So far i used magic only with one mic input. Works very well and any delay is no recognizable. Now i can only imagine but i guess frequency analysis is not that simple. How will magic perform if i had 10 mic inputs and on all of those came in different signals?
It would be pretty great for a band where you have at least one mic for each instrument. You could then seperate the visuals instrument wise and the drums in bass/snare/etc. Would be awesome, but well, is that even possible ressource wise?
Btw is the gpu also used for frequency analysis (using cuda or open cl) ? Or is that sth that can only be done sequentially on the cpu?
how many different live audio signals can magic handle?
Re: how many different live audio signals can magic handle?
From the beginning, Magic was designed to handle multiple live inputs. I've used 10 inputs before with no problems. There isn't any extra delay when using multiple inputs.
Every additional frequency range you use does take a bit of CPU to analyze, but you'd have to be using hundreds of different ranges before you'd notice any performance slowdown on most computers.
Every additional frequency range you use does take a bit of CPU to analyze, but you'd have to be using hundreds of different ranges before you'd notice any performance slowdown on most computers.
Yup, this is actually why I created Magic in the first place. I wanted each instrument to control a separate aspect of the visuals. In fact "Magic" is an acronym -- it stands for "Multitrack Audio and Graphics Interactive Composer". With an emphasis on the "multitrack audio" in this case.It would be pretty great for a band where you have at least one mic for each instrument. You could then seperate the visuals instrument wise and the drums in bass/snare/etc. Would be awesome, but well, is that even possible ressource wise?
No, the GPU isn't used for any audio analysis. That's why a good CPU is important in addition to a good GPU. CPU vector operations are used when possible, such as SSE/etc.Btw is the gpu also used for frequency analysis (using cuda or open cl) ? Or is that sth that can only be done sequentially on the cpu?
Re: how many different live audio signals can magic handle?
It's a pun AND an acronym XD, Awesome.
But thats pretty good news then, thanks.
So freqzency analysis isnt very ressource intensive after all?
But thats pretty good news then, thanks.
So freqzency analysis isnt very ressource intensive after all?
-
- Posts: 712
- Joined: Sun Sep 14, 2014 8:15 am
- Location: UK
- Contact:
Re: how many different live audio signals can magic handle?
A further thought, which may lessen the CPU load in the case of one input per instrument:
There's no need for a frequency-dependent response to isolate instruments from a mix.
Volume-response alone could then be sufficient for many aspects of an instrument's "scene", whereas otherwise frequency-selective processing would be needed. Less CPU load.
Even percussion could need less frequency-selective processing, as there would be no need to eliminate eg bass guitar.
There's no need for a frequency-dependent response to isolate instruments from a mix.
Volume-response alone could then be sufficient for many aspects of an instrument's "scene", whereas otherwise frequency-selective processing would be needed. Less CPU load.
Even percussion could need less frequency-selective processing, as there would be no need to eliminate eg bass guitar.
Re: how many different live audio signals can magic handle?
Nope not really. But like I said though, it will add up after a while if you use it a lot.So freqzency analysis isnt very ressource intensive after all?
True, it will use the CPU less in that case. But the main point is that manipulating frequencies will never give you the same isolation as having the tracks separate from the beginning.Volume-response alone could then be sufficient for many aspects of an instrument's "scene", whereas otherwise frequency-selective processing would be needed. Less CPU load.
Re: how many different live audio signals can magic handle?
I imagine the "tone" and "pitch" work much better with single instrument inputs?
Re: how many different live audio signals can magic handle?
Yup, definitely. Tone works best with percussive instruments (drums, electronic beats), and pitch works best with single-note instruments (human voice, guitar solo, violin, etc.).
Re: how many different live audio signals can magic handle?
Can confirm 36ch input works fine using MOTU hardware on OS X with current version of Magic (3x MOTU 828 MkII).
NB: I don't output audio via the MOTUs due to the current samplerate issue... I output via the MBPro headphone socket.
NB: I don't output audio via the MOTUs due to the current samplerate issue... I output via the MBPro headphone socket.
Re: how many different live audio signals can magic handle?
Hmm, are you referring to this issue? http://magicmusicvisuals.com/forums/vie ... ?f=2&t=298NB: I don't output audio via the MOTUs due to the current samplerate issue... I output via the MBPro headphone socket.
Re: how many different live audio signals can magic handle?
Yup... is it fixed now? (haven't updated my routing in a while to find out)
Re: how many different live audio signals can magic handle?
The sample rate issue with Macs is fixed in v1.54 .
For MOTU devices specifically though, I couldn't replicate any problems. I was able to obtain a test unit from MOTU, and I tried it on both my Macs, and playback worked normally. Can you let me know (in the other thread) if it works for you now, or if you're still having problems with it?
Much appreciated,
Eric
For MOTU devices specifically though, I couldn't replicate any problems. I was able to obtain a test unit from MOTU, and I tried it on both my Macs, and playback worked normally. Can you let me know (in the other thread) if it works for you now, or if you're still having problems with it?
Much appreciated,
Eric