Squeeze .vs Letterbox - Technical Theorem

Re: Squeeze .vs Letterbox - Technical Theorem

As fascinating a discussion this is, I'm trying to follow it with a pencil in hand to diagram it out but am having trouble doing so. Would someone mind diagramming out the resolution differences in scale as related to the DVX ccd and the up-rezzing path and posting it somewhere? I think it would greatly clarify the discussion.

Thanks , Joe Murray DGA/DP-IA600
 
Re: Squeeze .vs Letterbox - Technical Theorem

Ok this is where I still am losing you... I am not blowing up tthis footage (going to be shooting a documentary but i'm positive there will be no film out). I am just trying to get my settings straight before I commence shooting. If i shoot in regular format with no anamorphic lens but frame it for 1.85, and then throw black bars on in post... I'm not exactly altering the image or the quality of the image, and I could most certainly still make a DVD out of this and display it on a television with no problems... I dont see where you are getting this whole "it'll be fine for web format, but..." business. I've done this before and played a dv tape on a television screen and its come out fine.
how would applying a small area of black bar to the top and bottom of an entire project affect it... i'm not messing with the resolution of the image or size, its still 720x480 overall, the dimensions of the video haven't changed... so why would this not work or why would i do anything else instead of this. There will be a beta copy at the end and sure i'll try to get distribution but i'm not sure why in the heck this wouldn't be perfect. for like a regular tv screen, or even projected, it would simulate the widescreen look. why bother with pixel stretching blah blah.... and a second question to this, is if i need to pixel stretch or whatever you say... can I do this by shooting my regular 4:3 method but have my composition in a 1.85, and then put black bars on and THEN stretch pixels if necessary? too many questions, just trying to get this settled before i start. if i shoot my way, do i have the option of stretching properly in post if i choose to do so?
 
Re: Squeeze .vs Letterbox - Technical Theorem

Everything you're saying is just fine, you have a good handle on what you're doing.

The discussion at hand differs from your situation in that everyone here has been discussing how to get actual widescreen video. You're making 4:3 video, and adding black bars to give it a widescreen look. People on this thread have been discussing ways to get 16:9 video.

Most all DVD's are authored in 16:9, appropriate for showing on a 16:9 television. Your video would not be appropriate for showing on a 16:9 TV (which is, basically, not a problem, since about 98.5% of the TV's in the US are 4:3). But for people who want to get video that would be appropriate for a 16:9 TV, that's where the discussion of pixel stretching or anamorphic adapter comes in.
 
Re: Squeeze .vs Letterbox - Technical Theorem

Let me first thank you all for your quick and thorough responses, seriously this site and all of you are so helpful its unreal. Now Barry, thanks to you I dont feel completely clueless but I do have a couple further questions. So lets say I shoot the doc the way I'm talking about, and have the 4:3 with widescreen look... but that I decide I want an actual 16:9 version of this... First off, how would my 4:3 video, which has this widescreen "look" that i am going for, look on a 16:9 tv... would it not fill the whole screen because it isn't true 16:9? Second, if I did decide to do what you guys are saying instead, would stretching from the way I have it laid out work ok? u knoe the 4:3 with the black bars? i'm thinking i can render out 2 versions or at least have the choice to do so if what i'm thinking is correct... you know just in case after i've shot all this i wanted to go a different way and not have a complete friggin headahe. and once again, you guys are the best... thank you so!
 
Re: Squeeze .vs Letterbox - Technical Theorem

..yes. i see now that what your doing is fine. in your earlier post you has mentioned "cropping" a widscreen look out, and i misunderstood what you were trying to do. your just doing whats called "matting" for a 16:9, or 2.35:1 or whatever you want image left in the center of the 4:3 video, by matting the rest of the image out in black. which looks the same in the end as letterboxed footage..which is of course what you get if you had originally shot actual 16:9 or 2.35:1 dimension footage through the useage of anamorphic adaptors, or having that dimension of image sensor/film, and shrunk it down to fit in the center of a 4:3 screen, which is called "letterboxing". sorry for the confusion , but the thread as youve already been notified of was about how to get and deal with actual 16:9 footage. so i didnt know you were dealing with 4:3 letterboxed images. anyway, good luck with the project!, and again sorry for the confusion.
 
