Preserving vhs quality

[dcp]

I have two problems capturing from a vhs source. Both concerning quality:

--The first is to do with optimal video resolution. Is it true that a final (after all filtering, applying codec etc) video size of 352 x 288 (PAL) is okay and anything above would be unnecessary, since this is the approximate resolution of vhs? I want to be able to watch the final video at a size of at least about 600 x 480, should it be saved in that size or is it adequate to resize the media player window instead, in terms of quality?

--The second is the format/codec to use. With divx and mpeg there apears to be significant quality loss, when viewing the video at a high resolution. This occurs whether it is saved in a high resolution or just resized to a high resolution if it is saved in low res. I need to save it with minimal megabyte size, divx or mpeg would be perfect and I have seen video in both formats with excellent quality - eg if its taken from a digital video camera.

[jggimi]

in re: your resolution question:

FAQ Q9 discusses capture resolution very briefly. A more detailed discussion can be found as Chapter 4 of the Capture Guide, and, for a more technical discussion, all of Chapter 5.

in re: final format.

You don't have a question, but you are sharing that you get poor results with your final format.

Unfortunately, without knowing in detail your procedures, tools, tool configurations ... there is no way anyone can help you determine what your difficulty is, and what can be done to alleviate it. Here are some examples of the kinds of things that would be helpful to anyone who might be able to help:

Your capture hardware, drivers, and software. Settings used in that software.

Modifications made to the video stream prior to input to an encoder, such as cropping, resizing -- and which resizer you select -- noise reduction, deinterlacing ... the tools you use to make these changes.

And, of course, the encoder you are using, and the codec settings used during encoding.

That kind of information sharing would be a good start towards getting you the help you are looking for.

[FlimsyFeet]

--The first is to do with optimal video resolution. Is it true that a final (after all filtering, applying codec etc) video size of 352 x 288 (PAL) is okay and anything above would be unnecessary, since this is the approximate resolution of vhs?
This is not quite true. The fact that VHS has an equivalent horizontal resolution of 300 and something is I think correct, there's all that discussion on bandwidth and nyquist that the video experts talk about in the analogue capture guide.

However, PAL has 576 lines, and that is fixed. If you are downsizing to 288, then you lose half the detail.

A good output format would therefore be - I forget the name, half-D1 or something - which is I think 352 x 576. Anyway, it is apparently DVD compliant. And you don't need to deinterlace, which you would if converting to XviD.

[dosdan]

I do a lot of PAL VHS to PAL DVD transfers via miniDV ADC. Half-D1 PAL works fine. I do the NR using Convolution3D on the individual fields in the 688x540 frame (after edge cropping) and weave them back together before doing a fast HorizontalReduceBy2 and then ADDBorders to get 352x576.

I've tried leaving them as full D1 and I can't see a difference as regards resolution.

Using half-D1 means however, for the same MPEG-2 bitrate, more bits/pixel so the quality should be better, for the same bitrate,for VHS-to-half-D1 compared to VHS-to-D1.