Is SheerVideo more accurate than Motion JPEG?
Yes. SheerVideo is a perfect-fidelity codec, whereas Motion JPEG is a DCT-based approximating codec. For real-world RGB 8b content, Apple’s excellent Motion JPEG A and B codecs at their Best quality setting have an RMS (root mean square) error of 1.21 b/S (bits per sample), while SheerVideo has an error of 0.00 b/S. For Y’CbCr 8bv 4:4:4, Motion JPEG A and B’s RMS error is 0.80 b/S, against SheerVideo’s error of 0.00 b/S. Likewise, for Y’CbCr 8bv 4:2:2, M-JPEG A and B have an RMS error of 0.38 b/S, compared to SheerVideo’s error of 0.00 b/S. In spite of its lower error rate, SheerVideo compresses 3 times as fast as M-JPEG, and decompresses twice as fast as M-JPEG. Note also that SheerVideo supports alpha compression, which Motion JPEG does not. On the other hand, in situations where video file size is more important than quality and speed, Motion JPEG can compress far more than any perfect-fidelity codec.
