List of rack extensions and their CPU usage

[Advanced Suite]

You guys seemed to have covered everything so I am sure it's out there somewhere I just can not find it. Can someone point me in the right direction please?

[submonsterz]

Theres is no right direction concerning this all systems unless totally identical hardware throughout will be different on cpu/dsp .

[Advanced Suite]

Oh ok, I don't know enough about computers. But I do understand what you said.
Thanks submonsterz

[Dante]

Sometimes there are RE latency ratings - that would be just as useful guide.

[submonsterz]

Oh ok, I don't know enough about computers. But I do understand what you said.
Thanks submonsterz

I forgot to add that a sort of rough guide as in say for instance :-
Light
Light to low
Medium
And heavy.
As the catergories or something simular.
That would be helpfull especially to users with lightwieght
Systems that struggle more easy.
The tests would probally be best run on a very low system spec ie a single or dual core set up as a rough guide as that type of system will instantly tell if an re is heavy or not.
But as i say it would only be a very rough guide at most.

[tobypearce]

Here are the results from a recent series of stress tests I did. I'm on quite a fast computer, but the relative differences should hold across different devices. The number is the max number of separate instances playing a single chord before Reason stops playback.

Single note init:
Thor 191+
Parsec 97+
One 84
Predator 95
Quad 63
Veto 68
Antidote 72
Lithium 169
Tres 38
Subtractor 230+
Redrum - single kick 350+
Kong single kick 198
Kong single kick with 3 x onboard fx 122
NN19 (sampling the worst of the bunch here, then playing it back) 192+


Stack mode - full spread, nice amount of detune (v = voice)
One 1 Osc 8v 12
One Osc 4v 23

Quad 1 Osc + FX 19
Vecto 1 Osc + FX 30

Antidote 8v 34
Antidote 12 v 33
Antidote 4v 42
Antidote 2 Osc, 8v each 30
Antidote 2 Osc, 8v+12 dyad each 23
Antidote 8v, all FX bar reverb 23
Antidote 8v, all FX bar reverb 23 +12 dyad 23!
Antidote 8v, all FX reverb 17

Megasaur init 28v 35
Megasaue 3 osc 84v 35


Predator 1v spread/detune 42


Effects
One instance of One synth on init, typical use fx settings. Then duplicated devices (i.e. multiple One synths, each with it's own single FX insert)

RV7000 convolution 47
RV7000 arena long decay 77
RV7000 "Octaves" sample from Valhalla Shimmer 52
RV7 74
Saturation knob 56
Faturator 48
TSAR-1R 24!
None (to check) 84
Line 6:2 (in combinator) 64
Mixer 14:2 (in combinator) 68
FET Compressor 48
Trident 47
EVE AT1 (4 band) 51
EVE AT 4 (3 band) 50
EVE MP5 51
Trident and TSAR-1R 19
Echo bode, Dream Delay 49
Cakewalk RE-2A 41
Pulveriser 48
Echo 50
F-16 4 pole high pass 54
F-16 1 pole high pass 51!
Tube Tech full 33
GQ-7 single band high pass 44
GQ-7 5 band 41
Channel strip in full use 79 (i.e. channel strip is way more efficient than any other fx)
Sidechain using Selig Gain and Shape 31
Selig Gain 60
Shape 33
Scream 66
Plain combinator 76

M Class EQ full bands 73
M Class Compressor 77
M Class Limiter 74
Audiomatic 39
Decimort 2 28
Devastor 41
DCAM Envelope Shaper 49
Softube Amp 34
Softube Bass Amp 40
D-Filter High Pass 4 pole 47
D-Filter High Pass 1 pole 46
Mc E670 65
Mc C670 54
Mc FRG 4RE 54
Mc FRG EEE 58
Mc Moo Q 60
Mc Moo C 55
RP Distort high pass only 38
RP Distort full 33
Multiband compressor 24
TMC Phaser 35

[WongoTheSane]

Great!! Can you explain this:

> Channel strip in full use 79 (i.e. channel strip is way more efficient than any other fx)

...and the "+" and "!" on some items? (I guess "!" means "unexpected result"?)

