Potential underflow / overflow when converting from float to int16 #209

Meoo · 2022-08-10T16:21:20Z

In the example code "rnnoise_demo", the float output is converted to a short with a simple cast

I found by experimentation that the output can be outside of int16 range, so the output should be clamped to prevent clicky noises when the input has values near the int16's min or max

lufengjie051523 · 2023-02-20T02:07:41Z

Hi,did you solved the problem?

lguarda · 2025-05-01T10:56:39Z

Hello indeed clicky noise seems to be prevented by using something like this

void conversion(int16_t *data, float *buffer) {
    for(int j = 0; j < 480; j++) {

        if (buffer[j] > INT16_MAX) {
            data[j] = INT16_MAX;

        } else if (buffer[j] < INT16_MIN) {
            data[j] = INT16_MIN;

        } else {
            data[j] = buffer[j];
        }
    }
}

But i'm afraid that it's could prevent some compiler optimization,
(Actually using compiler explorer it seems that there's still some intrinsic instruction used so it should be fine)
It still doesn't seems right to do
if someone have a better solution, it could be nice (also i'm not even sure this is always the best thing to do for any cases)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Potential underflow / overflow when converting from float to int16 #209

Potential underflow / overflow when converting from float to int16 #209

Potential underflow / overflow when converting from float to int16 #209

Potential underflow / overflow when converting from float to int16 #209

Comments