VT3 CG Edges

[kburd001]

Greetings!

I am loving VT3, but I have a question about the CG. When I try to use different colors for the inner and outer edges on text or on a box, I get a gradient from one color to the other across the span of the edge. I cannot find a setting that gives me two distinct, sharp lines instead of a gradient, and the manual is no help. Any ideas or experiences?

Kyle Burdette
PixelEyes Media Productions

[Gordon]

Just a guess; but I suspect that this is done to eliminate 'jaggies'. The edges are antialiased, (the gradient from one colour to the next), on purpose. There were a lot of complaints previously about the edges not looking so good and the only way to get good edges is to antialias them.

[kburd001]

Possibly, but the gradient is an even transition from one color to the next even when the edge size is cranked all the way up to 30. I know an edge size of 30 is unrealistic - I was just testing. This seems like a pretty significant limitation. I hope there is a way around it.

[bradl]

You have your edge... good. Want more options, ahhh, color gradient... good. What you want now is two distinct edges of the same type but different colors. I have not seen a CG package alone do this. I suppose you could save the file then edge again.

[prospector]

yes but is the edge crawl fixed in the new CG ?

[Gordon]

Kyle;
There is a way to do what you want. In fact two ways; Aura and Lightwave.

Aura has a border George script that you can call as many times as you want for as many different color edges as you want.

With Lightwave you can bevel edges with as many bevels as you want at any angle and give a different name to every edges with different properites, including color.

There really are no limitations, however, you will have to choose the right tool for the job. The learning curve for all the tools is a bit steep and long, but this is the price to pay for having all the power and options available.

[kburd001]

I have used LightWave in the past as you suggested, I was just hoping for a faster solution. Someone in an earlier reply said he didn't know of any CG program that did this, but the Power CG that was (is?) so popular because of the weakness in VT2 CG does this very easily.

[SBowie]

I apologize I've not been paying more attention to this thread - could you please post an example of the effect you're after?

[kburd001]

The issue is multiple edges on text. Power CG lets you control the color (and other parameters) of two different edges. For example, black text can have white edge, and a red edge can be place on the white edge.

I know that VT3 CG lets you assign a different color to the inner and outer edges also, but the colors of the edges blend into each other across the span of the edges, regardles of how big the edge is. I want to have 2 distinct edges with no gradient between them.

[Paul Lara]

Have you thought about using the shadow layer as your second colored edge...and just not use any offset?

[SBowie]

other options... (oops, hit 'send' to soon: see next post)

[SBowie]

Here are some easy alternatives:

[prospector]

your example #2 is what I think he is talking about, because I tried it here.

the red edge and the white and blue edges are merged into each other so that there is a gradient on the TV sets even tho it looks crisp on a computer.

so it goes from blue thru a cupla shades of purple to the red then a cupla shades of pink to white, rather than crisp clean edges.

[SBowie]

Originally posted by prospector
so it goes from blue thru a cupla shades of purple to the red then a cupla shades of pink to white, rather than crisp clean edges. Yep - I got the picture, thanks ... but only the CG does this. That's why I'm offering a couple simple aternatives that don't have that ... 'feature' ... ;)

[tfrank]

What Steve has shown us from Aura is neat, but I think that the VT2 CG still isn't up to snuff in quality and the ability to design looks. You should be able to do these "Aura" produced graphics and more within the CG itself and do it quickly. I come from a "brodcast" background, and the VT CG is just not there yet for something that is billed as a studio in a box. Don't get me wrong, I love out VT2, I just dislike the CG. I have not seen the CG in VT3, but I hope there is a greater ability to design "looks" with more handles within it's software. I also hope the quality is better.

[SBowie]

It's improved a lot in [3], but there's still room to grow ...

[Danner]

Something I noticed on the CG, if you are doing a roll of white text over a black background add a 2 pixel border consisting of a very dark grey to the text, it will look much smoother without the border being noticeable.

[Paul Lara]

Originally posted by prospector
your example #2 is what I think he is talking about, because I tried it here.

the red edge and the white and blue edges are merged into each other so that there is a gradient on the TV sets even tho it looks crisp on a computer.

so it goes from blue thru a cupla shades of purple to the red then a cupla shades of pink to white, rather than crisp clean edges.

That's because computer monitors and TVs are completely different beasts. You are always advised to gradiate from one distinct color to another. Otherwise, a client viewing sharply-defined edges and/or colors on a composite TV will see a ton of dot-crawl. This can is minimized when viewing the output in Y/C or Component, but will never look like 'a computer screen'.

[prospector]

Danner..
yep but better way and faster (don't have to make border, is to lower Alpha to 175-200

this seems to help on all cg.


PLARA..yep I know

but did find out that video is running at 72 pixel/in. As reported by Photoshop if a frame is brought into it.

If I take the video saved out to frames in Aura, then do an action in Photoshop to increase this to 300 and keep the frame size same

720X480

it makes for a much cleaner video..even with the above examples by Steve.

Is there a way to change the pixel count in Aura alone? I haven't found it yet.

And

because the VT is for Telivision (the end results in 99% of the work done on it), shouldn't the results ( or tools in the VT) be based on Telivision signals and not Computer ?

So that the best results come on the TV and not necessarily the computer monitor ?

[SBowie]

Originally posted by prospector
but did find out that video is running at 72 pixel/in. As reported by Photoshop if a frame is brought into it.

If I take the video saved out to frames in Aura, then do an action in Photoshop to increase this to 300 and keep the frame size same720X480 it makes for a much cleaner video.. Video applications, like Vt-Edit and Aura don't 'speak' dots per inch. Unlike Photoshop, which is really primarily a print application, they don't know dpi from dipsticks. It has no relevance, since all that video pays any attention to is the raster size. :-p

What is happening with your Photoshop action is that you are upscaling the image to roughly 3200 pixels, then bringing it back down. This effectively results in something akin to an anti-aliasing pass, making the result appear smoother. If you feel it's worth the trouble, render on :)

[prospector]

OK raster sizes only then

I am really really tryin ta understand this stuff and not just screwin around.

Live video recorded in the VT and brought into Aura for painting has rather large blocky graphics if lookin at edges of anything, and output shows this on TV screen of frame I am working on.

Yet the paint brushes in Aura don't, and have a clean edge, and this is also seen that way on the TV screen. So there is some kind of relationship there on pixel sizes.


So am I recording the video on the wrong settings and making the video pixely ? (recording as Newtek RTV )

There 'must' be something that will change it someplace that I am not seeing.


An example is a pic that is 72 dpi (720X480)with a diagional line 1 pixel wide and the same diag line on a pic that is 300 dpi (720X480), shows up much much cleaner when shown on the TV side by side (via pic placement in TED)


HMMM
Camera maby ?
using Hi-8 recording directly to HD and not to tape first.

[tfrank]

I would think that the camera input has something to do with it. You are dealing with a small imaging surface...perhaps a 1/4" at best...and depending on the camera...perhaps as low as 1 chip for all three colors. Even though you are looking at your video direct out of your camera, you are still only going to get 525 scan lines. The larger and more dense the multiple imaging surfaces, the better the quality, but your still limited to a TV sets scan rate. I've also seen small chip "HTDV" cameras that looked worse than a decent "NTSC" camera because of the lack of a sizeable imaging surface. Due to it's inherient technical & physical limitations "video" will almost always be lacking in that you can more readily see the pixels. Some people like though. What it really comes down to in my opinion is the final audience of the project. Small screen...large screen....whatever. You can drive yourself nuts trying to deal with something that no one else may never see. This may not answer your question, but I aplaud your search for knowledge.

[SBowie]

Originally posted by prospector

Live video recorded in the VT and brought into Aura for painting has rather large blocky graphics if lookin at edges of anything, and output shows this on TV screen of frame I am working on. Could be several reasons- first, DV is this way by nature, and if you are recording using a DV camera, that could well be why. Oh, I see below you're using Hi-8, so that's not it. Hmmm, depending on your Aura project mode, you may only be looking at a single field, with or without aspect correction (again, depending on your settings.)

Originally posted by prospector

An example is a pic that is 72 dpi (720X480)with a diagional line 1 pixel wide and the same diag line on a pic that is 300 dpi (720X480), shows up much much cleaner when shown on the TV side by side (via pic placement in TED)
... using Hi-8 recording directly to HD and not to tape first. It's hard for me to tell from what you wrote here what the final resolution of the 'better' image you are placing in TEd is, whether TEd is providing the downscaling or you are doing it in Photoshop.

Regardless: at the end of the day, on output that image is going to be 720x486. The scaling can be done in realtime in VT-Edit, or before hand in another app. One would expect an image processing application (such as Aura or P'shop) to do a better job of reducing it from the larger size than a 'realtime' scaling operation in VT-Edit, no matter how clever the latter is.

But either way, if you upscale then downscale, you're going to get a little blurring of the line compared to an original created at 720X486. This will always look better on video output than a single pixel line, since, as everyone knows, single pixels are anathema to video.

I think you better come over for a visit. I'll put the coffee on :)

[prospector]

my brain is melting from overload :eek:

Scanlines
single fields
aspect correction
DV
upscaling
downscaling


Now I find that Photoshop is lying to me:mad:

It goes from 72 dpi to 3200 dpi THEN back to 300 ???
I thought it just went to 300 from 72.

This is tooo much pressure !!



I guess what I am finding it hard to understand is..
if the VT board can output video to the TV in clean pictures from Aura as a tga or other format, in whatever dpi I use there (either Aura, Photoshop and pic size is same (720X480), and it looks clean and smooth..why cant it put out the video at the same resolution??

