Rack plugin Performance test

[ArcoZ]

I've made a short comparison of RRack and corresponding VST plugins:
DAW: Reaper. All settings the same on both channels

1st Channel:
Serum
Rack (Tsar1 + Tsar1 + FET Compressor)

2nd Channel:
Serum
Tsar1
Tsar1
FET Compressor

Results in Performance Meter:
1st Channel: 3,12% - 5,01% - 0 PDC
2nd Channel: 1,85% - 2,62% - 256 PDC

Conclusion
So ..... Rack Plugin eats ca. 80% more power than corresponding plugins but deals better with Delay Compensation.

[dezma]

That's not my experience.

I did the test with Europa VST vs Reason rack with Europa with the CPU heavy poly/friendly keys patch both playing the same notes.



I also benchmarked my first track I converted from reason only to rack vst + studio one.

It got down from 46% CPU running in rewire to only 14% (this one is way less scientific as it's not a 1:1 comparison.

So with my "limited" testing, I'm more than okay with these results.

[Loque]

0 PDC is suspect...

[dezma]

I will need to correct my previous post.

I kind of hoped the rack overhead would become less when using multiple racks, yet they are a stable 4% per synth OR effect, which will eventually add up really soon if you run a lot of devices.

In this test example, there's already 12% cpu power lost due to rack overhead, which would equal a full blown unison lead from a synth flagship.

In the screenshot is:
Line 1 vst = line 3 Reason Rack Plugin
Line 2 vst = line 4 Reason Rack Plugin



Note this is on a battery optimized U- cpu-laptop running asio4all, not on my workstation. Assuming ArcoZ is on a serious workstation, results would indeed be closer to 2%.

I'm curious about other results.

[guitfnky]

don’t most other DAWs allow you to easily freeze tracks or otherwise handle high CPU overhead? of course, I don’t mean to question the usefulness of such testing. just thinking that from a practical perspective, any additional overhead from the rack VST may not be much of an issue.

[antic604]

I did the test with Europa VST vs Reason rack with Europa with the CPU heavy poly/friendly keys patch both playing the same notes.

This is misleading, because you're ALSO comparing Europa VST with Europa RE.

[dezma]

I did the test with Europa VST vs Reason rack with Europa with the CPU heavy poly/friendly keys patch both playing the same notes.

This is misleading, because you're ALSO comparing Europa VST with Europa RE.

Why is that misleading? I'm doing 1:1 the same with europa VST as I do with the rack vst europa. And can see again the consistent 4% difference?
What's the difference with ArcoZ' test besides its from Props?

[boingy]

This is misleading, because you're ALSO comparing Europa VST with Europa RE.

Why is that misleading? I'm doing 1:1 the same with europa VST as I do with the rack vst europa. And can see again the consistent 4% difference?
What's the difference with ArcoZ' test besides its from Props?

The RE is not identical to the VST. For example, the VST does not support CV connections. So I don't think it is a valid test.

[EdGrip]

Also Europa VST is made by Propellerhead, and is already a port from the Reason environment - I would expect Europa VST and Europa-in-Rack-VST to be more similar in performance than, say, Devastor VST Vs Devastor RE-in-Rack.

[guitfnky]

I did the test with Europa VST vs Reason rack with Europa with the CPU heavy poly/friendly keys patch both playing the same notes.

This is misleading, because you're ALSO comparing Europa VST with Europa RE.

he’s literally told you about it—how can that possibly be misleading?

[antic604]

This is misleading, because you're ALSO comparing Europa VST with Europa RE.

he’s literally told you about it—how can that possibly be misleading?

Because he thinks performance-wise they're the same, which they're not. And the topic is "Rack plugin performance test" so obviously he should exclude any other factors that might impact the results.

[dezma]

he’s literally told you about it—how can that possibly be misleading?

Because he thinks performance-wise they're the same, which they're not. And the topic is "Rack plugin performance test" so obviously he should exclude any other factors that might impact the results.

Ik also did the test with Rob papen products which are ports from vst to RE with exact same results being 4% overhead. It's still a lot for some cv i/o unless the re technology itself is inferior to vst.

[guitfnky]

he’s literally told you about it—how can that possibly be misleading?

Because he thinks performance-wise they're the same, which they're not. And the topic is "Rack plugin performance test" so obviously he should exclude any other factors that might impact the results.

no. misleading would be doing a test without letting you know the parameters/how it was done. you're free to take the results as you will, but there's certainly nothing misleading about it.