Trim Sample Start + Multi-channel

[d.healey]

Which mic position does the trim sample start feature read from in order to do its task? And with single stereo samples does it look at both the left and the right or just one channel?

[Christoph Hart]

To be honest, I didn't check if this feature supports multichannel at all (and if, it would use the first channel).

For stereo samples, it's supposed to look at both channels. However, looking at the code, I found this gem:

float l = analyseBuffer.getSample(0, sample);
float r = analyseBuffer.getSample(0, sample);

if (l > threshHoldLevel || r > threshHoldLevel)
{
	break;
}

Can you spot the mistake :)?

[d.healey]

For multi-channel it would be good if we could specify the channel it analyses because usually you want it to use the closest microphone to the instrument since that will have the shortest delay.

[Christoph Hart]

Yeah, good idea. It was a quick and dirty implementation for a particular use case but it may be smarter to make this more generic.

[d.healey]

I think it's a really neat feature, can allow me to be more lazy with sample editing :) Could the threshold also be adjustable or would it not add any real benefit?

[Christoph Hart]

I think currently it's hardcoded to something like 18dB below the max level, but it is certainly better if this is tweakable...

[d.healey]

Has there been any development with this feature since the last post?

[Christoph Hart]

Now there is :)

- added variable threshhold level & mic position for sample start tri… · christophhart/HISE@583b615

The open source framework for sample based instruments - - added variable threshhold level & mic position for sample start tri… · christophhart/HISE@583b615

GitHub (github.com)

[d.healey]

Excellent, thanks :)

[d.healey]

This seems to work well, is a good default -18?

[d.healey]

I might have stumbled across a bug but as yet I haven't been able to repeat it reliably. It occurs when I load some samples into a sampler, run trim start, then decide I'm not happy with the threshold I chose so I reset the start time of all the samples to 0 and run trim start again with a different threshold. What happens is some of the samples end up with their start time being set to the same as their end time.