Long Video / Half-D1 Question

[stevie1877]

Hi, I am trying to convert a 3-hour home VHS video to DVD (4.7gb) and trying to keep the bitrate as high as possible to avoid pixelation. I am capturing the video to an avi file through my digital camera using digital passthrough, so no video capture card involved. When encoding the avi to DVD format mpeg-2 the suggested 6000kbps predicts a file size of 8100Mb, so I have to (approximately) half the bitrate, which results in noticeable pixelation. I have been reading about half-d1 but am not sure quite how to use it. I'm trying to create an avisynth script that will half my resolution to 352x576 (PAL) as apparently there is no/little loss of resolution in doing this. However I'm not sure how to go about this. I have a script that essentially reads

avisource("D:\CAPTURE\capture.avi")
crop(8,0,704,576).BicubucResize(352,576,0,0.5)

and then using mainconcept mpeg encoder to convert it. I don't know if I've got it right to this point, but this is where I get lost. If I convert to DVD format, the suggested size (at 6000kbps bitrate) is 8100Mb - too big. If I leave at the suggested format of MPEG-2, it allows me to keep half-d1 as set in the avs but reduces the bitrate to 4200kbps, giving an estimated file size of 5600Mb - still too big to fit on one DVD, and therefore this hasn't helped me at all - I still have to reduce the bitrate further to the point where the quality is unacceptable.

Have I got the wrong end of the stick?
Can I not use half-d1 to fit 3 hours of video on a DVD?
Is my script wrong in the sense that I cannot use it "as is" to achieve this?
Am I doing something wrong with my encoder?
Should I be using a different app. (I'd rather avoid this if poss.)?

Apologies for posting a new topic but I've been doing some digging and am a bit confused / lost now.

[Nelson37]

Can you use a crowbar to fit 15 people in a VW? Absolutely. Will they be happy about it? That depends.

Your problem is conceptual. You can use nearly any resolution you want to fit any amount of video you want onto a DVD. Whether or not the calculated bitrate looks good to you is a subjective question.

Some things you can do, ordinarily IVTC would be first on the list but probably doesn't apply to your home video. Next would be smoothing filters to reduce detail and macroblocks at a given bitrate. Splitting to two disks or using DL media next, or reduce resolution still further to 352x240 to allow for the use of even lower bitrate.

Your procedure is correct, your evaluation of the results is different from others. Also, did you use a tripod to shoot the home video? This makes a dramatic difference in the required bitrate.

[Richard_G]

You are quite right in trying to use half D1. As you have found, full frame at 6000 kbs will allow over an hour and a half on a disc. If you reduce the frame size by half, you can reduce the bitrate by half with very little noticable difference in quality. Half D1 can sometimes look slightly soft, but that is all you should notice. The amount of compression that your mpeg encoder needs to apply will be exactly the same so you will get no more macroblocking and mpeg artifacts using 352 x 576 at 3000 kbs as you will at 720 x 576 at 6000 kbs.

How you would go about it with a avisynth script, I have no idea as I've never used one. Can't you simply import the DV AVI file into Mainconcept and tell it to encode to mpeg2 at 352 x 576 with a maximum bitrate of 3000 kbs? In fact, for three hours you could probably go up to nearly 3500 kbs and still end up with a file that will fit, particularly if you use VBR.

[Nelson37]

Was looking at the original post again.

You speak of seeing the macroblocks, but then refer to predicted file size as though the encoding had not actually been performed.

Particularly for the half D1, did you actually encode at the lowered bitrate and check the results? Predicted filesize not always accurate.

Good point on the VBR, Richard.

I just assumed you would be using that. Juggling the Avg, Max, and Min can often make the difference for marginal filesizes and/or quality. Different encoders will react in different ways, some will rigidly adhere to the avg, some will go both well under and over, some will tend to only one direction.