SX / Nuendo Performance Loss - A Myth ?

[susiwong]

Sascha,
coming from LAW I don't doubt your statement that LAW is more efficient than SX.

BUT, and that's a big BUT , :

1. LAW doesn't have PDC and it seems clear to me that this is a very CPU - hungry feature.
Relying heavily on DSP cards I won't ever switch back to a host without PDC, so I have to accept a certain performance loss.

2. Dynamic CPU management is cool - no question.
But don't you think that to shed light on Steinberg's possible CPU / Dongle - issues it's necessary to compare two almost identical projects with no host - specific EQs, instruments or plugs but identical (3rd party) plugins which are used from start to stop simultaneously ?

3. This is not meant to be a comparison SX vs. Logic in terms of real - life useability.
That*s a valid question in it's own right, no doubt, but not the topic here. I agree with you in so far as LAW still can keep up amazingly well with today's hosts and certainly can get a pro job done even years after it's untimely "death".

Do you have access to a machine running both SX and one or more of the other "big" hosts ?
I would very much like to see some numbers for a test project along the lines of (2.) .

Kind regards,
susiwong

[ericj23]

http://www.kvraudio.com/forum/viewtopic ... 11#1216011

did the test yesterday - cubase uses the same cpu as samplitude, fruity and traktion 2 on a p4 using the same plug-ins on each

it's a myth

PS sacha - logic is now (finaly) able to allow you to jam using more than one vst at a time so if the ones being compared were set up like that all would be square - this is after all how cubase works

oh how i don't miss the cludge of working round to layer up synths

[nuffink]

Well, if you look here, you'll see that they're pretty upfront about having tried the crack and compared it to the protected version.

What an insane thing to do. They might as well take out an advert...
"We know this one works, but it's no better than ours. Honest".

[ericj23]

Well, if you look here, you'll see that they're pretty upfront about having tried the crack and compared it to the protected version.

What an insane thing to do. They might as well take out an advert...
"We know this one works, but it's no better than ours. Honest".

they don't usually bother - but then the crackers implied that cubase would work better on there version - it is important to tell steinberg users that that is a lie

[Sascha Franck]

1. LAW doesn't have PDC and it seems clear to me that this is a very CPU - hungry feature.
Relying heavily on DSP cards I won't ever switch back to a host without PDC, so I have to accept a cetain performance loss.

Well, if you switch it off in SX (or "reduce" it), it still doesn't make up for much better performance.
Besides, when not using any FX tracks or groups, it shouldn't be any different as within Logic.

2. Dynamic CPU management is cool - no question.
But don't you think that to shed light on Steinberg's possible CPU / Dongle - issues it's necessary to compare two almost identical projects with identical (3rd party) plugins and no host - specific EQs, instruments or plugs which are used from start to stop simultaneously ?

Maybe.
So, what should that project look like?
If you come up with something using only a mixture of audio tracks and freeware VST instruments and FX I'll gladly make a comparison.
I own both SX3 and Logic 5.5.1.
I still think about trying out the Samplitude demo as well - but I wouldn't happen to know how well that would compare.

But anyways, no matter what the test results will look like, that still doesn't prove anything. It might still be that SX could be a lot more efficient without the encrypted CP stuff inside.

However, just set up a test project with proper descriptions and I'll compare the two.

[Sascha Franck]

oh how i don't miss the cludge of working round to layer up synths

I layer them using Energy XT. And I even do so in SX because it's way more comfortable.

[Josmoker]

I still want that dynamic plugin allocation thingie in Cubase
Now THAT's a feature that would make me upgrade instantly (still on SL1 here).
Pretty please?

Cheers, Jo

[Sascha Franck]

they don't usually bother - but then the crackers implied that cubase would work better on there version - it is important to tell steinberg users that that is a lie

No, the crackers didn't imply that.
All that was said was that it'd perform better without the copyprotection stuff being encrypted so deeply. And of course they didn't strip that out.

