On Mon, Jun 02, 2014 at 09:12:23AM -0400, Justin Ruggles wrote: > On 06/02/2014 08:53 AM, Diego Biurrun wrote: > > On Mon, Jun 02, 2014 at 08:30:42AM -0400, Justin Ruggles wrote: > >> On 06/02/2014 07:34 AM, Diego Biurrun wrote: > >>> -void ff_vector_clipf_neon(float *dst, const float *src, float min, float > >>> max, > >>> - int len); > >> > >> Might be a better fit in floatdsp? > > > > Hmm, maybe - do you refer to just vector_clipf or all the functions > > I'm splitting off? > > just that one. none of the others use floats...
I tried implementing that, but then stumbled across the issue that clip_coefficients() in libavcodec/ac3enc_template.c expects some sort of DSP context as first parameter. If vector_clipf() is in another context than vector_clip_int32(), then I get a conflict in the clip_coefficients prototype. What do you suggest? Use a void* in the clip_coefficients prototype, a union, something else? Diego _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel