PPro : turning 4:3 footage into 16:9

Monglane

Monglane

Well-known member
Hi all,

I usually shoot Squeeze, so my footage has a 16:9 ratio. I have a little project for which I'd like to try something else, i.e. shoot Normal (4:3) with the AspectCorrect thingies on the LCD, and then add black bands in Premiere Pro to obtain, at the end, the same 16:9 look.

I've read a couple of related threads on this forum but I'm not sure what is the best way to do that in Premiere Pro. Someone mentioned a "Matte" option in the Title menu, someone else said to just add 12.5 percent-high black mattes at the top and bottom (which would be 576 x 12.5 % = 72 pixels for each matte's width)...

As I've never done that (I'm trying it to see if there will be a visible increase in resolution over the Squeeze mode), I'm curious as to the best way to proceed.

Many thanks in advance for your suggestions.
 
Why don't you get hold of an anamorphic lens? Some claim there is an increase in resolution going this root, record it stretched, and 16:9 it in post.

Just a question,

Thanks,
Derrick
 
Derrick, using an anamorphic also causes some loss of resolution, because it is probably not of the optical quality of the Leica lens, and also because even if it were, any optical complement that was not designed as an integral part of the lens to begin with, is likely to cause some loss of resolution.

But anyway, and even if I were ready to fork out the money (which I'm not at this juncture), that was not my question (but thanks for your input all the same).
 
I have until now shoot 4:3 an just now trying to get some idea of what would be the best solution for 16:9 (I'm not going to put my money in the anamorphic adapter).

I shot three pictures of the same scene with 4:3, squeeze and in-cam letterbox then I combined these three clips in one footage side-by-side in a 16:9 project in Premiere. The 4:3 footage was streched to fit horizontally, the squeezed footage was streched to 16:9 horizontally (Interpreted) and the in-cam letterbox footage was upprezzed with Photozoom.

I thought the output would show some bigtime difference between the three images. There WHERE som difference but not THAT much. I tested this on my 100hz 32'' panasonic widescreen (PAL). My conclusion is that I will just use the setting that suits my project, and really don't bother too much about it as long as it's not going to be put on a big screen anyway.

By the way, I have made an accurate 16:9 Photoshop template. Myself I use Filters>Keying>Image Matte Key. To me this template is more accurate than the title-template. Just let me know and I can mail you this tamplate an also a tiff-file from my test above.

Espen
 
Thank you for your comments, Laffen. I would indeed be interested in looking at your Photoshop matte. If you want, you can email it to monglane at a2co.org.

Tks !
 
Monglane said:
Thank you for your comments, Laffen. I would indeed be interested in looking at your Photoshop matte. If you want, you can email it to monglane at a2co.org.

Tks !
Click to expand...

I have sent you the photoshop template.

Espen
 
Laffen, I see that in your Photoshop document, each matte is 72 pixels high. This corresponds to a masking of 12.5 percent of the original, 4:3 image, which is in line with was I've seen described as correct in another thread.

Thus, the unmasked portion of the image is 720 x 552 ==> (576 - [2 x 12.5]) = 552.

However, 720/552 = 1.304, whereas the 16:9 ratio is 1.777.

Therefore, since the "16" part of the equation is 720 pixels (image width), the "9" part ought to be 405 pixels (image height), oughtn't it ? If so, 84 or 85 pixels should be covered by the matte both at the top and at the bottom (576 - 405 = 171/2 = 84.5) ?

Or am I missing something ?

I know this discussion goes beyond the Premiere forum, but since it started here I'll continue it here as well.

Thanks in advance for your input.
 
Monglane said:
Derrick, using an anamorphic also causes some loss of resolution
Click to expand...
Actually, with tests using resolution charts, the anamorphic lens adds resolution. Enough so that it surpasses and slight loss from not being an integral part and lower lens quality. There's been several tests to prove this. The real issues I notice with the anamorphic are: 1. Difficulty to focus because of vertical/horizontal differences 2. Increased depth of field (for those looking for narrower) 3. Cost
Monglane said:
Therefore, since the "16" part of the equation is 720 pixels (image width), the "9" part ought to be 405 pixels (image height), oughtn't it ? If so, 84 or 85 pixels should be covered by the matte both at the top and at the bottom (576 - 405 = 171/2 = 84.5) ?
Click to expand...
Correct. But, to be clear, remove half that value (84.5) from the top and from the bottom.
 
