Aspact Ratio

[biferi]

OK I made all my JPGs 854x480 Pixels.

Then made my DVD and it came out great and filled the whole screen.

Just to see if I get what you told me about NTSC DV Video.
If an Image is 4:3 just the Pixels in the 720 Row Left to Right will be close to the Pixels next to it?

And in Widscreen the Pixels in the 720 Row Left to Right will be Farther away from the next Pixel next to it?

PS....should I Delete this Tompic?

[_Al_]

That's unfortunate how to think about anamorphic video. And it seams not helping you. If you say pixel are closer or further then that would implicate that something is missing in between them.

Depends how you look at it.
For widescreen, pixels would seem to be actually closer together and stretched only while viewing.
Yes, they would seam to be further away from each other for widescreen if you imagine final viewing scene. But wider angle is involved with widescreen at the same time.

What software are you using to get those images?

You had created a DVD that is proportionally correct, it works for you, but anyway, maybe even for others, there is a way to resize only one time, directly in software like Photoshop, usually you have to even zoom into photos and graphics, or designe an image directly in anamorphic DV image. No resize is involved all the way to making a DVD. With computer graphics it is quite important.

You are making a DVD. There is a way to make 720x480 anamorphic image directly in software like Photoshop, perhaps other softwares. Photoshop has DV NTSC (or widescreen project). It looks like square pixel while working on it, not 720x480, but you export BMP for example and you get 720x480.

Then you load it in Vegas, NTSC DV project (or widescreen project), put it on the timeline, click pan/crop and select "Maintain Aspect Ratio" to "NO". Because Photoshop already did that. So Vegas just fills the screen. No resize is involved while exporting mpg for DVD or making DVD directly.

[biferi]

Can I please ask if there is a way to Delete this Topic?
I want to start over and I do not want to keep Adding more to this Topic.

I do not want to get anybody Pissed off at me.

[davexnet]

Can I please ask if there is a way to Delete this Topic?
I want to start over and I do not want to keep Adding more to this Topic.

I do not want to get anybody Pissed off at me.

Topics are not usually not deleted because the contents may be of use to somebody in the future
searching for similar information ...

Just abandon this one and start your new topic

[biferi]

I thought I had it?

You told me 4:3 Digital SD is 720x480.
And 16:9 Digital is also 720x480.

But the Pixels for a Widescreen SD Video are farther away from the Pixel next to it.
So the 720 Wide Pixels are spread out.

But then you say thats not it because you would have Gapes inbetween.

[davexnet]

I thought I had it?

You told me 4:3 Digital SD is 720x480.
And 16:9 Digital is also 720x480.

But the Pixels for a Widescreen SD Video are farther away from the Pixel next to it.
So the 720 Wide Pixels are spread out.

But then you say thats not it because you would have Gapes inbetween.

It's just conceptual a way to explain it, that's all

[Cornucopia]

Let's pretend you have a house.
A round, small house. So do your neighbors.
You also each have a yard around them, with fences separating them all in a grid layout.

Your yard is very reflective so it mirrors the color of your house, as does those other yards for their houses (maybe they're used as solar cell farms).

The spacing/shape of the yards can be for example 20 meters by 20 meters and be square, or they can be 20 meters by 10 meters and be wide rectangles, or they can be 10 meters by 20 meters and be narrow rectangles.

As seen from the sky/space, it looks like different colored rectangles/squares, even though the color is really only coming from the small round houses in their centers.

In the NTSC normal neighborhoods, you only see real estate with narrow yards.
In the NTSC wide neighborhoods, you only see real estate with wide yards.
You never see NTSC square-yarded neighborhoods, because that isn't allowed in the city charter.
But if you go to the next city - Computerville - all kinds of neighborhoods are allowed, though the square-yarded ones are most common.
Or if you go the other direction to HDtropolis, almost all the yards are square, except a few of those expatriots from NTSC, but they always congregate in their own neighborhoods.

Scott

[biferi]

So then am I right that a DV Widescreen 16:9 video the space around the Pixels changes shape?

So just the space around the Piline xels in the 720 will be wider?
I get this if you want to make a Pixel look wider then high you just move the Pixel away from the one next to it.

Or is this not how they do it?

[davexnet]

How they do it is a technical issue depending on the specific playback equipment. Does it really matter
how exactly it's done?

[biferi]

I am so so sorry I see what is going on.

I am just talking about Digital Cameras and how they Record Widscreen.

To get a Widscreen 16:9 image on the CCD the Pixels that will makeup the 720 Line will be what captures the Wideth.
So the CCD will turn some of the Pixels OFF in the 720 Line to make it Record Wider.

Is this it??

[Cornucopia]

All real-life cameras actually do have real-world sensor wells with defined shapes and (empty) spaces between them. And all displays have real-world phosphors/crystal-filters/etc arranged in a grid with surrounding black masking.
But - here's the important part - IT DOESN'T MATTER either way (fatter-shaped sensors, or wider-spaced sensors, or summed/combined neighbor sensors). They will all give a nearly IDENTICAL resulting image (actual slight differences in noise, dynamic range & chroma saturation/accuracy aside), due to the mathematics of sampling and of the specifics of human perception.

