the way.

Unacceptable. A simple public beta prior to release would prob catch these major issues. Everytime i have issues, Steinberg points to “it must be some plugins” when the reality is: All plugins i have work fine in Live, Bitwig and whatever else I have. None of them even ever crashed due to one. But Cubase? God forbit. Im more than happy to beta but when I pay for an upgrade, I dont want showstopper bugs and performance issues. In 2023.


MacBook Pro M2 Max here, both Cubase 12 and 13. Very busy project runs fine if I have one or more of my VSTs record armed/enabled, but as soon as I don’t, ASIO Guard peaks and spikes and playback stutters and stops. It doesn’t exhibit this behaviour on my PC.

This is at a small buffer size of 64. If I raise it to 768 I can just about playback in the busiest parts of the track, but the ASIO Guard meter is completely full.
Again, on my PC (9900K processor) it pretty much plays well at 64/128 buffer size no matter what track I select and ASIO Guard stays calm around the 3/4 mark.

Oh and finally, if I select any Halion track, that simply makes it peak out regardless of whether it is record armed/enabled or not…

Ok, it seems Halion Sonic 7 is the culprit… disable this VST and I can even playback the project at 32 samples whilst selecting any track, record enabled or not. Is this a known thing that I’ve missed? It definitely seems very resource heavy in slightly odd ways compared to other VSTs.
ASIO Guard is still a little higher and stretched compared to my PC, but seems stable enough now.

The first project I opened in 13 tanked from measure 1. Nothing but stuttering. As a percentage, I can say there’s more silence from stuttering than there is music playing.

Interestingly, the projects from 12 mix down fine in 13. It’s just everything else.

And if I start a new project in 13, it’s fine up until I get to around 3 instrument tracks. Then it starts stuttering, again.

Ive been getting great results with my new install and C13 , even with older projects BUT the ASIO is still erratic and a little dicky bird tells me that there’s a maintenance update coming pretty soon

1 Like

Can anyone shed any light on to whether this has been resolved ?

1 Like

Yep, for me it did. I have almost identical behaviours in both performance meters now.
A Steinberg dev asked for a project of mine some time ago to check the differences. He confirmed, that they made some small adjustments to the core-affinity of the CPU in C13, which they tweaked again in C13.0.20, so that they should be identicall to how it was in C12.
Now C13 performs as well as C12, perhaps it even performs a little better, as the peak meter is a tiny bit lower in C13.

1 Like

I have not been able to interpret what’s going on with the Cubase metering, but I am currently focused on researching the DPC issues.

I will say that I’ve been maxing out the ASIO meter with Dom’s stress test, yet the “real time” meters show very little activity while the CPU cores are actually near 100% occupied.

I have not looked at this closely in 13.0.20, but it appeared again not to make any sense. For now, I want to stay focused on the DPC issues.

Although I can pretty reliably get a high channel count with Dom’s test, the 15-minute DPC spikes are causing intermittent issues with dropouts. Sorry I don’t have a solution yet. However, I did want to confirm odd metering and uncertainty about how to reliably interpret the meters relative to audio issues.

The realtime meters should be low, as long as you don’t record anything or have any monitoring activated on a track.
The manual says:

Shows the average audio processing load of the latency critical processing path. This is the case for all live inputs or monitoring activated channels.

But its interesting, as if my cores are nearly 100% used, all three performance meters peak higher. At least “Asio Guard” and “Peak” should be higher here.

Ummm, interesting , ive made the jump , the skiing issue seems to be solved all thou the Asio AG now acts very erratic according to the info that’s being played and this is with the 'Suspend plugins ’ disabled . It’s like the Asio AG should be on just over half way but the lowers and instantly jumps up into like a peak limiter . It’s go , no where near peaking ,i think the Asio AG seems to be acting more real time now than average

1 Like

This is far from resolved . the asio guard is just sooooo erratic

I had the same experience. C 13.0.02 made a huge performance gain over the previous version on my test mix. It now performs on par the the macOS silicone version I’m running on Sonoma.


Yeah, again I’ve written elsewhere, but I’m chiming in here as it’s def a bit odd.

ASIO Guard is def not behaving right… I’m running a MacBook Pro M2 Max (UAD Apollo X) and on a busy project AG runs very close to max with the odd spike pushing into the red and stopping playback regardless of buffer sizes. This project and others run absolutely no problem at very low (64/128 sample) buffer sizes and AG running at around 3/4 regardless of if I have an instrument or audio track selected on my previous PC (i9 9900k).

One other thing I have noticed whilst doing these back and forth tests is that Halion seems very unoptimised compared to its native Intel counterpart. Very processor hungry…

Anyways, really hoping this gets resolved. In theory my M2 Max really should be smoking my old 2018 i9 but sadly it isn’t as yet…
Oh, and one thing I forgot to say, these projects were originally created on my PC machine. Could this possibly cause any issues with how the plugins are maybe loaded and in which order, etc etc?!



This is definitely not okay.
I just got the Cubase 13 demo (still working with V12…).
Here is a project with a classical orchestra using VSL Prime Samples in the Vienna Ensemble plugin. Freshly created with Cubase 13.
There is an effect group channel with reverb and sends to that channel.
For testing, I use an Izotope Neutron 4 plugin in a woodwind group output.
I also tried an Aqua Plugin, but then it was almost impossible to work with because of overloads
This is what the audio performance looks like:
Cubase 13_VE_Ochestra_No_ASIO-Guard-Meter
The ASIO guard doesn’t show anything here (of course the ASIO guard is activated on normal and in other projects it also shows something…), the peak moves back and forth a lot when you play a few tracks, and you hear peaks and ‘scratches’, although there is rarely a red overload indicator on the right.
My CPU is only at about 16% usage and all 16 cores are evenly utilized and far away from overloaded.
Actually, you can’t work properly with this version.

I hope that there is still an attempt at improvement, but to be honest I have little hope.

1 Like

Might be the plugins.

No, it’s not the plugins.
What I didn’t see myself: in the picture you can see that the audio guard is not active at all. You can tell this because the line in the display is very thin. The Asio-Guard simply switches itself off, even though it is actually switched on in the settings.
For example, I have this project where the ASIO guard switches itself off as soon as I connect the output of a MIDI track to a rack instrument, if that is a Vienna Ensemble instrument.
In another project, the ASIO guard turns off when I set a track to record ready. This always happens, for example, when the connected instrument is an Ample Bass Y.
… and so on… after hours of testing I now need a break…

Good time to read the documentation: ASIO-Guard
Restrictions: The ASIO-Guard cannot be used for: Real-time-dependent signals
… and …
If you activate the monitoring for an input channel, a MIDI instrument, or a VST instrument channel, the audio channel and all dependent channels are automatically switched from ASIO-Guard to real-time processing and vice versa


I can fairly say THIS issue has been resolved , i can now work with the meter on 98% and still not glitch out , it’s very stable

1 Like