This forum definitely rocks... Iots of interesting discussions and little tips/tricks for Vegas.

I think this is most concise way of explaining color banding I have heard.
In the case of the gradient map filter, I believe the banding results from how they programmed the filter (may be some sort of performance tweak).





Well it mostly depends on what the codec is doing. For broadcast, I just kind of made a blanket statement there. But chances are the codec expects 16-235.

In terms of allowable colors, DVD has a greater range of colors possible than broadcast. Transmission limits means you should keep colors broadcast safe. When the digital signals get converted to analog, the composite voltage shouldn't go too high. It's hard to get information on this, and it seems to vary from broadcaster to broadcaster. Usually around composite voltage peaking at no more than 110-115IRE is what you should aim for.

For NTSC:
The main signal contains luma, which is like the black and white part of a signal.
The color subcarrier rides on top of that. It contains color as the name suggests.
Composite is when both signals are combined.

With the NTSC encoding scheme, 1.33X the maximum luma voltage is the limit of the composite signal. With 7.5 IRE setup, this drops a few IRE. However, to cram more channels into the broadcast spectrum, the engineers decided to limit the maximum composite voltage. This is about 1.2X the maximum luma voltage.

In old TV sets, the decoders aren't that great so really high composite voltages will bleed into the audio. So, 1.15X is a more practical limit.

*This is according to Charles Poynton's book (poynton.com). This is what I understand of it anyways... it is fairly technical.

If you look at the sky near the horizon, as it's getting closer to dusk for example, you'll see a beautiful gradient.

If you shoot a video of this with 8 bits per color, you can reproduce 256 different levels in each color, which for a mostly monochrome color means just 256 colors total.

This can show as stair steps or bands in the sky, and can sometimes be even more visible in skin tones.

By going to 10-bit, you can reproduce 1024 levels per color (R,G,B). This is not unlimited either, but in practice it has a visible effect of reduced or eliminated banding.

As Glenn indicated, the cheap workaround for lower bit depths is to use dithering.

Worst case this looks like Windows in Safe Mode. Best case it can be very decent.

Of course, if imaging math is always done at 8-bit for speed, then even dithering won't help (because the output can only have 256 steps regardless of how those steps would have been calculated if a higher precision had been used).

When you print a piece of paper in a laser printer for example, you only have black toner. So how do you get the shades of gray?

You print dots (that can be different shapes) fairly close together. The closer the dots, the darker the gray.

Ditto for newspapers, magazines, etc. Kinda works :O).

The difference between pure gray and dithered gray becomes more apparent when you get close to the piece of paper. Then the pure gray stays pure gray, while the dithered gray looks like an ant colony.

In video, this means that dithering isn't likely to be much of a problem on a 27" TV, but on a very large screen, it can start looking like Windows in Safe Mode. If you've ever seen DV footage projected onto a full size theater screen, you may have seen this effect (at least if your eyeglass prescription is current :O).

So for inkjets, with 3 colors you should be able to reproduce any color, right? No, if you mix all three equally, you get a muddy gray-black. So you add a fourth color, pure black.

So that's it, finally? No, as you have probably noticed, the world is full of 6-color inkjets and beyond.

In offset printing there is the Hexachrome process that looks reeeeaaally good for nature photos, for example. It can reproduce colors that ordinary 4-color printing could only dream about.

And there is plenty more beyond even that. Color is a huge field of science (and engineering)...

bigsole,

Aw give me a hug, you big softie.

tee hee

1- If the material already has noise in it, then dithering isn't necessary.

2- What I wrote about the Gradient Map filter may be wrong. I think I remembered something wrong or erroneous.

Well I just moved to a new company... Vision2Media and they are strictly Avid. They have 1 Media Composer & a new Adrenaline. Although I feel that I had to learn Avid to get decent employment, not many VEGAS jobs around, I absolutely hate Avid's workflow. Neither of these Avid systems offer me any more 'Real Time' performance than Vegas. I am yet to capture & edit SD in Vegas though.

I always hear my boss praising Avid. Everytime it just sounds like, "I paid $50K for my Avid." That seems to be the only BIG point that makes Avid better. I would much rather pay only $10K for a system that will give me the same, if not better, results. -He already stated that my DV work coming from VEGAS looked cleaner than his SD work from the Avid. I'm still yet to get a clean export from either Avids.

-For some reason he keeps calling it Video VEGAS. I hate hearing that. Everytime I correct him and say, "It's Sony VEGAS."

Now if I could only figure a way to get a VEGAS system in there so he can see what it's capable of.