analog to mpeg2, bitrate and vbr vs cbr

[codemaster]

Before I will capture from SD analog source, to mpeg2, in CyberLink PowerDirector I will have to choose the CBR or VBR mode, the bitrate for CBR mode or the min and max bitrate for VBR.

I will encode the captured mpeg2 file to h264, using avisynth and the x264 encoder. Therefore, I think it will be a good idea to choose a higher bitrate for capturing to mpeg2, since the mpeg2 file will not be used on a standalone dvd player. But I don't know what would be the treshhold bitrate, above which any higher bitrates will not have a significant impact on maintaining a visual quality as close as possible to the source analog signal.

For video capture, using the mpeg2 codec, which mode is more suitable: CBR or VBR ? If VBR, then how do I determine the min bitrate and the max bitrate ? Or what bitrates and modes would be recommended by compression specialists ?

[DB83]

Choice of bitrate can depend on the type of footage you capture. Little movement can be encoded at a low bit rate whereas something with fast movement requires more. Even so, I would not go below a minimum of 3500 kbps

Personally, I always capture at constant bitrate since this puts less strain on the hardware/software.

But......

If you intend to go the h264 why bother with mpeg2 (if a dvd source I can understand). If you can, capture lossless to lagarith.

[2Bdecided]

If the final file will be something else, and your only option is to capture to MPEG-2, use CBR max bitrate (usually 15Mbps). If lossless is available to you as a direct capture option, use that instead.

[jman98]

The point of VBR is to try to squeeze somewhat higher quality into a small space, like a CD or DVD. For example, if you have 45 minutes of video and want to put it on a single layer DVD disc and make a standard DVD out of it, you can use CBR at the highest bit rate possible for DVD (including the bit rate of audio in your calculations) and you still won't use all the space on the disc. VBR encodes some parts of the video at lower bit rates than the average and some parts at higher than the average and the idea is that you'll get slightly better quality because you won't notice the lower bit rate parts looking worse and the higher bit rate parts may look better than if you just used a constant bit rate on everything. Since you have no worries about space as your goal is to re-encode to H.264, follow 2Bdecided's instructions.

[Nelson37]

Since you are re-encoding, size considerations go out the window. Crank it up as high as it will go without causing dropped frames. If you can, use I-Frame-only. Stay with CBR.

A bitrate of 15,000 for SD video is usually considered transparent, but if you can go higher, do it. It's a temporary file.

IF you can do uncompressed (Huffy or Lagarith), do it. Be aware that the massive filesize may cause slight playback problems that will not be present in the re-encoded file.

[PuzZLeR]

MPEG-2 capture alone can be demanding on resources, and sounds furthermore demanding with VBR, like DB83 pointed out.

As for what bitrate, or making it more efficient with VBR, this can also depend on another factor - are you keeping the capture Source after you encode it with x264?

If not, then I say go as HIGH a bitrate as possible, and CBR, etc, - heck even I would say to go lossless to get even more quality out of the Source capture.

If you do wish to keep, and archive, the capture Source (recommended), unless you don't mind capturing twice (one for archival and one for your encodes), or don't mind accommodating the extra space, and your capture setup supports it, then a good compromise for archival SD MPEG-2 for size and quality would then be VBR between 10mbps-15mpbs and interlaced.