Tweet
MPEG1/2 comparison
Hi, all!
FWIW, I've just done a stringent comparison between MPEG1 and 2, as generated with TMPGEnc v. 1.12c.
For input, I took an AVI (MJPEG) as generated by Marvel and edited with MSP6. This is PAL 704x576 25 fps, from a mini-DV camera, S-video.
In both cases, I used the default TMPGEnc settings of CBR 1150 kbs, sound layer 2 44100 Hz (this is actually higher than captured at 22050 mono)192 kbps.
I played the 2 files back side-by-side on a 21" monitor, using MS MP 6.01, so I could get a direct and simultaneous comparison. There were no hitches with either playback, despite the load of the two files.
The video playback codecs were the standard MP one for MPEG1 and the Ligos one for MPEG2, set to Fast mode.
I observed the following:
1) the colour of the MPEG1 file was slightly greener, most noticeable on a clear sky in a few shots. I judge this colour shift not to be severe for ordinary viewing. It was not sufficient to significantly affect flesh tones. Comparison with the MJPEG original, the MPEG2 was slightly closer, although both showed a slight shift in opposite directions.
2) the artefacts were more noticeable with rapidly moving objects with the MPEG1 file. Particularly on vertical edges, aa "tearing" effect was noticed. With MPEG2, this was reduced to a slight "fuzziness", compared to the original.
3) no difference was observed with the sound, either between them or the original. There was no sync. offset with either
4) There was a slight change in frame rate and frame time jitter. The MPEG1 was converted to average 24.98 fps with zero jitter and the MPEG2 to average 24.72 fps (average frame length 40.45 ms) with an SD jitter of 3 ms which is higher than I would have expected. This did not convert to anything visible, that I could observe, so is really academic.
------------------
Brian (the terrible)
FWIW, I've just done a stringent comparison between MPEG1 and 2, as generated with TMPGEnc v. 1.12c.
For input, I took an AVI (MJPEG) as generated by Marvel and edited with MSP6. This is PAL 704x576 25 fps, from a mini-DV camera, S-video.
In both cases, I used the default TMPGEnc settings of CBR 1150 kbs, sound layer 2 44100 Hz (this is actually higher than captured at 22050 mono)192 kbps.
I played the 2 files back side-by-side on a 21" monitor, using MS MP 6.01, so I could get a direct and simultaneous comparison. There were no hitches with either playback, despite the load of the two files.
The video playback codecs were the standard MP one for MPEG1 and the Ligos one for MPEG2, set to Fast mode.
I observed the following:
1) the colour of the MPEG1 file was slightly greener, most noticeable on a clear sky in a few shots. I judge this colour shift not to be severe for ordinary viewing. It was not sufficient to significantly affect flesh tones. Comparison with the MJPEG original, the MPEG2 was slightly closer, although both showed a slight shift in opposite directions.
2) the artefacts were more noticeable with rapidly moving objects with the MPEG1 file. Particularly on vertical edges, aa "tearing" effect was noticed. With MPEG2, this was reduced to a slight "fuzziness", compared to the original.
3) no difference was observed with the sound, either between them or the original. There was no sync. offset with either
4) There was a slight change in frame rate and frame time jitter. The MPEG1 was converted to average 24.98 fps with zero jitter and the MPEG2 to average 24.72 fps (average frame length 40.45 ms) with an SD jitter of 3 ms which is higher than I would have expected. This did not convert to anything visible, that I could observe, so is really academic.
------------------
Brian (the terrible)
Comment