Re: [Hampshire] Video editing - 1st steps

> DV format instead of MPEG2 would have been better.
> When you edit MPEG2 you loose a bit of quality.
> When you edit DV, you do not loose any quality.

This is not the case.

Both DV and MPEG-2 are DCT-based lossy compressors - they *both* lose
quality compared to the uncompressed video.

MPEG-2 uses both intra-frame and inter-frame compression, giving it a
better compression rate for the same quality than DV, which uses
intra-frame compression only.

DV video is often recorded at a higher bitrate than MPEG-2, which means
that it will often be less compressed than an MPEG-2 stream - but this is
not some inherent fault in the compressor, it's just that the MPEG-2
copmpression was set to be lossier. The maximum bitrate that can be
supported by an MPEG-2 stream is significantly higher than that supported
by DV. Coupled with the inter-frame compression mentioned earlier, this
means that the best quality available from an MPEG-2 system significantly
exceeds the best quality available from a DV system.

A side-effect of inter-frame compression is that sequences are generated -
called Groups of Pictures (GOPs). Each GOP contains the first image of the
sequence, the last image of the sequence (which is compressed by a
unidirectional prediction), and all the intermediate inages (which are
compressed with bidirectional prediction). Historically, some editing
software has struggled with putting edits inside a GOP - again, this is
not a fault of the compressor, it's down to how the editing software has
been written. Splitting GOPs isn't pretty, but is perfectly possible.
Given sufficient resolution in the calculation software, any additional
quantisation noise is negligible, except in the pathological case of
synthetic edges that weren't split into separate GOPs by the original
compressor.

Vic.