TRV33: Widescreen mode or 4:3 with black bars?

PainterPaul wrote on 3/12/2004, 5:05 PM
Greetings to the Vegas community on this forum!

I have been an avid reader of this forum for several months, and have posted a few times here with issues and have had most of my questions answered with good results. That said, I am not a professional, but more of a hobbyist but with future aspirations at some point – perhaps.

Before I purchased my Sony TRV33 camcorder, I read that the 16:9 feature was somewhat “superior” to the competition because it had some sort of technology which in some way added more pixel information to the image which was different than the stretching other similar cams do (which for them, equals a quality loss over shooting in straight 4:3.) Since then, for me the question has been percolating, and now, since I read here, I want to know if this is true and if so, how.

I have been reading a few threads regarding the issue of shooting in 16:9 vs. 4:3 (adding black bars to the 4:3 in order to have widescreen footage). The issue has been which is the best quality over all when shooting with a cam which does not have a true 16:9 lense, etc. Since I love widescreen over full-screen, and since my last shoot was in the TRV33 widescreen mode, I thought I’d get this figured out once and for all: which is the best for me in terms of quality?

My last shoot was a darkly lit series of barroom scenes of a birthday party. In my opinion, considering the low, natural barroom lighting, I am very pleased with the over all quality of the final DVD. I have had none of the issues some have talked about with “widescreen” footage; blurring, etc. With my cam, the only blurring came from the focus itself. Everything else is quite sharp, in fact. The final edited DVD has cross-fades, and a few color-fades only. The beginning 15 – 20 seconds or so is a composite of the opening clip with 10 fast, 1 –2 second clips of a sampling of the evening superimposed over that one clip. Nothing else fancy.

I would really appreciate if those who are willing would take the time and help me understand what it is my camera is doing that is “superior” to others in 16:9 (widescreen mode). HOWEVER, now I’m thinking that if I shot in 4:3, and added black bars instead, I might have had better quality, especially when considering the low lighting. I would greatly appreciate some tutoring on this, because, as stated, I will be doing mostly everything I do (for myself) in widescreen mode one way or the other.

I realize I will need some advice from someone who is familiar with my cam and how the 16:9 mode is supposed to be different/better. Maybe Sony can comment or point me/us to a link or source – I can’t for the life of me find where I read this before, and I have been looking – even to the Sony website where I found nothing on the TVR33! But there is supposed to be some way in which this cam deals with the so-called “stretch”, in a different way, which yields superior quality over its competition. Having not seen what other cams yield in their respective wide mode compared to what they’d yield in full mode (with black bars), I can’t say. I want to know which road to take now, because I am beginning. Obviously I can’t re-shoot the same footage to make a comparison (which I would if I could).

Thank you! I am looking forward to a learning experience!

Paul

Comments

farss wrote on 3/12/2004, 6:28 PM
I don't know the specifics of your camera and if you can find out no more than some throw away marketing line then I don't fancy my chances of finding anything either. But first let me say this, if you're happy with the results what the hell else matters. The pros will wince but hey you're not tying them to a chair and making them watch it are you.
There's only one prosummer camera with true 16:9 CCDs, the Sony PDXC10 or some confusing number like that.
The argument is that when you go into electronic squezze mode in the camers it's not using all the available CCD elements so vertical resolution has to suffer. If you've got more elements in the vertical direction than scan lines I cannot see how it would matter that much. In other words if they were being overscanned in 4:3 by 30% then you still shouldn't see a hit in 16:9.

Now if you try doing the job in post, this is a different story, You only have 480/576 lines to start with, you have to throw 30% of them away and interpolate the remainder back to 480/576. Something gotta give! So doing it in camera has got to be a better way to go.

Best method is an anamorphic lens for a number of reasons but they are expensive, probably more than your camera. So as I said at the start, if your happy with the results, don't fret. One thing to be certain of, each step up the ladder in quality with this kind of stuff is about a 10 times step up in price and difficulty in use. You put a big hunk of glass in front of the camera, next issue is lens flare, need a good matte box, hmm how to mount on camer? Need rods and mounts, oops just blew another $3K. And now the damn things getting heavy, need tripod or more likely even bigger tripod, another $1000. Next thing you need someone to help carry all that gear.
PainterPaul wrote on 3/12/2004, 10:51 PM
Hey farss,

I’ve read many (I should say *many*) of your posts.

>>The argument is that when you go into electronic squezze mode in the camers it's not using all the available CCD elements so vertical resolution has to suffer. If you've got more elements in the vertical direction than scan lines I cannot see how it would matter that much. In other words if they were being overscanned in 4:3 by 30% then you still shouldn't see a hit in 16:9.

