Jan Stary <h...@stare.cz> writes:
> Yes, this diff fixes it, and looking at sox -V4
> it does the right thing:
>
> $ sox -c 1 -b 16 -r 1000 -n file.wav synth 1 sin
> $ sox -V4 file.wav -n spectrogram -x 5000
>
> [...]
> sox DBUG spectrogram: duration=1 x_size=5000 pixels_per_sec=5000 dft_size=1024
> sox DBUG spectrogram: window_density=0.5
> sox INFO spectrogram: actual pixels/s = 1000
> sox DBUG spectrogram: step_size=1 block_steps=1
> [...]
> sox DBUG sox: start-up time = 0.001902
> sox DBUG spectrogram: cols=489 left=0 end=24
> sox DBUG spectrogram: flushed cols=1001 left=0 end=-512
> sox DBUG spectrogram: signal-max=-1.1134
>
> Thank you!
Thanks for testing. I've pushed the fix to the SF repo.
> Jan
>
> diff --git a/src/spectrogram.c b/src/spectrogram.c
> index 3dcda69c..89ae1fbf 100644
> --- a/src/spectrogram.c
> +++ b/src/spectrogram.c
> @@ -431,7 +431,7 @@ static int start(sox_effect_t *effp)
> actual = make_window(p, p->last_end = 0);
> lsx_debug("window_density=%g", actual / p->dft_size);
> p->step_size = (p->slack_overlap ? sqrt(actual * p->dft_size) : actual) +
> 0.5;
> - p->block_steps = effp->in_signal.rate / pixels_per_sec;
> + p->block_steps = max(effp->in_signal.rate / pixels_per_sec, 1);
> p->step_size =
> p->block_steps / ceil((double)p->block_steps / p->step_size) + 0.5;
> p->block_steps = floor((double)p->block_steps / p->step_size + 0.5);
>
--
Måns Rullgård
_______________________________________________
SoX-devel mailing list
SoX-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sox-devel