If you didn't get that analogy, I think you're going to have a hard time getting further unstuck. Reread what has been written. Then move on to some other problem which hasn't been solved.

Scott

[_Al_]

I am so so sorry I see what is going on.

I am just talking about Digital Cameras and how they Record Widscreen.

To get a Widscreen 16:9 image on the CCD the Pixels that will makeup the 720 Line will be what captures the Wideth.
So the CCD will turn some of the Pixels OFF in the 720 Line to make it Record Wider.

Is this it??

Older CCD digital DV camcorders?
Imagine using wide angle lens that " funnels" wider image into "older" 4:3 sensor. Distortions could be introduced in the corners. So for example you cannot compare 4:3 to 16:9 because in real world you cannot compare it. if shooting that way. To get the same width you'd have to step closer with wide lens. Hope it helps explaining like this.

16:9 DVD's are made from film (digitized into square pixel) or digital camcorders of higher resolution, in both cases downresize is involved. Very different. Again, you cannot compare 4:3 vs. 16:9, 4:3 versions are pan/crop anyway.

[biferi]

OK I went back and Read the Posts you gave me. image are more wide we meen the space

And Both DV 4:3 and DV 16:9 use 720x480 RES.

Now if Pixels can not change shape they have to make them seam Wider then High.
So the space between them is widened Right?

But just the Pixels from Left to Right?

And Pixels are the same shape so wen we say Pixels in a 16

[jagabo]

For purposes of aspect ratios it doesn't really matter if you think of pixels as rectangles or points with spaces between them. When the DAR is wider than the SAR the pixels (or spaces between them) are rendered wider than they are tall. When the DAR is narrower than the SAR pixels (or spaces between the points) are rendered narrower than they are tall.

[biferi]

ok I think I get this part.

When they Record DV 16:9 and DV 4:3 they both use a RES. of 720x480.

But when they Record 16:9 the 720 Wide Pixels are the ones encoded as 1.212 am I Right?

[_Al_]

nothing gets encoded as 1.2121, it is encoded as 720x480, encoder does not care if it is 4:3 or 16:9, it just analyzes 720x480
EDIT: aspect ratio is included in container or in stream as a flag only after encoding, it is there but encoder does not need that information while encoding video

pixels are pixels, no size, no shape, you need to forget this at this point, you seam to be stuck right here,

You should actually start first what is happening when DV is shown on screen. The math. That only matters. For example you have CRT and viewing 720x480 16:9 DVD. So player comes up with letterbox. That is familiar for you. Black top and bottom, renders all available pixels horizontally, that is 720. So you see, nothing is missing right there.

How many vertical lines player comes up with? How many lines it will render on CRT in this case?

[biferi]

OK so any Digital Video 4:3 and 16:9 both are Recorded at 720x480 they just put a Flag to tell the DVD Player what Aspact Ratio to Display it back at.

So say you do have a Widescreen TV.

If you play a 16:9 DVD I get that the Flag tells the DVD player how to keep the Aspac Ratio.

Now a Pixel is a Point Of Light and are some what Rounded.
Now to keep the Video 16:9 it must Display the 720 Pixels with a space of 1.2121 mm.

is this right?

[_Al_]

How many vertical lines player comes up with? How many lines it will render on CRT in this case?

You first, how many?

[jagabo]

If you play a 16:9 DVD I get that the Flag tells the DVD player how to keep the Aspac Ratio.

In the case of DVD the DAR is specified in the MPEG 2 data. Either 4:3 or 16:9.

Now a Pixel is a Point Of Light and are some what Rounded.

A theoretical pixel has no shape at all. It is not round. It is not square. It is not rectangular. It is a dimensionless point.

Now to keep the Video 16:9 it must Display the 720 Pixels with a space of 1.2121 mm.

Only if the vertical spacing is 1.0000mm. A small TV might have 0.5 mm vertically so it will space the pixels 0.606 mm horizontally. A big TV might space the pixels 2mm vertically so the horizontal spacing will be 2.4242mm horizontally.

In practice, of course, the TV puts something between those dots. No modern TV fills the area with constant intensity (which would literally give rectangular pixels). They interpolate in some fashion. This is what scaling algorithms are all about. Even an old fashioned analog player/TV will use low pass filters to smooth the image.

[_Al_]

Math is needed here, because imagination does not work. To get rid of idea that there is a room between pixels or that they have dimension.

above mentioned number 1.2121 (40/33) seams to center of the problem, because Vegas shows that info for you, and even messing up with the things, because it is ITU 704x480 based, Vegas is set for ITU not mpeg2 specs 720x480. That number would be 1.1852 (32/27) for mpeg2 specs. If you just load NTSC DV and export for DVD it is not involved, so peace. And this proves that again that coefficient is complicated to understand by you somehow but in your case and workflow Vegas does not even use it. Only if loading square pixel JPG. You were told what to do in that case though.

