Hi guys,

My question concerns DV avi. Does the (mini) DV camera
use some kind of compression when recording to its tape, or, rather, the capturing software does it, while the camera stores data in raw format?

I have read somewhere that the best practice is to edit the DV avi file and later encode it as mpeg 2, etc. and
feed back the avi to the camera to store the clip in its original format. If the camera stores raw data then there is no sense in this procedure!

Cheers,
Laszlo

A miniDV camera records to DV25 tape. In doing so the signal is compressed by a hardware DV codec, so the resulting footage on the tape is 1:5 compressed.

When the footage is tranfered to computer harddrive by firewire, the process is essentially lossless, with no futher compression occurring. (There is also no loss when shifting a DV avi from the computer back to the camera's tape.)

While the footage is within the computer (as a DV avi file) it can be viewed and edited and this process utilizes a software DV codec (usually the one that is supplied within DirectX - there are others for example by Mainconcept). Simple cuts to footage do not cause re-rendering but fades, color-correction and transitions and titles do. Rerendering involves a decompression/recompression cycle and so loses some quality.

Converting the 1:5 DV avi to MPEG2 involves a large amount of further recompression and loss of quality - depending on the bitrate you use (e.g. 2300kbps for SVCD or up to 9000 for DVD). The resulting MPEG will be only about 1/10th the size of the avi.

So what the advice you got was getting at was to do all your editing while the file is still in avi format. Compress the final edit to an MPEG for your DVD or SVCD, and in addition store your edited DV-avi footage back to tape via the camera. That way you retain the best possible master footage in case you want to re-edit later and/or recompress using some other codecs (e.g. DiVX;).

Coby,

Thank you for your clear answer.

I am doing exactly what you described. I create a title, arrange clips, add some transitions and sound sometimes and render the whole project in DV avi (I think the procedure which renders only the changed parts is called as "smart rendering"). I feed back the rendered DV avi to a master tape in the camera and use mpeg2 encoders to produce, mostly, SVCD. I was not sure, however, that there was a general quality loss beyond what was inevitable due to editing, i.e. the hardware and sofware was identical or different. Now I know they are the same.

Thanks again for your valuable help,
Laszlo