Please Help Me Set Up A Camera With The Correct Aspect Ratios

[flak-monkey]

I want the final image output by maya in my animation project to look like an anamorphic widescreen dvd. The target medium will be dvd. What are the correct device aspect, pixel aspect, and resolution settings that I need to use to get the image that I want?

[joojaa]

It depends, on wether you want to use square to interpolate ofrmat or squeeze format. but basically the idea is that the number of pixels is exatly the same as for pal or ntsc just the aspect is different 16:9.

However the filming devices do this in many different ways so you should match the comp softwares plate.

[joojaa]

It depends, on wether you want to use square to interpolate ofrmat or squeeze format. but basically the idea is that the number of pixels is exatly the same as for pal or ntsc just the aspect is different 16:9.

However the filming devices do this in many different ways so you should match the comp softwares plate.

[dryo]

For anamorphic NTSC DVD, the settings are

width 720
height 480
device aspect 1.777
pixel aspect 1.185

And for non-anamorphic NTSC DVD

width 720
height 480
device aspect 1.35
pixel aspect 0.9

Also, the number of pixels is NOT the same for PAL and NTSC. For anamorphic PAL DVD, the settings are

width 720
height 576
device aspect 1.777
pixel aspect 1.422

It sure is strange that there are no presets for DV or DVD within Maya, whereas there are presets for CCIR-601. Must be a cultural bias against consumer technology?

[joojaa]

Its not that simple! Nor that straightforward. That works just as well as few other options. (tough gtanted for ntsc that seems to be 6the most common size)

The device itself has as many- or more pixels as the normal signal, yes it does. However its filtered a bit weirdly due to the fact taht ist the same signal so a modern direct show lcd will use that resolution and then scale it to another but a old tv works differently). Now a pixeled image that we have here is not perfectly corresponding to howe the signal works. (according to standard). The dvd disk on the other hand has more pixels yes. (some compressors account for the wirerdness others dont)

TV's just works plain weird! And today every brand works differently to enhance the image.

Now the transfeer may be done in many ways and the conversion rate depend s on the hardware. lets take something that has persets for dvd images youd notice theres 2-6 formats to account for.

But the thing is for best image its not said that one size fits all im afraind. It depends on how your doing the conversion into mpeg 2! As it happens the comprssed signal can be better if it gets more resolution or less, depends on compressor.

So the size depends on your output device! And since maya does not do output compression they dont want to confuse the user with many presets.

Chek your devices settings and add those presets.

Now since it will be mated with smething shot then it need to match that brands settings.

PS: Theyve fixed this total insanity of wild west formating in hdtv formats

[dryo]

We're not concerned with arguing the vagaries of random display devices such as LCD monitors. The only question on point here is what settings to plug into Maya for an anamorphic NTSC DVD.

Old-school PAL and NTSC have very well-defined standards, it's just the nonsquare pixel madness that confuses people. And for that we can thank Quantel and the rest of the broadcast industry for doing its part to prevent WYSIWYG video on the personal computer desktop.

[joojaa]

QUOTE(Aaron Ross @ 09/27/07, 02:01 AM) [snapback]273920[/snapback]
We're not concerned with arguing the vagaries of random display devices such as LCD monitors. The only question on point here is what settings to plug into Maya for an anamorphic NTSC DVD.

And im telling that you cant just render out the images blind, the output devices have quite free hand in deciding how the convert into the signal stream. So if it needs to match a plate you need to know how that plate will be built.

here are several factors for this such as whether the system wants to interleave the frames or not, because its quite impossible to do certain comp tasks on interlaced footage, yet undeniably the video feed is interlaced.

So you wouldn't even know if you should render 30(25 for pal) frames per second or 60(50 for pal) half frames per second and wether these half frames come upper or lower first. (also for filering reasons i render full frames then it dont matter)

Secondly you don't know hoe the comp artists will work, he might desire to work on square images thet he then intepolates flat

So ultimately theres no better way than to get sample footage up fornt or ask the person whio is going to do the comp.

So doing anything drastic up front is ill advisable. Or you can just tip thet the comp artists knows hios stuff better than you do. (in wich case asking him up front is far easiest)

So it dam right matters for the discussion!

[dryo]

Of course there are many variables and many possible workflows. It might be good to render "flat" square pixels even if the target platform uses nonsquare pixels. It might be good to render at double the frame rate for proper compositing for output to fields. It might be good to render at double resolution if the compression software works better that way. However, it may be that none of this is relevant to the situation at hand.

Certainly, it's a good idea to know what the requirements are if one is working with a background plate, or with any special compositing considerations. I don't want to oversimplify, but I also think it's counterproductive to make things more complicated than they need to be. When someone asks a simple question, one should try to give a simple answer. Occam's Razor and all that.

[joojaa]

And thats why my original asnwer was breif as i suggested to take look ate the video plates aspects.

It was just taht you oversimplified the problem way too much. You just cant go to rewnde blind like this. If he just puches those values it he might have to rerender a potenttialy expensive flaw.

Anll this is why you should look at the plate first before you touch any rendering. And yes theres no way to avoid doing comp (because you need to clip colors and color correct to match) so and the comp app knows the aspect and other settings most likely nativetly.

[flak-monkey]

What I've done before is render in 720x480 and add the black bars in after effects, helps me compose better anyhow.

Now to figure out this really backwards depth of field settings menu. Don't remember my photography classes teaching it this way.

[flak-monkey]

So then comes the question of how to output an image onto a dvd like this.

I popped in the Departed, which I know is the frame size that I like, and did a screen capture and found that the resolution is 720x405??? This is on a computer screen.

What is that? I'm still terribly confused. Tell me how to make it look like this when I put it in a dvd player.

