First off, FFmpeg’s Interplay MVE decoder finally supports 16-bit video mode in addition to the original palette mode. I would post a pretty picture showcasing this (as I often do for new game-related formats) but Kostya has already done the honors.
I added some new FATE tests today based on systems that have recently appeared in FFmpeg:
Nice work :)
Please just do not forget also the less used cases of MPEG2 and H264 that I added to the TODO section of the FATE wiki page.
There seems to be a small discrepancy with the cdgraphics test: the test spec refers to $SAMPLES_PATH/cdg/BrotherJohn.cdg while the collection on mphq has fate-suite/cdgraphics/BrotherJohn.cdg.
Thanks. I just updated the test spec to be cdgraphics rather than cdg (and changed my local path).
Isn’t there an official test-suite for als?
That’s a very good question, Carl Eugen. MPEG codecs usually do have suites and this would be very straightforward to test. I am unaware of such a suite, however.
I found the following in http://elvera.nue.tu-berlin.de/files/0782Liebchen2004.pdf
The test material was taken from the standard audio sequences for
MPEG-4 Lossless Coding. It comprises almost 1 GB of stereo
waveform data with sampling rates of 48, 96, and 192 kHz, and
resolutions of 16 and 24 bits.
(Note that the paper states that Monkey’s Audio Codec is better for lower sampling rates.)
Is that a gigabyte before or after compression? The nice thing about testing lossless coding is that I don’t need to store the uncompressed waveforms (like I will need to with my 1-off testing idea, if I ever get it deployed).