ΣSigmaStudio Ask AI Русский Log in

Dynamic Range Question

SigmaDSP & SigmaStudio mdot Jun 14, 2011 00:00 5 replies Original on forum ↗
Subscribe
mdot Question
Jun 14, 2011 00:00

Hi All,

I have a quick question that I have been unable to really find an answer to.

Is there a way to measure/detect perceived volume levels? What I mean is, we all know that music, movies and TV use a lot of compression on their audio, which increases the perceived volume, while not actually raising the output level. I'd like to try and measure this.

I have rumintated on trying to use timers to detect when the peak value of an audio signal is not varying much (decreased dynamic range). Or, maybe split out several different frequency ranges, measure their individual levels and if they are all within a certain amount of each other, it's probably because of decreasd dynamic range.

Is something like this even possible programmatically? Any information/suggestions would be greatly appreciated.

Shahin Reply
Jun 14, 2011 00:00

I believe what you want to be doing is comparing the peak to average level of the signal. The smaller the ratio (i.e. the closer the average is to the peak), the more compressed the content is.

mdot Reply
Jun 14, 2011 00:00

That's interesting, I hadn't thought about it that way.

I'm going to try that out. Thanks.

mdot Reply
Jun 15, 2011 00:00

Just wanted to update, in case anybody stumbles across the thread in the future.

The method Shahin is definitely on the right track. Because of the nature of voice audio, it's a bit tricky to setup a flow that can tell the difference between voice, where compression is not as much of an issue and more complex audio signals.

The direction I'm heading in now is trying to design a signal flow that makes use of the information from the ITU-R 468 noise weighting graph.

I'll also be starting a discussion about setting the bandwidth of a bandpass filter to try and apply some weighting to the incoming signal, then compare the peak of the filtered signal to the average level. I think that this will yield the results I'm looking for.

ADIApproved Reply
Jun 16, 2011 00:00

Thanks, mdot. Keep us posted.

I think the idea of pre-weighting your signal is a good approach here.

EZAdministrator Reply
Aug 2, 2018 00:00

This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

Thank you,
EZ Admin