mirror of https://github.com/AlfredoSequeida/fvid
e8362ec7e6
This results in a nearly 20% speedup for the get_bits_from_image function. Cython is now only 40% faster than Python, and FFmpeg is now the major bottleneck in code. I have reason to believe this function is close to maximal optimization, unless it is rewritten to scan multiple pixels at a time (which might not work because bits need to be in sequential order) or use a different method to load pixels. |
||
---|---|---|
.. | ||
__init__.py | ||
__main__.py | ||
cythonizer.py | ||
fvid.py | ||
fvid_cython.c | ||
fvid_cython.pyx |