From the original SynthMaker forum (2004–2014), with its members’ names and dates.
datsound ·
What is the maximum CPU useage for a synthesiser so that musicians can use it happily? before they think: "this VSTi is crap it uses too much processor."
based on 44100 SR with a 2GHZ processor:
if you make a synth that uses 20% on one voice, then it would use 40% on 5 voices. and 97% processor would be eaten up at 96000 SR.
My conclusion is that: if your synth uses more than 15% percent on one voice, a studio musician working at 96 SR would very soon think "this is shit" unless you limit to 4 voices, unless he has a 2.8Ghz processor. if you have more than 25 % on 1 voice @ 2ghz you wipe out 2/3rds of your users?
Do any of you change the number of voices dynamically according to CPU levels?
MichaelBenjamin ·
if the synth/effect runs well, and the code is somewhat clean with no excessive waste, then it does not matter how much cpu it uses. either it runs, or the cpu is not fast enough. however you easily get the impression of every plugin that uses nil cpu to be the best and most awesomely coded brilliance - while somewhat true the other side is that all those savings and approximations also go into the sound. i would say if you can do interesting things with the cycles, use them all.
Cyforce ·
datsound wrote: What is the maximum CPU useage for a synthesiser so that musicians can use it happily? before they think: "this VSTi is crap it uses too much processor."
based on 44100 SR with a 2GHZ processor:
if you make a synth that uses 20% on one voice, then it would use 40% on 5 voices. and 97% processor would be eaten up at 96000 SR.
My conclusion is that: if your synth uses more than 15% percent on one voice, a studio musician working at 96 SR would very soon think "this is shit" unless you limit to 4 voices, unless he has a 2.8Ghz processor. if you have more than 25 % on 1 voice @ 2ghz you wipe out 2/3rds of your users?
Do any of you change the number of voices dynamically according to CPU levels?
I would say it belongs to the sound-quality which the synth offers.
Best example U-He ACE, on highest setting he nearly eats my complete cpu, but the ACE sounds superb, and so the most users approve the cpu-usage for a real good sound.
For "normal" synths on my system(i7-920 on 32bit) would be my limit about 5-10% - if the sound is right^^
trogluddite ·
I agree, each user balances quality, CPU use, and cost.
I think that on the whole, professional users will choose quality, and they will most likely have a faster, more efficient PC (because it is their job to make music, they invest in the best hardware too) - and will pay a premium price too (so developers can still make money from less customers)
Amateur bedroom musicians (no insult, that is what I am), mostly prefer cheap and low CPU, but also don't have the speakers/environment to hear any extra quality, so will be satisfied with a few compromises to the sound.
That's a gross simplification, I admit, but I think it is a case of choosing the market for your plugin - many of the big plugin makers sell pro and 'amateur' versions of their plugins for precisely this reason.
BTW - My first music PC was a Pentium MMX 200MHz - I had to bounce tracks to a new audio file every time I even used one mono synth plugin - and it wouldn't run a reverb that was worth listening to at all (external audio hardware was essential).
Really we should be thankful for the amazing things that our PCs can do - if we have to do a little 'track freezing' and bouncing to audio, that is not so hard - not compared to editing tape with a razor blade anyhow!
attic ·
My first music PC was a Commodore C64 so we indeed have come a long way. However I am now searching for a used C64
for making music. Ironic isn't it.
infuzion ·
trogluddite wrote: I think that on the whole, professional users will choose quality, and they will most likely have a faster, more efficient PC (because it is their job to make music, they invest in the best hardware too) - and will pay a premium price too (so developers can still make money from less customers)
I'd say pros don't upgrade their computers that much; takes too long to upgrade everything, & workaround the "broken" older software that worked fine on their older OS version. But if a user hasn't upgraded in the last 5 years I'd say forget about them.
I expect VSTs to break the quantum space-time continuum, & actually provide a nill CPU drain.
Dell ·
I limit myself to 10% sitting on a Pentium 4, 2.8 GHZ using 44.1 kHz. Higher than 10% are not offered publicly. So far the highest I've offered publicly is 7%.
If you can achieve 15% or less on a 2.0 GHZ, it should keep those who use your plug-ins happy.
Cheers!
djbrynte ·
Maximum 10-15% depends on al effects. Mine is using maximum 10% when im finished. Its 4 osc mega synth 😀 cheers bud!
MegaHurtz ·
Im at about 3% max for the most.(FM8 is about 1.5/2%).. The last litlle granular delay ran unoptimised at about 1.5% here.
Im using about 10% on my latest project.. but its HUHGHE, nevermind the sequencer proportion ram cosumption 🙂.
It aint a modular couple of oscillator little thinghy. But now im bragging 😊 .
but 3-4% for a lot..
djbrynte ·
Depends on waht pc you have? mine synth use half on my friends pc because of mine is old.
MegaHurtz ·
Barely, depends on crappy vs good programming TBH.
these are based on a 2.3gz dual core laptop wit 4 gigs of ram.
djbrynte ·
ine is slower
Dell ·
MegaHurtz wrote: Barely, depends on crappy vs good programming TBH.
these are based on a 2.3gz dual core laptop wit 4 gigs of ram.
I believe datsound is using a single-core 2.0 GHZ.
I design mine around a single-core 2.8 GHZ.
Cheers!
MegaHurtz ·
Not to be confusing dsp with actual processing kids.
These are streaming instructions. SIMD 😉
Wrote something here? To have a post removed or your name hidden, write to [email protected].