[ericj23]

ive already done the comparison with samplitude sascha - its cpu useage is exactly the same as cubase's during playback

admittedly i had playback on all channels so no plugs were getting dynamically turned off

PS so logic can dynamicall turn off plug-ins inside another host (ext) - now that is clever

[ericj23]

they don't usually bother - but then the crackers implied that cubase would work better on there version - it is important to tell steinberg users that that is a lie

No, the crackers didn't imply that.
All that was said was that it'd perform better without the copyprotection stuff being encrypted so deeply. And of course they didn't strip that out.

but if you folow the link in my first post i compared the cpu useage fo cubase with a few other pc hosts - it's the same i.e. there is no cpu hit for the copyprotection on playback - the protection is only there when you are opening and closing windows - changing editor tools etc - ie things that can wait a milisecond or so while a call is made to the dongle

the notion that playback performance is affected is nonsense im afraid

[susiwong]

Sascha,
it seems ericj23 did exactly what I was talking about, a pity he doesn't own LAW though.

ericj23,
thanks a lot for your test, sorry, I must have overlooked your thread. I can live with your results quite comfortably unless somebody comes up with new findings .

For the sake of completeness, Sascha, could you possibly get in touch with ericj23 and try his test on LAW and SX3 ?

Thanks to all of you guys for your contributions so far, seems it is really possible to have a civilized discussion on a controversial subject.

quite impressed,
susiwong

[cold c]

There's no way to benchmark things.
As said, Logic is saving up CPU power when nothing is playing back. How would you benchmark that?

Of course that can be benchmarked.

If logic is utilising less CPU when a plugin is idle (but transport is playing) but no audio is running through it then this can simply be benchmarked by having some plugins in the project which are not used for most of the project, and benchmarking the playback CPU utilisation.

If logic is utilising less CPU when the transport is in stop mode, then this can be benchmarked by measuring the CPU utilisation in stop mode.

Simple enough.

Also, why is it necessary to keep starting new topics on this subject when threads already exist?

And to the original question, "is there significant performance loss because of the dongle"-

The facts available indicate that there is no playback performance loss due to the dongle, this is because it can be assumed that SX3 has more CP embedded than SX2 (which is why it took longer to break), however, there is no performance difference between SX3 and SX2 (from my own testing), therefore the CP does not significantly affect playback performance.

Here it is again-

SX2, significant copy protection/encryption embedded
SX3, much much more copy protection/encryption embedded
Both the same playback performance.

Got it?

Simple enough.

Yes in theory newer compilers could have compensated for inefficiencies in the newer version, but that is simply theory until proven.

Please stop spreading the FUD until you have some solid reproducable evidence, I hate dongles as much as the next person but for different reasons as I stated in one of the other threads,

And yes, I would rather SX3 did not require a dongle, but this is because of Steinberg's lost dongle policy, and the fundamentals of 'pure' native software with the least dependance on an unnecessary piece of hardware (for example: whereas software can be backed-up as many enterprise environments do, a hardware dongle can't and is therefore an added 'risk').

Please stop starting new threads on the performance myth.

[uselessmind]

but if you folow the link in my first post i compared the cpu useage fo cubase with a few other pc hosts - it's the same i.e. ther is no cpu hiot for the copyprotection on plyback

The crackers didn't say that because of the cp sx3 performes worse than other hosts,be it sx2,sx1,samplitude or whatever.
They didn't even say that cp affects playback in sx3.
They only took a somewhat educated guess that sx3 without that kind of cp would gain 50% performance.

Again,there is only one way of knowing:
Test sx3 with and without cp.Everything else is pure speculation.

[susiwong]

cold c,
I did my best to read everything on this topic I could find before starting this thread, on several forums.
But things are moving so fast atm , my apologies for starting this thread, but I still think I had a valid question a bit different to other threads.
Anyway, I'm quite happy with the way this thread has turned out.
cheers,
susiwong

[ericj23]

but if you folow the link in my first post i compared the cpu useage fo cubase with a few other pc hosts - it's the same i.e. ther is no cpu hiot for the copyprotection on plyback

The crackers didn't say that because of the cp sx3 performes worse than other hosts,be it sx2,sx1,samplitude or whatever.
They didn't even say that cp affects playback in sx3.
They only took a somewhat educated guess that sx3 without that kind of cp would gain 50% performance.

Again,there is only one way of knowing:
Test sx3 with and without cp.Everything else is pure speculation.

but this 50% perfomance loss can be shown to not exist on plyback - frankly who gives a f**k about performance at other times - nothing else in useing a host is taxing the cpu to any great extent

2/3 of nothing is still nothing