Does rendering DV1 to DV 1 cause quality loss?

[chloh]

I always assumed that working in the same project settings would result in no quality loss, for example when saving part of a DV1 avi clip as a new file (smart render was turned on, there seemed to be no rendering as the process was v quick). However upon close examination the saved clip has a slight change in quality from the original (I observed how white texts, which is good for checking, appeared in the original clip and the saved clip - in the latter the edges were just a bit more jagged and some colour distortion crept in).

I did this in ULEAD Video Studio 7 and wonder if it is the same in MSPro 6.5?

More important 2nd Question:

I used the Videostudio Paint to paint out some background in my clip, and when I rendered it I noticed the quality seemed poorer. The question then that came to mind is this - if I render a clip where I had done some editing on (eg used mask, paint etc) but I use Smart Render and the same project settings (ie I render from DV1 backt o DV1) will the unaffected parts of the video be rendered as well, THAT IS, is there any degradation in the parts of the frame that were not touched by the edit?

Cheers
ch

[Devil]

Ideally, with Smart render on, the parts with the red and green lines on the timeline are only copied and there cannot be any loss of quality.

With it off, there is a minute loss of quality with DV to DV (same settings) but it becomes visible on a TV screen only after about 5-10 generations. A single (or even three or four) re-rendering should not produce any visible difference on a TV, which is what counts.

Many persons, myself included, have done multiple generation tests and then examined comparatively single frames. After 20 generations, things change to the extent that someone who is blind in one eye and couldn't see with the other, could see the differences. I think you can find examples if you Google for them.

[chloh]

Thks v much, "devil".

Just a follow up question - if my source is Type1 DV, when I render (I assume this means use the Save > Video command, right?), should I choose DV1 or AVI uncompressed? I always thot DV1 was uncompressed, but if u peek into the settings DV1 has 11% compression. Rendering to AVI gives me a file that is four times larger.

A bit confused with this - in Ulead VS7 I simply choose "same as project settings" and it seems not to recompress or re-render untouched frames. I am a regular VS7 user who is a bit of a MSPro newbie.

As mentioned in original question, in VS7 even when I use smart render and just save a trimmed DV1 to a new file (DV1) I notice slight colouration distortion, especially clear in white text titles. I dont know why this is so.

Cheers
ch

[Devil]

If you re-render, do so into DV. You cannot magically gain what isn't there. It's like expecting a worn, pre-electric, 78 shellac record of Dame Nellie Melba or Caruso to sound like a DDD CD! (OK, perhaps I exaggerate!)

You don't say what system you are using. NTSC DV is worse for colour artifacts than PAL DV. In both cases, there is one set of colour sampling for 4 pixels, but the way it is done is very different. If you want to delve into it deeper, NTSC has a 4:1:1 colour space, meaning that on each line the colour is shared between 4 pixels. In PAL, it is 4:2:0, meaning that the colour is shared between 2 pixels on every other line, but, because it is interlaced and the colour "burst" is phase alternated between lines, it gives the appearance of a much better colour resolution. When a DV signal is taken into any app (or a TV), the 4:1:1 or 4:2:0 is converted, as best it can, into 4:4:4, i.e. every pixel has full luminance and chrominance. When this signal is re-rendered (not using Smart Render), the codec you use will do its best to restore it exactly as it was, but small variations, normally invisible for ordinary TV viewing, could occur in chrominance encoding, usually poorer for NTSC (hence Never Twice the Same Colour! ). This could also be a function of the codec you use; some are created more equal than others. Normally, the only time you would normally be bothered with NTSC colour space is when keying a blue screen or green screen, where you might see some horizontal overlap fringeing.

Anyway, because uncompressed shares the same chrominance signal over each block of 4 pixels, you gain nothing by repeating it 4 times, rather than recording it just once, and you lose a lot of disc space plus the risk of the larger required bandwidth causing video glitching if your HDD has difficulty in coping with the bitrate.

Sorry, I can't compare with VS7 as I've never used it.

Hope this helps.