Arrggghh another word !!! resolution

Is it because the VT records it at low res?
can that be adjusted?

there HAS to be something that makes video frames look blocky and paint programs look clean, both being output to the same TV.

there just has to be....:(

[SBowie]

Originally posted by prospector

Now I find that Photoshop is lying to me:mad:

It goes from 72 dpi to 3200 dpi THEN back to 300 ??? No, please note: I didn't say 3200 dpi. Forget dpi, it has no relevance. Seriously. If you say "dpi" again I'm going to have to have someone go over and beat you. :-p

Here - let me try again:

For *print* purposes, some filetypes (and applications) support an embedded dpi setting. This is so that a suitable print program (and its user) can have some reference to the amount of detail can be squeezed out of a file in relation to print size in inches - but the inches have no bearing on what you will see on a TV, and they are hence meaningless, and justly ignored. If you're printing a magazine, you can talk about dpi.

In reality, a 'Photoshop image' at 900 X 600 @300dpi is NOT actually 900 pixel by X 600 pixels (resolution). It's really 2700x1800 pixels... the 'dpi' info is letting you know there's enough data there to print it at a 3" X 2" size at 300dpi quality. But you're not printing it, and the inches go out the window as soon as you show it on a different size TV set. Forget about the dpi, think in terms of raster size.


Originally posted by prospector

I guess what I am finding it hard to understand is..
if the VT board can output video to the TV in clean pictures from Aura as a tga or other format, in whatever dpi I use there (either Aura, Photoshop and pic size is same (720X480), and it looks clean and smooth..why cant it put out the video at the same resolution?? Your video, if recorded at 720x486i RTV quality, should look just fine - no, let me take that back - it should look superb. If it's 'blocky' there's something else going on.

[prospector]

WOO HOO I GET IT !!

What is viewed as image info isn't what is being broadcast to TV:D

K I forgot about _p_ ;)

so I am looking for something else.
Lemme see where this trail leads.

Thanks

[SBowie]

Originally posted by prospector
WOO HOO I GET IT !!Thanks WOO HOO indeed :)


so I am looking for something else.
Lemme see where this trail leads.
May I suggest a) starting a new thread, and b) posting a screenshot showing the blockiness that you're seeing.

[prospector]

HMMM
difficult at best
only seen on the TV

and only when shuttling frame by frame thru video (camera shots, not computer video)

In Aura, tried to load even first, odd first, and full frame...
samey same.

video looks ok on the Aura computer screen (well, for the area it takes on computer monitor as compared to the 32in TV it goes out to).

I try to take pic of TV but get scan lines running thru pic.:mad:

Video looks clean when running at realtime.

ONLY when doing a bluescreen (the other thread I had going) can the blockyness be seen clearly.

and that's what got me off on the _p_ theory. ( those letters you forbid me to type :D )

[SBowie]

Originally posted by prospector
HMMM
difficult at best ... only seen on the TV, and only when shuttling frame by frame thru video (camera shots, not computer video)...
_____________

ONLY when doing a bluescreen (the other thread I had going) can the blockyness be seen clearly. Hmmm...

Problems not seen in motion may not really be problems at all, but simply the nature of interlaced video. Anything with inter-field motion is going to show a 'comb-like' effect on high contrast vertical edges when viewed as a still. You can grab the output from your VT by capturing a single frame from Program Out if you want to show us what you are seeing. If you're not sure how to post it here, or how to convert it to a JPEG, you can e-mail me directly with a single frame RTV as an attachment.

Bluescreen problems are likely a separate matter.

[wvp]

If I may either A) Help clarify or B) totally screw up how far you have come:
I believe the the "quality" differences you are referring to have to do with different formats. lemme splan.
If you are working with a still image (say a picture taken with a good quality digital still camera) and you are comparing this or an image that you electronically created with Photoshop or Aura to a still of a video image, then you cannot compare these!
As Steve mentioned, one of the big reasons is video is interlaced while the computer screen is not.
If you START with a non-video image THEN output this image to a video monitor it will look better than an image captured via a video camera because it is (most likely) an image that started out at a higher resolution than video.
While the rtv format is lossless (does not loose quality) it cannot give you a higher quality than what you started with.
If you compare a video still shot with a cheep 1 chip camera vs. a 3 chip pro camera you will see the 3chip image looks better. The rtv output is still the same image size/quality BUT the incomming image is much better with the 3chip camera. Hope this helps!

[bradl]

...the old GI-GO maxim...

[prospector]

so both have mentioned the camera now and that's probably what is going on.

With the Amiga Toaster/Flyer I used just VHS for filming and that was fine with just video.

Then I went to Hi-8 with the VT1 and that was fine with just video and lots better than the VHS.

But now I am getting into bluescreening and the blockyness of the video is REALLY showing up.

I took a frame from the video that comes with the VT and it looked much better than mine for resolution.


So i guess that now I get to the REALLY hard part....
explaining to wife how I gotta get ANOTHER set of cameras....:(

Thanks guys for setting me straight on this aspect of video, as I didn't run into it till I got to bluescreen stuff.


Any camera suggestions that won't get me hung :confused:

Like 3/4" betacams will get me past the hanging stage for sure.