Measure input audio volume

addr3ss

How can I measure input audio volume?
I want to detect when input audio volume is lower than some level and do something.
There is no problem if volume unit is decibel or other things.

Kent-Dorfman

"lower" is harder than "higher" because with higher you just need to read the pcm amplitude values and compare to a high limit. Lower would involve an algorithm that looks at mean average amplitude and decides.

addr3ss

@Kent-Dorfman Ok.How can I get input volume and compare it to the high level?

Kent-Dorfman

https://www.springer.com/us/book/9783030117801?gclid=EAIaIQobChMI2ZfKkcma6AIVlpOzCh1mTQGgEAQYASABEgK3ofD_BwE

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=10&ved=2ahUKEwi-ksORyZroAhVtj3IEHWg9BfEQFjAJegQICRAB&url=http%3A%2F%2Fcfile21.uf.tistory.com%2Fattach%2F181FA6174A8C7AD652D80D&usg=AOvVaw1BML3PHYHummXEtx7IvMv8

Read it
Learn it
Live it

addr3ss

I dont want to be an audio processing expert.I just want some qt c++ code.

mrjj

Hi
Maybe this example can be used
https://doc.qt.io/qt-5/qtmultimedia-multimedia-spectrum-example.html
as a base for reading the data and then apply some algorithm to get what you want.

addr3ss

I found this. A simple example:
https://github.com/soramimi/QtAudioInput

addr3ss

I wrote this code but it does not detect continuously. It just detect the level at the beginning and after, the level begins to reduce.How can I solve this?

#include "Bot.h"

void Bot::start() {
  format.setByteOrder(QAudioFormat::LittleEndian);
  format.setChannelCount(2);
  format.setCodec("audio/pcm");
  format.setSampleRate(8000);
  format.setSampleSize(16);
  format.setSampleType(QAudioFormat::SignedInt);
  source = std::shared_ptr<QAudioInput>(new QAudioInput(format));
  device = source->start();
  connect(device, SIGNAL(readyRead()), this, SLOT(onReadyRead()));
}

void Bot::stop() {
  source = nullptr;
  device = nullptr;
  peak = 0;
  emit volumeChanged(peak);
}

void Bot::onReadyRead() {
  peak = 0;
  n = source->bytesReady();
  n /= 2;
  if (n >= 80) {
    for (int i = 0; i < n; i++) {
      v = 0;
      device->read((char *)&v, 2);
      if (v < 0) {
        v = -v;
      }
      if (peak < v) {
        peak = v;
      }
    }
    peak = peak * 100 / 32768;
    if (peak < 0) {
      peak = 0;
    }
    emit volumeChanged(peak);
  }
}

Kent-Dorfman

@addr3ss said in Measure input audio volume:

> void Bot::onReadyRead() {
>   peak = 0;
>   n = source->bytesReady();
>   n /= 2;
>   if (n >= 80) {
>     for (int i = 0; i < n; i++) {
>       v = 0;
>       device->read((char *)&v, 2);
>       if (v < 0) {
>         v = -v;
>       }
>       if (peak < v) {
>         peak = v;
>       }
>     }

Two things:

for the type of programming you are doing, truncation, overflow, and signedness will bite you in the hind quarters. Please explicitly type your vars.
please comment your algorithms. I can assume that n/2 is accounting for stereo samples, but I should not have to make assumptions about your algorithm. Explaining it to others often causes YOU to see the error in your process.

I still say you need to do some real research into digital sound processing, and that the "I don't want to be an expert" is why there is so much crap software out there.

Good luck!

addr3ss

I typed my vars in header file:

  Q_INVOKABLE void start();
  Q_INVOKABLE void stop();
  std::shared_ptr<QAudioInput> source;
  QIODevice *device;
  QAudioFormat format;
  int peak = 0, n;
  int16_t v;

And I copied this code from https://github.com/soramimi/QtAudioInput with some edit.
So actually I dont know how it works. But I hope someone can understand the code and help me.