[FFmpeg-devel] [PATCH] avfilter/af_astats: calculate audio bit-depth

[wm4]

On Tue, 14 Jul 2015 22:01:48 +0000
Paul B Mahol <onemda at gmail.com> wrote:

> Signed-off-by: Paul B Mahol <onemda at gmail.com>
> ---
>  libavfilter/af_astats.c | 21 +++++++++++++++++++--
>  1 file changed, 19 insertions(+), 2 deletions(-)
> 
> diff --git a/libavfilter/af_astats.c b/libavfilter/af_astats.c
> index a89ac6d..21cec8c 100644
> --- a/libavfilter/af_astats.c
> +++ b/libavfilter/af_astats.c
> @@ -33,6 +33,7 @@ typedef struct ChannelStats {
>      double min, max;
>      double min_run, max_run;
>      double min_runs, max_runs;
> +    uint64_t mask;
>      uint64_t min_count, max_count;
>      uint64_t nb_samples;
>  } ChannelStats;
> @@ -121,6 +122,15 @@ static int config_output(AVFilterLink *outlink)
>      return 0;
>  }
>  
> +static unsigned bit_depth(uint64_t mask)
> +{
> +    unsigned result = 64;
> +
> +    for (; result && !(mask & 1); --result, mask >>= 1);
> +
> +    return result;
> +}
> +
>  static inline void update_stat(AudioStatsContext *s, ChannelStats *p, double d)
>  {
>      if (d < p->min) {
> @@ -151,6 +161,7 @@ static inline void update_stat(AudioStatsContext *s, ChannelStats *p, double d)
>      p->sigma_x2 += d * d;
>      p->avg_sigma_x2 = p->avg_sigma_x2 * s->mult + (1.0 - s->mult) * d * d;
>      p->last = d;
> +    p->mask |= llrint(d * (1LLU<<63));
>  
>      if (p->nb_samples >= s->tc_samples) {
>          p->max_sigma_x2 = FFMAX(p->max_sigma_x2, p->avg_sigma_x2);
> @@ -177,7 +188,7 @@ static void set_meta(AVDictionary **metadata, int chan, const char *key,
>  
>  static void set_metadata(AudioStatsContext *s, AVDictionary **metadata)
>  {
> -    uint64_t min_count = 0, max_count = 0, nb_samples = 0;
> +    uint64_t mask = 0, min_count = 0, max_count = 0, nb_samples = 0;
>      double min_runs = 0, max_runs = 0,
>             min = DBL_MAX, max = DBL_MIN,
>             max_sigma_x = 0,
> @@ -203,6 +214,7 @@ static void set_metadata(AudioStatsContext *s, AVDictionary **metadata)
>          max_count += p->max_count;
>          min_runs += p->min_runs;
>          max_runs += p->max_runs;
> +        mask |= p->mask;
>          nb_samples += p->nb_samples;
>          if (fabs(p->sigma_x) > fabs(max_sigma_x))
>              max_sigma_x = p->sigma_x;
> @@ -217,6 +229,7 @@ static void set_metadata(AudioStatsContext *s, AVDictionary **metadata)
>          set_meta(metadata, c + 1, "Crest_factor", "%f", p->sigma_x2 ? FFMAX(-p->min, p->max) / sqrt(p->sigma_x2 / p->nb_samples) : 1);
>          set_meta(metadata, c + 1, "Flat_factor", "%f", LINEAR_TO_DB((p->min_runs + p->max_runs) / (p->min_count + p->max_count)));
>          set_meta(metadata, c + 1, "Peak_count", "%f", (float)(p->min_count + p->max_count));
> +        set_meta(metadata, c + 1, "Bit-depth", "%f", bit_depth(p->mask));
>      }
>  
>      set_meta(metadata, 0, "Overall.DC_offset", "%f", max_sigma_x / (nb_samples / s->nb_channels));
> @@ -228,6 +241,7 @@ static void set_metadata(AudioStatsContext *s, AVDictionary **metadata)
>      set_meta(metadata, 0, "Overall.RMS_trough", "%f", LINEAR_TO_DB(sqrt(min_sigma_x2)));
>      set_meta(metadata, 0, "Overall.Flat_factor", "%f", LINEAR_TO_DB((min_runs + max_runs) / (min_count + max_count)));
>      set_meta(metadata, 0, "Overall.Peak_count", "%f", (float)(min_count + max_count) / (double)s->nb_channels);
> +    set_meta(metadata, 0, "Overall.Bit-depth", "%f", bit_depth(mask));
>      set_meta(metadata, 0, "Overall.Number_of_samples", "%f", nb_samples / s->nb_channels);

So why would everything use _ as separate, except bit depth?

>  }
>  
> @@ -276,7 +290,7 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *buf)
>  static void print_stats(AVFilterContext *ctx)
>  {
>      AudioStatsContext *s = ctx->priv;
> -    uint64_t min_count = 0, max_count = 0, nb_samples = 0;
> +    uint64_t mask = 0, min_count = 0, max_count = 0, nb_samples = 0;
>      double min_runs = 0, max_runs = 0,
>             min = DBL_MAX, max = DBL_MIN,
>             max_sigma_x = 0,
> @@ -302,6 +316,7 @@ static void print_stats(AVFilterContext *ctx)
>          max_count += p->max_count;
>          min_runs += p->min_runs;
>          max_runs += p->max_runs;
> +        mask |= p->mask;
>          nb_samples += p->nb_samples;
>          if (fabs(p->sigma_x) > fabs(max_sigma_x))
>              max_sigma_x = p->sigma_x;
> @@ -318,6 +333,7 @@ static void print_stats(AVFilterContext *ctx)
>          av_log(ctx, AV_LOG_INFO, "Crest factor: %f\n", p->sigma_x2 ? FFMAX(-p->min, p->max) / sqrt(p->sigma_x2 / p->nb_samples) : 1);
>          av_log(ctx, AV_LOG_INFO, "Flat factor: %f\n", LINEAR_TO_DB((p->min_runs + p->max_runs) / (p->min_count + p->max_count)));
>          av_log(ctx, AV_LOG_INFO, "Peak count: %"PRId64"\n", p->min_count + p->max_count);
> +        av_log(ctx, AV_LOG_INFO, "Bit-depth: %u\n", bit_depth(p->mask));
>      }
>  
>      av_log(ctx, AV_LOG_INFO, "Overall\n");
> @@ -331,6 +347,7 @@ static void print_stats(AVFilterContext *ctx)
>          av_log(ctx, AV_LOG_INFO, "RMS trough dB: %f\n", LINEAR_TO_DB(sqrt(min_sigma_x2)));
>      av_log(ctx, AV_LOG_INFO, "Flat factor: %f\n", LINEAR_TO_DB((min_runs + max_runs) / (min_count + max_count)));
>      av_log(ctx, AV_LOG_INFO, "Peak count: %f\n", (min_count + max_count) / (double)s->nb_channels);
> +    av_log(ctx, AV_LOG_INFO, "Bit-depth: %u\n", bit_depth(mask));
>      av_log(ctx, AV_LOG_INFO, "Number of samples: %"PRId64"\n", nb_samples / s->nb_channels);
>  }
>