Hi,
I wanto to find some hardware that alows me to input some analog channels (microphones and instruments, around 18) and then mix them.
I want to access each channel for processing it in matlab, python ...
Is there any hardware to do that or do i have to build something with a DSP processor ?
Thank you !
I wanto to find some hardware that alows me to input some analog channels (microphones and instruments, around 18) and then mix them.
I want to access each channel for processing it in matlab, python ...
Is there any hardware to do that or do i have to build something with a DSP processor ?
Thank you !
Hardware exists. Recording studios use most of it. It is likely to get rather expensive, hopefully you know that. How many thousands of dollars is the budget? What sound quality is desired? Is digital processing of mic channels to be in real-time, or off-line? How much computational resources does the digital processing require, say, per channel?
If you are using matlab or python the presumably you can do the mixing on the host PC. That being the case all you need is some digital i/o. Here is an example that may meet your spec. Note that the price goes up from here.
Basically I want to analyze each channel and control its volume in real time to mix them for example in a live show. If there is some cards that allows me to do that I would prefer it. If there's no way I want to know what's the solution.
I'm an electrical engineering student who wants to research in that area. What do you think is the best solution for me?
Thank you for your time!
I'm an electrical engineering student who wants to research in that area. What do you think is the best solution for me?
Thank you for your time!
How much real time do you need ? Beware of latency. Btw, the umc1820 above only has 8 actual analog input. You need to add a ADA8200 to get 16 analog inputs. Focusrite has similar hardware (18i20+octopre), but at about twice the price.
What exactly do you want to do with python or matlab ? Is it anything more than a soundcraft ui24r (899€) or a berhinger xr18 (385€) could do natively ?
A final option might be something like a A&H Zed-r16. It's a mostly analog mixer with firewire/adat multitrack ability. But that's a 2000€ console.
What exactly do you want to do with python or matlab ? Is it anything more than a soundcraft ui24r (899€) or a berhinger xr18 (385€) could do natively ?
A final option might be something like a A&H Zed-r16. It's a mostly analog mixer with firewire/adat multitrack ability. But that's a 2000€ console.
This is a totally standard thing to do, but typically we don't use matlab (Except while designing the digital filters)!
Basically you just need an audio interface that has an appropriate number of the appropriate types of IO, a mess of software and (most critically for a live show) a control surface that does not suck....
There are plenty of 24 channel and up sound cards around, but the elephant in the room is usually the need for multiple mic preamps.
These days I would be looking to a suitable Dante equipped stage box, and a copy of "Dante Virtual soundacard" or just writing an AES67 (Ethernet) based processor to do whatever you need done, then hooking it to an AES67 capable stagebox over a 1gig ethernet link.
The audio doings are not ever particularly interesting, with most of the tricky stuff really being about having a control surface that works and (particularly as numbers of outputs go up), having a workflow that makes producing a stereo and a 5.1 and a 7.1 and... Mixes at the same time reasonably easy.
Do remember that while the audio processing core is a realtime thing and should never allocate any memory or contend on a lock for any reason, the UI mearly needs to be quick, a much simpler thing to pull off, making these separate processes is usually a good idea.
Basically you just need an audio interface that has an appropriate number of the appropriate types of IO, a mess of software and (most critically for a live show) a control surface that does not suck....
There are plenty of 24 channel and up sound cards around, but the elephant in the room is usually the need for multiple mic preamps.
These days I would be looking to a suitable Dante equipped stage box, and a copy of "Dante Virtual soundacard" or just writing an AES67 (Ethernet) based processor to do whatever you need done, then hooking it to an AES67 capable stagebox over a 1gig ethernet link.
The audio doings are not ever particularly interesting, with most of the tricky stuff really being about having a control surface that works and (particularly as numbers of outputs go up), having a workflow that makes producing a stereo and a 5.1 and a 7.1 and... Mixes at the same time reasonably easy.
Do remember that while the audio processing core is a realtime thing and should never allocate any memory or contend on a lock for any reason, the UI mearly needs to be quick, a much simpler thing to pull off, making these separate processes is usually a good idea.
Thank you for your answers.
Im new at this, thats why im asking those trivial questions. im just an engeenring student who is very intrested in music.
I want python to read each channel and i understand now that any digital mixer or soundcard allows me to do that.
I want python for some machine learning stuff that will determine the wanted volume of each track.
I understand now too, that the volume can be controled on the soundcard directly and so, I dont need the signal pass througth the computer (only need the inputs to analyse them) so the real time situation im talking about is not so crytical, right ?
I see that Berhinger x32 as a github repository where i can control the channels volumes and let me access all the channels doing a live show right ? I like it because we can see those fader knobs moving "alone" which is good for demonstration. Is there any other cheaper solution (even without the knobs) ?
I don't know how many money i can spend, its the university's decision...
do you understand ?
Thank you very much !!
Im new at this, thats why im asking those trivial questions. im just an engeenring student who is very intrested in music.
I want python to read each channel and i understand now that any digital mixer or soundcard allows me to do that.
I want python for some machine learning stuff that will determine the wanted volume of each track.
I understand now too, that the volume can be controled on the soundcard directly and so, I dont need the signal pass througth the computer (only need the inputs to analyse them) so the real time situation im talking about is not so crytical, right ?
I see that Berhinger x32 as a github repository where i can control the channels volumes and let me access all the channels doing a live show right ? I like it because we can see those fader knobs moving "alone" which is good for demonstration. Is there any other cheaper solution (even without the knobs) ?
I don't know how many money i can spend, its the university's decision...
do you understand ?
Thank you very much !!
Python is a horrible choice for this, slow and garbage collected. You want something much more deterministic, C or C++ being the usual tools of choice for audio realtime processing.
Of course there is nothing saying you could not do your control layer in python, but the audio needs to be in something very much faster and more deterministic.
Now if the machine learning is what actually matters then just get some multi track recordings form the music department, load them into a PC and run them against your routines in whatever language works for you then listen to the result after the computations are done.
If the objective is to build a mix engine, then soundcard and a mess of programming plus a suitable mixer control surface.
Of course there is nothing saying you could not do your control layer in python, but the audio needs to be in something very much faster and more deterministic.
Now if the machine learning is what actually matters then just get some multi track recordings form the music department, load them into a PC and run them against your routines in whatever language works for you then listen to the result after the computations are done.
If the objective is to build a mix engine, then soundcard and a mess of programming plus a suitable mixer control surface.
...like an op-amp.
Add motordriven pots, if you need remote control or the ability to replay a mix in real time.
SSL SL9000J ?
You need to actually explain why you want/need to re-invent the wheel.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.