What memory usage am i looking at here? This audio system odd?

[damonb82]

Hi all! I have a few questions that I couldn't solve myself. You can answer them all or answer one lol, but I greatly appreciate your time equally. Because of ya'll helpful people I've been progressing & learning heaps.

What type of memory am I looking at operating here? Do you have any suggested learning material where I can better understand how to assess power requirements? I know each module states it at the bottom of the info tab but I can't understand it that much. I've also read that the order of how things are laid out affects the memory usage; can anyone clarify if I'm causing issues here? Sort of a microcontroller newb. I would essentially only be having raw environment noise being processed through this audio system, but not much more than that



2) For using filters, can I route the output of BOTH Band and high pass, like I did on the "filter3Reverb" to the reverb itself? Would that double the input or blend them?



3) I have our initial input going to three mixers, one for left side processing and one for right side processing, and then a third mixer which gives a noise & reverb, which then routes back to the output mixers of Left and Right. Is this an efficient setup? Or am i just adding too many steps here?

Thank you so much for your time!

 

[el_supremo]

1) I think the memory requirement will be determined primarily by how long the delays need to be and how much memory you allocate to the granular effects. That in turn will determine whether you need a T4.1 or can get by on a T4.0.

how to assess power requirements? I know each module states it at the bottom of the info tab but I can't understand it that much.

Which module and which info tab are you talking about?

2) I've not tried that but my guess is it isn't going to do what you want. You should feed both into a mixer and combine them that way.

3) Efficient in what sense? If you need those connections, I don't see that you can make them more "efficient".

Pete

 

[MarkT]

The power requirement of the microcontroller depends on the clock speed its running at, principally,
not what code is running - you can optimize power by reducing the clock speed and using sleep
modes, but you'd have to arrange that the relevant interrupts woke it up.

For audio modules you need to ensure that when combined they use < 100% of the CPU time.
This depends on each module's cpu use and the speed of the Teensy involved. With the T4
you'd be struggling to even remotely max-out the processor in any realistic scenario unless
using a very large convolution.

Lesser Teensy's and you'd have to be more aware of the cpu use - there are functions to
measure audio lib cpu usage and memory usage you can use to instrument your setup.

 

[MarkT]

The power requirement of the microcontroller depends on the clock speed its running at, principally,
not what code is running - you can optimize power by reducing the clock speed and using sleep
modes, but you'd have to arrange that the relevant interrupts woke it up.

For audio modules you need to ensure that when combined they use < 100% of the CPU time.
This depends on each module's cpu use and the speed of the Teensy involved. With the T4
you'd be struggling to even remotely max-out the processor in any realistic scenario unless
using a very large convolution.

Lesser Teensy's and you'd have to be more aware of the cpu use - there are functions to
measure audio lib cpu usage and memory usage you can use to instrument your setup.

If you want to do heavy-lifting, the T4.x is the clear choice to use as its far faster than the others.