Thanks dude, it should turn into a very useful tool.

Easy: 1920 x 1080, and 1280 x 720. Both are 16:9 square pixel formats. 720 is always 720p; 1080 can be either 1080i or 1080p. Both can be various framerates, but you don't need to care.

@citizen:

This calculator's rockin'. I like the AviSynth code. Makes it almost impossible to screw up. If you're going to that level of detail already, would you consider adding the LanczosResize() line you'll need, as well? There's a new syntax for Lanczos that I haven't figured out yet. Even with the old one, I can never remember which numbers come in which order...

Blue skying a bit for a second:

Imagine a sort of "AviSynth for Dummies" automatic script-generating page, where a nearly-non-user of AviSynth could enter the dimensions of the source and desired output, click on a few checkboxes to pick stuff like sharpening or smoothing or saturation increases/decreases, and have a pretty much full-blown AviSynth script magically appear in a box at the bottom, ready for cutting and pasting.

I'm in no way suggesting you *should* or even *should want to* put together something like that. I'm just sayin' it would be cool, and it was inspired by your already-cool AR calculator makin' life simpler.

I think (without going to the trouble of checking...) that while Lanczos4Resize() is pretty new (v2.55, maybe?) that there's an even newer, more general and flexible syntax for v2.56 that will allow functionality like either LanczosResize() or Lanczos4Resize() or something more like LanczosNResize().

I say you go from where it's not black. That is, if your whole image fits in 704, then use 704 (or whatever the actual edges are). If it fits in 720 (like I hope most things do, eventually), then use that.

I could be missing the point; but if I'm not, then that's my 2¢.

I think I have a little problem with that. I know you've been looking at a super-technical page (from Finland) and have made adjustments accordingly, but I think I'd like to ask you to reconsider. Here is my rationale:

For a fixed-pixel display (such as a projector, which I don't own yet but will someday), the aspect ratio is fixed, and there is effectively no overscan. So the source material should ideally be full width (i.e. 720, say) with no black at all. Given that, then the height is the only variable that can affect the AR, and it does so in a non-CCIR way.

Once things are digital, then we have the flexibility to make sure round things stay round -- and I don't see a lot of hand-wringing about CCIR specs built in there. For example, many projectors are 854 x 480, which is 16:9 square pixel. It's clear that the source is meant to be DVD, with no overscan.

Now, for a real TV, getting the AR "right" is sorta pointless, because a CRT can be (and is) adjusted to mess with the AR, and the edges hidden by a bezel. So the minute differences made by CCIR-compliant calculations sorta make no difference at all, when some doofus at the factory wasn't all that careful, and your TV is actually displaying 4.2:3 already!

Given the additional fact that an increasing number of masters never actually exist as NTSC or PAL until they're converted to DVD (since they come via an HD master), and the people doing the conversion are probably computer nerds and not broadcast engineers, isn't it likely that a lot of the CCIR rules are sort of irrelevant, and therefore things are being done strictly on a straight math calculation (i.e. 720 x 480 is 3:2; figure out the PAR for NTSC at 4:3; convert on that basis)?

That's my take. Don't know if I'm right. Hope it contributes to the project, by at least making you think and/or explain why I'm out to lunch!