[FFmpeg-devel] [PATCH 3/4] Add setpts filter by Victor Paesa.

[Stefano Sabatini]

On date Tuesday 2010-11-02 17:37:42 -0400, Alexander Strange encoded:
> 
> On Nov 2, 2010, at 5:33 PM, Stefano Sabatini wrote:
> 
> > On date Tuesday 2010-11-02 22:16:30 +0100, Michael Niedermayer encoded:
> >> On Tue, Nov 02, 2010 at 08:45:07PM +0100, Stefano Sabatini wrote:
> >> [...]
> >>> +/**
> >>> + * @file
> >>> + * video presentation timestamp (PTS) modification filter
> >>> + */
> >>> +
> >> 
> >>> +#define DEBUG
> >> 
> >> ?
> > 
> > Commented out, I'd like to keep it in case I need to debug, but I
> > don't want to spam the log everytime I enable -loglevel debug.
> > 
> >>> +
> >>> +#include "libavutil/eval.h"
> >>> +#include "avfilter.h"
> >>> +
> >>> +static const char *var_names[] = {
> >>> +    "E",           ///< Euler number
> >>> +    "INTERLACED",  ///< tell if the current frame is interlaced
> >>> +    "N",           ///< frame number (starting at zero)
> >>> +    "PHI",         ///< golden ratio
> >>> +    "PI",          ///< greek pi
> >>> +    "POS",         ///< original position in the file of the frame
> >>> +    "PREV_INPTS",  ///< previous  input PTS
> >>> +    "PREV_OUTPTS", ///< previous output PTS
> >>> +    "PTS",         ///< original pts in the file of the frame
> >>> +    "STARTPTS",   ///< PTS at start of movie
> >>> +    "TB",          ///< timebase
> >>> +    NULL
> >>> +};
> >>> +
> >>> +enum var_name {
> >>> +    VAR_E,
> >>> +    VAR_INTERLACED,
> >>> +    VAR_N,
> >>> +    VAR_PHI,
> >>> +    VAR_PI,
> >>> +    VAR_POS,
> >>> +    VAR_PREV_INPTS,
> >>> +    VAR_PREV_OUTPTS,
> >>> +    VAR_PTS,
> >>> +    VAR_STARTPTS,
> >>> +    VAR_TB,
> >>> +    VAR_VARS_NB
> >>> +};
> >>> +
> >>> +typedef struct {
> >>> +    AVExpr *expr;
> >>> +    double var_values[VAR_VARS_NB];
> >>> +} SetPTSContext;
> >>> +
> >>> +static av_cold int init(AVFilterContext *ctx, const char *args, void *opaque)
> >>> +{
> >>> +    SetPTSContext *setpts = ctx->priv;
> >>> +    int ret;
> >>> +
> >>> +    if ((ret = av_parse_expr(&setpts->expr, args ? args : "PTS",
> >>> +                             var_names, NULL, NULL, NULL, NULL, 0, ctx)) < 0) {
> >>> +        av_log(ctx, AV_LOG_ERROR, "Error while parsing expression '%s'\n", args);
> >>> +        return ret;
> >>> +    }
> >>> +
> >>> +    setpts->var_values[VAR_E          ] = M_E;
> >>> +    setpts->var_values[VAR_N          ] = 0.0;
> >>> +    setpts->var_values[VAR_PHI        ] = M_PHI;
> >>> +    setpts->var_values[VAR_PI         ] = M_PI;
> >>> +    setpts->var_values[VAR_PREV_INPTS ] = NAN;
> >>> +    setpts->var_values[VAR_PREV_OUTPTS] = NAN;
> >>> +    setpts->var_values[VAR_STARTPTS   ] = NAN;
> >>> +    return 0;
> >>> +}
> >>> +
> >> 
> >>> +static int config_input(AVFilterLink *inlink)
> >>> +{
> >>> +    SetPTSContext *setpts = inlink->dst->priv;
> >>> +
> >>> +    setpts->var_values[VAR_TB]    = (double)inlink->time_base.num/inlink->time_base.den;
> >> 
> >> av_q2d
> >> 
> >> otherwise lgtm
> > 
> > Fixed and updated.
> 
> Whatever does M_PHI have to do with pts?

Solution in search of a creative problem, and the same could be said
to E and PI, but PHI given its nature seems to fit better than them.
-- 
FFmpeg = Formidable and Fantastic Meaningless Proud Elegant Gymnast