1.2121 (ITU specs) or 1.1852 (mpeg2 specs) or and 16:9 ARE THE SAME THING basically, they are information how to get correct proportion on screen, you need one of them only

to make it simple for that 16:9 DVD for example :

720 rendered on CRT ...... unknown height
16 ....................... 9

unknown height=720x9/16=405
So player most likely renders 406 (rounded up to even number) horizontal lines for widescreen DVD if movie is 16:9
480-406=74
So on CRT, there is black bar 37 pixel high on the top and bottom

1.2121 or 1.1852 was not involved here for calculations, no pixel dimension was needed or calculated, that 16/9 ratio takes care of it

If you need to use that 1.1852 or 32/27 and forget about 16:9 info:
480/(32/27)=405

So that 1.1852 (DVD does not work with 1.2121) or 16:9 is the same thing to get correct proportion on screen.

to render that DVD on CRT player needed to interpolate only vertical resolution, that height, it needed to calculate 406 pixels out of 480. It interpolates it. He needs to get rid of some pixels and perhaps change color of other pixels right next to that one that was deleted to preserve edges, keeping overall color effect, whatever is in TV settings. So you see it is about interpolation, making up completely new pixels, pixel dimensions and shapes are not involved

If you play that DVD on you big screen say full HD TV 1920x1080, then again, this is calculated in a fraction of a second by player and it interpolates like this for every video frame:
Player knows target TV is 16:9 (1920/1080) and DVD flag is 16:9 as well, so all it does is interpolate 480 into 1080 and 720 to 1920. Because aspect ratio is the same. It does not care what shape or distortion it was stored or how weird it looks. No pixel dimension or shape is involved.

Of course real world algorithms are different, this just shows a principle.

[biferi]

Is this part Right?

Say you have a Standard NTSC 16:9 DVD and when you watch it the Movie is in LatterBox.

I do know what the Bars are for and the DVD has a Display Aspact Ratio of 720x480.

Now I think I have one thing down.
Even the Black Bars on the Top and Bottum will be part of the 480 Lines.

Is this Right just for what I think on this Part?

[manono]

...and the DVD has a Display Aspact Ratio of 720x480.

No, the DAR is 16:9, which is meant to tell the DVD player how to resize the 720x480 video upon playback. And the word is 'aspect'. I'm getting tired of seeing you butcher the word post after post.

Even the Black Bars on the Top and Bottum will be part of the 480 Lines.

If watching on a 'widescreen' television set, yes. Not if watching on an old-fashioned 'square' television set.

If I were you I'd forget about trying to figure all this out. You don't seem to be any closer to understanding it than you were more than two weeks and 51 posts ago.

[_Al_]

You were told that 406 lines is interpolated from 480 on CRT (16:9 DVD). Rest of those pixels (74 top and 74 bottom) is generated black be player to fill the screen with black nothing.
EDIT: 74 together, top and bottom

So you ask about it right next question.

[biferi]

Well I do understand onething.

If I am taking Video at 720x480 and then I do my Edit to it.
If I want to Export it as 16:9 Widescreen I pick Pixel Aspect Ratio of 1.2121:1

I do have this right??

[Cornucopia]

If when you export, your resulting file is ALSO supposed to be 720x480 AND 16:9DAR, then your PAR should be near 1.21212121...
If you were using ITU settings, it would be EXACTLY 1.21212121...
For MPEG2 settings, it is exactly 1.185185185...

But you are correct enough to have gotten your setting choice right.

Now, could you please change the title of your subject to read "Aspect Ratio"? The misspelling is really starting to bother me.

Scott

[biferi]

When they made CRT TVs the Pixel shape was like a Diamond.

But then Computers came out and they can not see Diamond Shaped Pixels.
So when they made the CRT Screen for the Computers the Pixels had to be Square so they just put the 3. Phosphor colors R G B right next to one another and this is how they made a Square Pixel?

PS....I know all Pixels have a Black Mask.
Rhanks

[Bernix]

I know bayer is using on chip blue, green, next row is green, red. So whole 1/2 is green and rest is 1/4 red + 1/4 blue. But it is case of photographic chip to keep square...

[Cornucopia]

No, when they made CRT tvs, the pixel shape (or more rightly phosphor shape because analog is CONTINUOUS and there is no such thing as pixels until you get to digital) was a number of different types.
First round triangularly laid out, then hex (closer spacing less mask), then vertically long rectangles of various dimension ratios (incl. Trinitron), then some staggered rectangles, etc.

Scott

[biferi]

Now as far as if I make a JPEG and I want it to be 16:9 Widescreen I thought it as 720x540 Pix.

But I know this is not right.

[manono]

Now as far as if I make a JPEG and I want it to be 16:9 Widescreen I thought it as 720x540 Pix.

I thought they spoke English in the US? I thought everyone went to school and learned basic mathematics in the US? Apparently not.

16/9=1.7777...
720/540=1.3333....

If you want your video to be 1.7778:1, then give it a 1.7778:1 resolution. 640x360, ~854x480, 1024x576, 1280x720, 1920x1080.

This ignores any SAR a video might have to resize it upon playback. It's only for square-pixel videos.