I'm pretty sure, but not positive, that there was 14% more vertical when using the wide-mode with this camera, when compared to others of similar value. We're not talking about a high end cam. Slightly higher end of the basic consumer cam (unless Sony wants to add something).

So if I get you right (and I want to know) the vertical lines in the normal cam when shooting in wide mode, are stretched (beyond the 4:3) hence, a loss in quality because of an Interpolation from the original 4:3? In my sense of understanding, that would be similar/equal to a digital zoom effect? In other words, it would be like once you pass the 10x optical zoom, the cam would start to digitize to the scene in order to bring it closer, interpolate (add crap which is not really there) in order to bring the shot "closer". (This is why I turn off digital zoom altogether on my cam – and maybe why I might rethink shooting in widescreen-mode).

So the “normal” best shot would be with 4:3, same as if it were below the 10x zoom, in my example above . Anything beyond that 10x zoom would be a quality loss because we are going beyond what the camera is capable of in terms of its ability?

I’m trying to put this into words/concepts that I understand…. Just thinking here…

Okay, so this really helps me a LOT!

If what I have said, based on my understanding of what you have said, and of course from what I limitedly know, I’d need 30% more over-scanning in wide-mode to get the equivalent quality had I shot in 4:3. If this is true, and I don’t doubt you for one second, then even if my cam DOES have 14% more vertical over scanning, (I think but not sure) then I’m still 16% Sh*t/out of/luck!

Very interesting indeed, and hopefully I have this concept down!

Anyone/everyone. I’m not at all pissed if the above is true, as I hope it is because want to learn, even at the expense of screw-ups. I am glad to learn for future projects. Does my digital-zoom analogy apply here?

(Farss) Conceptually, am I on the right track? I’m a visual person in terms of learning. I need that concept. I’m really bad with figures/data/calculations. And if I need to learn the figures/data/calculations, I will, but I will still need a concept for that as well!

Thank you Farss. I really appreciate your input.

Paul
farss wrote on 3/12/2004, 11:52 PM
WHat I'd keep in mind is this. Assuming you're in NSTC land then DV is only 720x480 pixels, not exactly huge resolution. That doesn't change if it's 4:3 or 16:9, so when you go to view it the same number of pixels have to fill up a wider screen, something got to give, horizontal resolution in terms of line per inch of screen has to suffer. Just how much difference that makes to our percieved resoltion I don't really know but it sure can't be getting any better!

Now on the camera side, despite what the marketing guys would have us believe anymore than a certain number of pixels is actually a negative thing, in fact you'll notice broadcast grade cameras have less pixels in their CCDs than consummer ones. Main reason for sticking more in the consummer ones, they also take stills and OK they're only single CCD but in general for a given sized CCD more than enough pixles makes each element smaller and hence less light sensitive.

Git that bit out of the way, not much to do with 16: versus 4:3 but what the heck! Your digital zoom analogy does apply BUT, simply analogy, take a 720 x 480 still into Vegas and zoom in 200%, looks pretty bad. Now take a 1440x960 still and zoom in to 200%, no change in resolution because we had more pixels than we needed to start with.

Now same thing should happen with the video camera, if the CCD has more pixels than you can record onto DV and then you crop some of them off so long as theirs still as many as need to fill the DV frame you'll not see much difference. Now I'm certain it's all a little bit more complicated than my basic knowledge but that's my theory and so far no ones been able to shoot it down! I've heard the opposite argument on this topic, if you have a 16:9 CCD then its got to be bad for 4:3 because you're cropping off the sides, poor camera manufacturers cannot win.

So it sounds like in your cameras case it scans more of the CCD in 16:9 than it does in 4:3, if they mean 14% more area that would compensate for the 29% or whatever effective reduction in vertical size so I'd sleep easy.

Simplest answer of course would be to get a standard resolution chart, print it out on a good printer and have a look yourself. Nothing beats doing a test. Should find one at pixelmonger.com if you don't spend too long drooling over the Viper, think he's got some good words on how to read the results properly too.

I'd add you still need to be a bit more careful about a few things shooting 16:9 than 4:3. Keeping your horizontals horizontal would be my first concern and avoid moving the camera as much as you would shooting 4:3.
PainterPaul wrote on 3/13/2004, 4:25 PM
Thanks, farss, will comment further after the weekend!
farss wrote on 3/13/2004, 6:12 PM
I'm heading OS for a couple of weeks so I'll have to leave further assistance upto others.