That's the size of the frame that I want. Maybe I'm not using the correct terminology and it's throwing you guys off or maybe this has been made to be confusing on purpose. Even if I have to fake the image, that's the size of the picture I want when I put it in a DVD. Doesn't have to be anamorphic, most things that this will be played on are sadly 4:3 anyhow.

[joojaa]

QUOTE(flak_monkey @ 10/01/07, 08:39 PM) [snapback]274163[/snapback]
That's the size of the frame that I want. Maybe I'm not using the correct terminology and it's throwing you guys off or maybe this has been made to be confusing on purpose.

No its just made to support technology form 1940's and you know how confusing it gets when you talk to your granddad Same problem. You should actually reefer to the output options of your MPEG compressor/edit software such as premiere not the uninteralced output of your dvd on screen (because its not going to be making the disk).

So idea here is to take sneak peek on the step your going to make when putting it on the disk. Before you start that process then work backwards because thats what you need. This kind of go ahead check then come back is normal approach to things.

You have to understand that tv signal is not pixel by pixel the same a s computer screen its pixel for pixel unlike your computer screen where each vertical line is offset form the previous line by 0.5 pixels to the side, also each alternating line is half a time cycle ahead of the second line, so whet computer screens do is they processes this interlace, so what you see is notwhat you get. In fact its physically quite impossible to get the exact same thing as you gave on your computer screen in both terms of quality of image as well as depth of color bacesue TV's just plain suck on those aspects. (ntsc si known by people who work on it as being a acronym for not same color twice). Then theres the very wacky gamma and color space of the images themselves (which cant be don properly without having a tv signal output image input to a broadcast monitor, but can be guesstimated quite well with with icc).

So your option is just acceptable - good transfer... nothing perfect.

Simply put yes you could use those dimensions. And calculate the aspect ratio form the image on your screen

PS: yes its complex as hell. because you actually need those interlaced images or it may look off. Theres actually no way of conveying this well without experimentation.

[dryo]

Don't trust the screen capture at all. The DVD software player may rescale the image however it feels like doing.

The bottom line is this. The NTSC DVD format has exactly 720 pixels wide and 480 pixels tall. Those may be of two pixel ratios:

1. 0.9 pixel ratio for a DVD with a so-called "flat" 4:3 aspect ratio. This may be letterboxed if the content is not 4:3, but this will result in a lower resolution image. Not all of the 480 horizontal rows are being used.

2. 1.185 pixel ratio for an "anamorphic" or "16:9 enhanced" DVD. All of the 480 horizontal rows are used for picture information. When the 16:9 enhanced DVD is displayed on a 4:3 NTSC device, the image is scaled down to fit, and letterboxing is introduced at the playback stage. When the 16:9 enhanced DVD is displayed on a 16:9 HD device, then the image is scaled up to fit. Software players may do whatever they wish, but they usually output 16:9 enhanced DVDs to a screen resolution of 854x480 or 720x404.

Note that the term "anamorphic" is actually misleading. Both DVD formats use nonsquare pixels, so neither of them are truly flat. The term "anamorphic" is inherited from the world of 35mm cameras and projectors.

If you are outputting to a widescreen format, it's a bad idea to render to 4:3 and crop it later. It's better to have the render settings in Maya set to the actual output format, that way your resolution gate is accurate and you know what's in the frame and what's not WHILE YOU'RE ANIMATING. You also don't want to waste time and CPU cycles rendering something that no one will ever see.

[joojaa]

And if you still want those crop bars on top of that add them in post! Just render the section you need.

QUOTE
Don't trust the screen capture at all. The DVD software player may rescale the image however it feels like doing.

And it will becasue it wants to delete the interpolation and fit it to your screens pixel amounts for example 720/1024 is not even si if it wants to fit widt it may not have a choice.

But as i said don't trust the standard settings (because while the standard setting has a intelligent pixel size its not made out of one image but 2 superimposed offset ones, so the teriminology dont translate 1 to 1) look at the next stage application you have it might well want to stretch the images itsetself. Reason being uits often easier to work if what you see on screen matches to what will be output. Since the tv signal and thus dvd output is little more than that it does not matter if you blur a bit by stretch scale down. In fact if you render a image thats sharp it will look very disturbing on tv screen and helps to match those alternate lines.

But its easily cheeked because your need to get those apps in sooner or later why not test immediately up front. You dont work in void.

Theres lots of technology down always pin your entire work flow form start to end before you start. And what your work flow will be depends on what software and hardware you happen to have. Not chekking is stupid since it takes you 2 seconds to do it.

Analogy: if im going to print something i first check what kind of paper i loaded into the printer (or the printhouse uses before crop) before i design my aspect. Its the same kind of situation. Also you always calculate target dpi based on the output device if known ahead of time it saves you 10% in printing cost easily.

PS whatever happens you output wont look pixel perfect, when it does your in trouble, unless you use a hd p format! Because it is in fact impossible for the signal.

This is what i personally use: I use the siganl carrier size but to do that i render 50 fps (i live pal land) 720 576 frames, also 1 expression that slightly offsets film fit sideways each alternating frame. The reason i use full frames is that i want to be able to properly filter the images. i then interlace this footage to one 720 576 frame aqfter getting in the gamma correction and reducing the colors to safe ranges and blur (it makes wonders to the quality of your final image), after integration i blur slightly again. I them batch a scaled up reference footage for a square pixel size (time is almost no consideration is 0 as i do this during the night when i sleep anyway before is start the comp task). I use this square image as reference for all motion graphics animation on top. I then squish this to a layer of real interlaced size, integrate the plates with bg and color correct and add difference noise.

Not entirely straightforward but produces about 35% better image quality (more imporatantly ahs no interlace flicker thats really common to beginners) than not doing thisway