HansK said:
Actually, with tests using resolution charts, the anamorphic lens adds resolution. Enough so that it surpasses and slight loss from not being an integral part and lower lens quality. There's been several tests to prove this.
Click to expand...

Add resolution ??? I'd be curious to know how that can be, optically speaking, and to see those tests, too ! I'm not doubting you, just extremely surprised and curious... To me, it's like when you get white-out pixels because of overexposure : there's no way you'll ever get color back into those pixels.

But as I said, I'm curious to know more...

HansK said:
But, to be clear, remove half that value (84.5) from the top and from the bottom.
Click to expand...

Why ? If I remove half of each of the two 84.5 pixel-masks, I'll end up with two 42.25 pixel-masks, i.e. a picture 720 wide by (576 - [2 x 42.25]) = 491.5 high, in other words a ratio of 1.46, and not 1.777 or 16:9... ??
 
Monglane said:
Add resolution ???
Click to expand...
Ok, I read your comment again. If your comment is simply comparing a lens to a lens then you are correct. The stock Leica lens in the camera is higher quality than the anamorphic lens. If you are comparing the resolution of the stock lens and squeeze mode to an anamorphic lens and non-squeeze mode then you will find that the resolution of the captured image is higher.

Monglane said:
Why ? If I remove half of each of the two 84.5 pixel-masks
Click to expand...
Dang, I was tired when I did that reply.

Ok, one of the items you are not taking into account it aspect ratio. For ntsc 4:3 the width of the pixels are 90% smaller. So:

(720 * 0.9)/480 = 1.35 which is approx 4/3 ( the actual vert rez is a little more)

For ntsc 16:9 the width is 120% larger so we have: (720 * 1.2)/480 = 1.8 (again rounding error since vert rez is a little higher)

Then to find the vertical height just divide the 480/1.2 = 400. Therefore, a 16:9 ntsc image is masked to 720x400.

For a PAL image the aspect ratio is 1.42, so:

(720 * 1.42)/576 = 1.775 And your mask would then be: 576/1.42 = 406

Does this help?
 
Or for a cheaper and more productive solution, purchase Photozoom Pro and use Disjecta's method of uprezzing your 4:3 letterox to HQ 16:9.
 
HansK - good call, it's important not to forget about pixel aspect ratio.

As for masking - myself, I usually either use the title preset (if I want a really wide cinemascope kind of look), or go the photoshop route. However, instead of using Image matte, I just two black bars with a transparent middle section out of photoshop (in something that holds transparency like a tiff or png), and then drop it onto my highest video track.

One nice thing about shooting 4:3 and letterboxing instead of shooting anamorphic is that you can reframe in post (by adjusting the vertical position of your clip so that a different part shows through the matte). While some purists will say that your DP should have framed things right in the first place, we all know that this isn't always the case :)
 
HansK said:
Ok, I read your comment again. If your comment is simply comparing a lens to a lens then you are correct. The stock Leica lens in the camera is higher quality than the anamorphic lens. If you are comparing the resolution of the stock lens and squeeze mode to an anamorphic lens and non-squeeze mode then you will find that the resolution of the captured image is higher.
Click to expand...

HansK : OK, I agree with you and understand what you're saying.

HansK said:
For a PAL image the aspect ratio is 1.42, so:

(720 * 1.42)/576 = 1.775 And your mask would then be: 576/1.42 = 406

Does this help?
Click to expand...

Yes it does, and once again we are in agreement. :thumbsup:

Laffen said:
Think I have to redesign my photoshop mask. So then the top and the bottom should be 84 pixels??
Click to expand...

Laffen : yes, I believe so. I have changed it myself and it looks fine.

Also, the pixel aspect ratio doesn't come into play here, since overall the image is and remains a 4:3 image. It only happens that the non-black part of it is 16:9-sized.

It makes sense to me, but please *do* correct me if you think I'm wrong !
 
Yes Monglane, I think we have some kind of conclusion on the black bars. I have tested it a little bit now, and it seems correct. 84 pixels it is.

Thank to all you guys helping out with these problems. Without you we all would be lost. :)
 
You must log in or register to reply here.
Back
Top