On 12/11/12 7:57 PM, Justin Ruggles wrote:
---
Hi,
This patch does a few things differently the previous patches.
For simple rectangular and triangular dither without filtering, the
conversion of LFG values to float is done as SIMD along with the
sample conversion.
I added SIMD for separate conversion of LFG values to float.
I added SIMD for sample conversion for use with the triangular dither
with high-pass filter.
I included the input scaling with the conversion functions. Having it
separate is not much advantage since the conversions use SIMD. This
also avoids copying the whole frame if the input is not writable.
The C part looks fine to me.
_______________________________________________
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