[tobypearce]

No problem.

Channel strip in full use = compressor, gate, HP, LP and EQ all active and moved away from default settings.
+ = at least this many (stopped counting)
! = interesting

[WongoTheSane]

Oh right. Very interesting, thanks!

[Advanced Suite]

Here are the results from a recent series of stress tests I did. I'm on quite a fast computer, but the relative differences should hold across different devices. The number is the max number of separate instances playing a single chord before Reason stops playback.

Single note init:
Thor 191+
Parsec 97+
One 84
Predator 95
Quad 63
Veto 68
Antidote 72
Lithium 169
Tres 38
Subtractor 230+
Redrum - single kick 350+
Kong single kick 198
Kong single kick with 3 x onboard fx 122
NN19 (sampling the worst of the bunch here, then playing it back) 192+


Stack mode - full spread, nice amount of detune (v = voice)
One 1 Osc 8v 12
One Osc 4v 23

Quad 1 Osc + FX 19
Vecto 1 Osc + FX 30

Antidote 8v 34
Antidote 12 v 33
Antidote 4v 42
Antidote 2 Osc, 8v each 30
Antidote 2 Osc, 8v+12 dyad each 23
Antidote 8v, all FX bar reverb 23
Antidote 8v, all FX bar reverb 23 +12 dyad 23!
Antidote 8v, all FX reverb 17

Megasaur init 28v 35
Megasaue 3 osc 84v 35


Predator 1v spread/detune 42


Effects
One instance of One synth on init, typical use fx settings. Then duplicated devices (i.e. multiple One synths, each with it's own single FX insert)

RV7000 convolution 47
RV7000 arena long decay 77
RV7000 "Octaves" sample from Valhalla Shimmer 52
RV7 74
Saturation knob 56
Faturator 48
TSAR-1R 24!
None (to check) 84
Line 6:2 (in combinator) 64
Mixer 14:2 (in combinator) 68
FET Compressor 48
Trident 47
EVE AT1 (4 band) 51
EVE AT 4 (3 band) 50
EVE MP5 51
Trident and TSAR-1R 19
Echo bode, Dream Delay 49
Cakewalk RE-2A 41
Pulveriser 48
Echo 50
F-16 4 pole high pass 54
F-16 1 pole high pass 51!
Tube Tech full 33
GQ-7 single band high pass 44
GQ-7 5 band 41
Channel strip in full use 79 (i.e. channel strip is way more efficient than any other fx)
Sidechain using Selig Gain and Shape 31
Selig Gain 60
Shape 33
Scream 66
Plain combinator 76

M Class EQ full bands 73
M Class Compressor 77
M Class Limiter 74
Audiomatic 39
Decimort 2 28
Devastor 41
DCAM Envelope Shaper 49
Softube Amp 34
Softube Bass Amp 40
D-Filter High Pass 4 pole 47
D-Filter High Pass 1 pole 46
Mc E670 65
Mc C670 54
Mc FRG 4RE 54
Mc FRG EEE 58
Mc Moo Q 60
Mc Moo C 55
RP Distort high pass only 38
RP Distort full 33
Multiband compressor 24
TMC Phaser 35

This is great! Exactly what I was looking for.
I may copy and paste this to the top of the thread if that's ok with you. I will give you credit still.

Thanks man

[Advanced Suite]

I want to add sychronous to that list how do I do the test with the effects?

Which instrument did you pair with it?

[tobypearce]

Hi.
I paired it with KHz One.
Since you have a different computer it might be difficult to compare results. I did happen to test the reverb on synchronous today. With just the reverb I got 38 instances.

[jonheal]

The Thor, he seems to be an efficient little bastard, doesn't he?

[satyr32]

I made something similiar some months ago.

