KVR Audio

jens · Post by **jens** » Mon Oct 12, 2015 1:40 pm

Omkar wrote: While Disto shows 20-25% i.e. on S1, Reaper is at 5-7%

With a quadcore CPU this would be about the same though - S1 Shows per core, Reaper the percentage of available processing power...

but I just tested with S1 and could run significantly less instances - at 24 there was some crackling already... this is however by far not as severe a difference as what has been reported here. My money is still on some nasty bug...

Omkar · Post by **Omkar** » Mon Oct 12, 2015 2:51 pm

jens wrote:S1 Shows per core, Reaper the percentage of available processing power...

Ah, I wasn`t aware of it, thank you.

jens wrote:but I just tested with S1 and could run significantly less instances - at 24 there was some crackling already...

So at least the results between S1 and Reaper are in common, I bet on Vsti`s it is similar.

quintosardo wrote: Did you see if the load is the same while the song is playing? If it is worse when the song is stopped it may be the usual "denormal" issue.

Yes, the load stays all the same high cpu usage and I would exclude a "denormal" issue.

quintosardo wrote:It isn't good at all if you cannot use it.

This is one of the rare plugins I accept to have it printed to continue mixing,
the quality is superb and worth the extra steps.
Also, to blend the print with original mimics nicely a parallel mode.

However, I wouldn`t say no if you come up with an Eco Mode solution

Compyfox · Post by **Compyfox** » Mon Oct 12, 2015 2:59 pm

zeep wrote:
Omkar wrote:While Disto shows 20-25% i.e. on S1, Reaper is at 5-7%
Those differences are huge, between 3 and 4 times as fast (light on CPU) compared to other hosts.

With the other comments like "3-4 instances on i5 (from 2012)", this is indeed the most CPU intensive plugin out of this house.

Something I criticized years ago already with the forced 8xOS (that started with GTS-39) and no way to turn it off (the code was/is there in theory with Roundtone - render HQ, realtime usage LQ). And that sadly was also the case with the C165A (beta, mind you!).

Even Klanghelm's recent compressor (MJUC) doesn't crap on my CPU/ASIO load if I have more than 24 instances running (realtime, LQ mode, not even 50% ASIO load at 128 samples, x64). So this is definitely not a good value in direct comparison. So thanks for the feedback zeep/jens. Definitely appreciated. Even if I would be able to buy this plugin (Distro) - it's not CPU efficient. So I'm also going to pass.

Furthermore - Reaper and S1 can't be compared due to the "latency engines". Reaper will always show way lower values while introducing a load of latency (that was my experience with Reaper 4 at least).

quintosardo wrote:Did you see if the load is the same while the song is playing? If it is worse when the song is stopped it may be the usual "denormal" issue, too low levels send CPU crazy.

Something that should not be happening these days anymore?

But then again - what do I know, I'm just a plugin user - not a programmer.

jens · Post by **jens** » Mon Oct 12, 2015 3:18 pm

Compyfox wrote: Even Klanghelm's recent compressor (MJUC) doesn't crap on my CPU/ASIO load if I have more than 24 instances running (realtime, LQ mode, not even 50% ASIO load at 128 samples, x64). So this is definitely not a good value in direct comparison.

On my machine both Disto and MJUC have a similar CPU load.

Compyfox · Post by **Compyfox** » Mon Oct 12, 2015 4:55 pm

HQ mode of MJUC or Normal Quality mode?
Just out of interest.

You're also on a i5 Haswell, I'm still on a i7 920 Bloomfield. So results can differ.

jens · Post by **jens** » Mon Oct 12, 2015 7:14 pm

MK1 Stereo HQ (which is the most CPU intensive) tends to use a tad more than Disto Mid/Side with everything enabled...

I think it's safe to say that there is some bug in Disto which causes extremely high load under certain conditions... however there is no reason to doubt that Quinto will fix that, so any judgement currently would be premature.

Compyfox · Post by **Compyfox** » Mon Oct 12, 2015 7:38 pm

I just hope the best for the users...

In my case even with the most recent version of GTS-39 (loaded couple of weeks ago), the CPU load if MJUC is still considerable lower. So if it's an optimization thing, I hope the right part of code will be addressed. For all plugins.

jens · Post by **jens** » Mon Oct 12, 2015 8:05 pm

Can't say anything regarding GTS-39 - don't have it...

kvaca · Post by **kvaca** » Tue Oct 13, 2015 8:11 pm

hi there,
I have 2 basic questions /since theres no demo version/:
-are there any noise added which cannot be switched off?
-are there any oversampling used which cannot be switched off?

quintosardo · Post by **quintosardo** » Wed Oct 14, 2015 8:08 am

Yes (comes from the modeling, about -96 dBfs)
Yes (4x)

kvaca · Post by **kvaca** » Wed Oct 14, 2015 9:37 am

quintosardo wrote:Yes (comes from the modeling, about -96 dBfs)
Yes (4x)

I was afraid of it, thanks

Omkar · Post by **Omkar** » Wed Oct 14, 2015 11:15 am

Compyfox wrote: In my case even with the most recent version of GTS-39 (loaded couple of weeks ago), the CPU load if MJUC is still considerable lower.

I completely forgot it`s in my account and downloaded it a few days ago,
do you see on your version three options at the bottom left to run the plugin:
draft, oversample, oversample on export ?

The draft mode performs very well in S1 (lowest value the meter can show)
and the last option is something I can think of having on Disto as well.

Not exactly what I have on mind with an Eco Mode,
but a switch of draft and oversample on export may be enough
to keep Disto as a permanent instance while mixing.

Compyfox wrote:Reaper and S1 can't be compared...

Sometimes life is simpler than one may think.

I compare daws by comparing them
and when I see Reaper (5) performs better, then Reaper performs better.

transmetropolitan · Post by **transmetropolitan** » Wed Oct 14, 2015 12:28 pm

I prefer to commit to a sound and print it rather than run an inferior version of a plug when CPU gets tighter.

I suppose it's just a preference thing, but for me this often has a way of focusing the mind as a track comes together.

MJUC and Disto are about on a par here in terms of CPU.

Compyfox · Post by **Compyfox** » Wed Oct 14, 2015 2:25 pm

Omkar wrote:I completely forgot it`s in my account and downloaded it a few days ago,
do you see on your version three options at the bottom left to run the plugin:
draft, oversample, oversample on export ?

Yes, I see these features "now" - I seem to have ignored them on first re-impression (due the the high CPU load, I didn't use SKnote tools in months, then the fall-out last December happened).

Omkar wrote:The draft mode performs very well in S1 (lowest value the meter can show)
and the last option is something I can think of having on Disto as well.

The questions are:
Why does the software constantly have to run in fixed OS mode?
Why don't all plugins have an OS on Export mode, while the realtime mode is pretty much set on default?
And on a sidenote - why is the GUI either so dark that you can barely read anything, or so bright and "angled" as if you're working on a photo and the knobs are hard to read, or it is so small that you can barely read the values (Strip 3 comes to mind)?

Omkar wrote:I compare daws by comparing them
and when I see Reaper (5) performs better, then Reaper performs better.

But Reaper is still a world in itself with it's ASIO/Audio engine. It will always show lower values and can not(!) be compared to other hosts.

quintosardo · Post by **quintosardo** » Thu Oct 15, 2015 7:46 am

This one is offered as part of five bundles for Save The Children Charity 2015:

http://www.kvraudio.com/forum/viewtopic ... n#p6258632

Get it there for some music but also something really good!

Disto released! (Distressor & Fatso in 1)