Re: Squeeze .vs Letterbox - Technical Theorem

If you want to see what the scaling looks like in creating anamorphic 16:9 from a 4:3 source, apply a 3D effect to the footage and scale the vertical. If you shot with letterbox. scale the vertical until the black bars are no longer in the image. Becuase this is being done on the vertical, quality is an issue since lines need to be doubled etc. Using THIN when shooting will help in this endeavor. Some 3D applications have diffirent levels of quality when rendering so make sure to use the highest quality.

This is sometimes referred to as a "tilt & scan" since you are keeping the width and moving the target image up and down as needed. A "pan & scan" is the typical term for creating 4:3 from a wider aspect ratio where most of the moves are left to right.

As the digital editing consultant on "November" they used the resize effect on their Xpress Pro and Adrenaline to properly place the intended 16:9 area in the center of the frame (using the grids). So there was no resize involved, just moving the image up and down as needed. When it was complete, the uprez was done via a Teranex with a center crop for the 16:9 via a digibeta output from the Adrenaline to the 1080p/23.976p master. Shots needing to be recomposed were few and far between as most of it was properly framed during the shoot.

Color correction and all effects were done in Adrenaline and After effects after recapturing the sources at 1:1 in a 23.976p project so they were not subject to the multigeneration render issue as well as being in 4:2:2 color space for the color correction. There was some final color balancing done in HD space before the Sundance screenings. Sundance screened both the HD as well as a 720 x 1280 HD in Windows Media 9.

Michael
 
Re: Squeeze .vs Letterbox - Technical Theorem

alright michael, so if i am understanding what exactly it is that you're saying... i can shoot in 4:3 with no adapter, and just frame for 16:9 mode, and in post for regular tvs i can add the black bars... or I can do the stretch they did in "november" and have bare minimum quality loss and have a 16:9 image... so basically shooting the way i am now (in regular 4:3 but framing for wide) leaves me the most amount of options?
 
Re: Squeeze .vs Letterbox - Technical Theorem

[quote author=John Brandis link=board=Events;num=1076431425;start=0#23 date=02/15/04 at 11:28:37]First off, how would my 4:3 video, which has this widescreen "look" that i am going for, look on a 16:9 tv... would it not fill the whole screen because it isn't true 16:9?  [/quote]

Depends on how the TV is set up. I would guess that what you would see would be stretched-out-wide video with black bars on the top and bottom. But, some TV's may automatically enlarge the center part, cutting off the top and bottom, so what you would see would be "correct" -- except that it will have been zoomed-in-on, so it'd be low-res. Finally, some TV's may choose to "window-box" 4:3 footage, showing the full height but adding black bars on the sides. Seeing as your footage already has black bars, the net effect would be to have a small widescreen image within a big black frame.

Second, if I did decide to do what you guys are saying instead, would stretching from the way I have it laid out work ok?
Click to expand...

Yes, it'll be somewhere around the same quality as having used "squeeze" mode in the first place... DEPENDING ON WHAT PROGRAM YOU USE TO STRETCH IT. I got lousy results in Premiere, excellent results in Vegas, I've never used FCP but heard reports that it didn't work out so well. Use a good-quality stretching program and you can get decent 16:9 footage. It would help you very, very much to have shot in progressive/thin line detail to get the best results from post-stretching to widescreen.
 
Re: Squeeze .vs Letterbox - Technical Theorem

John - correct in your workflow. Just test the stretching for the 16:9 anamorphic version for quality puroses.

Michael
 
Re: Squeeze .vs Letterbox - Technical Theorem

MichaelP - Thanks for the insider look at November's pathway from 4:3 to 16:9. Answered a few procedural questions I had lurking in my mind for sure.

John - You can indeed setup for 4:3 and then take any route you would wish to go for other formats. It's all up to you.

If you have no desire to go to either HD or film, you could realistically do your shots in either standard 4:3, letterbox, or squeeze. Pathways exist to letterbox, and 16:9 from any of the 3 modes. My contention here is that of the 3 choices, the best technical quality, when final output is destined for 16:9 (or for that matter widescreen letterbox), should be via squeeze mode.

BTW - Did some shooting yesterday over at Vedauwoo and had some nice 16:9 stuff come out. Boy, there are sure alot of settings to try, to run through the entire process to 24p DVDs.