On each instance of a device I did a different preset to average it out.
Number of instances till stoping playback. On each synth instance I also had a matrix running:

12 Softube TSAR-1R
15 Etch Red
16 DC-9 Overdrive
18 T2 Phaser
20 C1-L1
20 Steerpike
22 DCAM Bus Comp
22 Synapse DR-1
23 Chenille
25 Softube Tube Delay
26 Polar Dual Pitch Shifter
32 DCAM Chan Comp
33 Audomatic
46 Synchrounous
47 Alligator
49 3Plex
50 Titus
51 Softube Spring Reverb
55 Synapse Ninety Vintage Phaser
58 Pulversiser
58 RE 200 Bass Enhancer
77 Distort Chain
79 Resampler Bitcrusher
92 Glitch
96 Bitspeak
98 The Echo
100+ SB Filter Pattern
100+ Scream 4

Synth, each with Matrix
22 Parsec
25 Additive Oscillator
25 Antidote
26 Korg Monopoly
30 Table Oscillator
31 FM4
31 Revival
31 Tres Mono Synth
31 VK-1 Viking Synthesizer
61 Radical Piano
66 Re-Tron
67 Redominator
100+ Ammo 400R Modulation Oscillators
100+ PX7

[Advanced Suite]

Nice Satyr32 thanks for sharing.
That px7 is pretty efficient as well
It was missing from the previous list.

[tobypearce]

Interesting. Thanks.
It's good to see that we have got similar results.
The TSAR-1R is quite disappointing isn't it.
It sounds great so I have it one a single send channel, but its high toll puts me off using it as an insert.

You've also reminded me to dig back into the PX7, which I do have but haven't used for ages.

[satyr32]

Interesting. Thanks.
It's good to see that we have got similar results.
The TSAR-1R is quite disappointing isn't it.
It sounds great so I have it one a single send channel, but its high toll puts me off using it as an insert.

You've also reminded me to dig back into the PX7, which I do have but haven't used for ages.

You can use twice as many TSAR as with my computer. What do you have? I have a Macbook Pro 2.2Ghz i7 4770HQ which not really suitable for music production.

[tobypearce]

Mine is faster than yours I think.
Remember I used it with One, with a large decay, and then duplicated both The synth and fx, not just chaining the reverbs. Is this what you did also?

[satyr32]

No, I had a subtractor and then chained the FXs. But it is well possible that chaining uses more CPU than single FX as you did it.

[mcatalao]

No problem.

Channel strip in full use = compressor, gate, HP, LP and EQ all active and moved away from default settings.
+ = at least this many (stopped counting)
! = interesting

Can you tell what cpu you have? Just because the last test i did with mine would only complain with the computer too slow at 400+ stereo channels with audio.

The mixer channel loads a fraction of any other devicei reason. As a matter of fact that would be the performance i expected from from a surface.

Additionally, it is important to define the source you are using (audio, a device, a sampler, or a synth). I would expect better performance from an audio channel, then a sampler, then a synth with the worst performance), and above all, specially for the compressor and gate, that any kind of audio content is passing through the channel, or else the compressor and gate thresholds will never be passed.

[mcatalao]

Oh and btw, I don't see a single reference to NN-XT. I find that odd, because i almost NEVER use NN-19, NN-XT is the De Facto sampler for multi sampled and multi layered sounds in reason.

I'm not complaining, Kudos for your work here, but I'd really like to see some answers. for example, the NN-XT with A Grand Piano, on my Surface, with complex chords (5/6 finger chords with pedal), i can have 120+ instances. Mind this is a simple mobile CPU, thought an I5, but only 2 cores.

[tobypearce]

Hi,

I never intended my list to be exhaustive. My main goal was actually to identify the most intensive devices rather than catalogue each and every one. That's why I just tested one sampler, for reference. Practically speaking, I assume that I can have as many instances of samples as I need on all my projects. (In contrast, I have to think twice before including too many Rob Papen synths!) Same goes for pure audio channels - I'm sure I could have loads.

Channel strip - you're technically right, of course. I actually can't remember the source, but I do remember that the compressor and gate were both active and working - not with quiet (or loud) audio not meeting the threshold. Again, my conclusion was that I can use as many of the channel EQs and compressors etc as I practically want - no CPU penalty.

I'm running a 4GHz i7, 24GB ram.