Sorry for this, but I think I might be missing something here. I want to make a patch that will alter a signals dynamics, and in this case I think it would make more sense to work on a sample by sample basis.

So, is my understanding of how digital audio works in Max correct?

- signals are represented by a series of samples (44,100 a second, obv.).

- The further a sample is from 0, the higher it's amplitude.

- a sample at 0.5 has the same amplitude of a sample at -0.5

Let me know if I'm missing something heree.

- signals are represented by a series of samples (44,100 a second, obv.).
- Each sample is a value from -1 to 1.
- The further a sample is from 0, the higher it's amplitude.
- a sample at 0.5 has the same amplitude of a sample at -0.5

simple-digital-audio-concepts

See the MSP tutorial files "How Digital Audio Works", "How MSP Works", and "MSP Tutorial 2 - Adjustable Oscillator".

2. no, they are float32 and can exceed the range of -1 to 1

3. yes, but amplitude is not relevant to dynamics processing, you need to deal with power here,

and power is not present for only one sample, you´d need two or more.

dynamics processing sample by sample ... not at all. classic concepts would use 40-200 milliseconds.

1. yes
2. no, they are float32 and can exceed the range of -1 to 1
3. yes, but amplitude is not relevant to dynamics processing, you need to deal with power here,
and power is not present for only one sample, you´d need two or more.
4. see 3., otherwise "yes".

dynamics processing sample by sample ... not at all. classic concepts would use 40-200 milliseconds.


Ah, okay...I've been trying to use an avg~ object to get a volume measurement, before I thought about trying it this way. Not sure if that's where the issues I'm having lie or not.

And yeah, I'll reread those MSP tutorials. I was just looking at the numbered ones, and wasnt finding much helpful information.

EDIT: question...So, audio samples can exceed the -1 to 1 limit, true. But doesn't any output that exceeds that range get clipped?

A Signal beyond the -1 and 1 range gets clipped only when it is sent to the DAC. Before that, there is no clipping, just the maximum value of a 32 bit float (so before going to the dac, an amplitude of 4 billion would be fine).

Simple digital audio concepts