-Rodger
 
Re: Squeeze .vs Letterbox - Technical Theorem

Ok then another question, how do I go about doing that progressive thin line business? where do i set this? how does it effect the image quality?
 
Re: Squeeze .vs Letterbox - Technical Theorem

Set it in the menu under Scene file (3rd screen).

I just shoot in 24PA and leave it thin.  It switches to thick in 60I mode automatically (I don't use 60I though). I guess if you get twitter for some other types of artifacts in progressive, you could try med or thick.

I used these settings yesterday outdoors in bright sun and got some really nice stuff:

Start with f6 (squeeze mode) and adjust detail level to -4 and v detail to -4, detail coring to +4 and master ped to -6.  Use ND filters as needed and ATW.  Burned to 24p DVD widescreen (16:9) and it just looks stunning.

-Rodger
 
Re: Squeeze .vs Letterbox - Technical Theorem

Ok so I now know how to set it but... I have no clue what this does and in what situations it would be appropriate for? is this a setting to use in 24p mode for all situations as a general guideline that makes your images look better some certain way or easier to mess with in post? I'm just wondering what it is and how it effects things. Thanks in advance...
 
Re: Squeeze .vs Letterbox - Technical Theorem

Can someone please tell me if what I'm getting out of this is correct? To get true 16x9 footage from shooting 4:3>

- Crop the 4:3 footage in a 4:3 sequence (I'm using FCP) to 16x9 dimensions
- Import the now cropped footage into a 16x9 sequence. At this point it should not fill the entire "screen" area, but will be smaller
- Resize/Stretch the footage to fill the "screen" area in the 16x9 sequence.

Walla, I now have 16x9 footage.... right? If this is correct, do you lose resolution by stretching the image? If loss is minimal, how far can you stretch it?

Thanks for your infinite patience,
Chico
 
Re: Squeeze .vs Letterbox - Technical Theorem

thats the way it goes.. but you can do itall at once, by using a 16:9 project, and cropping to the 16:9 image, and then expnding it to fill the frame. theres no reason to do that in a 4:3 project. your pixel info resolution will be degraded in the end. basically your trying to strecth the info of the 360x480 pixels over 864x480 pixels. some programs will smooth and blend, and interpolate to get a pretty decent image still though.
 
Re: Squeeze .vs Letterbox - Technical Theorem

basically your trying to strecth  the info  of the 360x480 pixels over 864x480 pixels.
Click to expand...

Shouldn't it be stretching 720x360 footage into 853x480? Your just 'losing' the vertical resolution, right?
 
Re: Squeeze .vs Letterbox - Technical Theorem

If you're making it for computer display, 853x480 may work, but if it's for DV, it must be 720 x 480. So in order to make 16:9 video, you have to stretch the center 75% to be full height. And that's why you lose resolution: because you're using only 75% of the frame to fill the full frame.
 
Re: Squeeze .vs Letterbox - Technical Theorem

The point is that if you don't use anamorphic optics on the DVX and you want to end up with real 16:9 images, you will loose resolution eitherway. In non anamorphic mode, the effective vertical dimension of the CCD capture area is reduced by 25% and so is te vert.rez. The difference can only come from the in-camera uprez processing performance (before compression) vs the uprez performance in post when using the (4:3) matte approach.
 
Re: Squeeze .vs Letterbox - Technical Theorem

The point is that if you don't use anamorphic optics on the DVX and you want to end up with real 16:9 images, you will loose resolution eitherway. In non anamorphic mode, the effective vertical dimension of the CCD capture area is reduced by 25% and so is te vert.rez. The difference can only come from the in-camera uprez processing performance (before compression) vs the uprez performance in post when using the (4:3) matte approach.
 
Re: Squeeze .vs Letterbox - Technical Theorem

..mike.. i may have gotten the #s backwards lol. but as far as the dimension of 16:9 video. well thats really a big bucket to swim in . 853x 480 is much closer to a true 16:9 but alot of programs will make "16:9" somthing that is close to that. for example premiere edits 16:9 @ 864x480. does it matter? no. i personally preffer the slightly wider frame dues to overscan on the tv, it helps preserve the full width of the image being viewable.
 
You must log in or register to reply here.
Back
Top