Rendering

[Renton Maclachlan]

Just a question about rendering: Is there any loss of video quality when a video is rendered, or rendered a number of times? Is there a difference in such situations between rendering a video without any effects, as opposed to one with effects?

[Edward Troxel]

How about yes, no, and maybe?

You are going to need to be MUCH more specific in order to get the correct answer. So start by answering some questions:

1. What is your source format(s)?

2. What format(s) are you rendering to?

3. Are these for pre-renders? Or final renders?

4. What is your ultimate destination for your video?

[Seth Bloombaum]

Edward is, of course, correct, and asking the right questions.

Another way to look at it is in terms of designing workflows.

Acquisition Codec
Something that comes from your camera will (almost always) be compressed with a lossy codec. However, the manufacturers have carefully selected codecs and datarates that allow lowest manufacturing costs and highest picture quality.

Intermediate Codec aka Digital Intermediate
If your NLE doesn't deal well with your camera codec, or, if you're needing to do some prerenders to bring scenes over to After Effects, or maybe to a specialty encoder to create a delivery format Vegas doesn't directly support (flash comes to mind...), you want a codec/bitrate that preserves all possible quality, but doesn't bring your system to its knees or tie it up for a couple days. Many Vegas editors are using Cineform AVI, Lagarith AVI or DNxHD QT. It's more complicated if you're going to another NLE or platform...

Delivery Codec
This matches the delivery system - it could be anything from a 50Mbps MPEG2 TS to a 0.5Mb wmv. The large one could be fine for basing further work on, the small one is too compressed to base any future work on. A 2Mbps MP4 for youtube would also count as a delivery codec...

[Renton Maclachlan]

Hmmm...I've done the filming - Panasonic GS400 and GS500 (four cameras) - which I understand is uncompressed avi - on MiniDV tapes.

I've downloaded the footage without editing using the Panasonic supplied program SweetMovieLife, which I understand brings the footage straight over into the computer as it is on the camera.

I've now got two situations:
1. I've done some editing - on all the footage I've simply cut some sections out of the footage.
2. On several clips I've had to do very slight cropping as well because the edge of the green screen was showing. (Camera screen shows a smaller area than what is captured. :-( )

I understand that any cropping will lower resolution when rendered, but will the rendering itself lower quality in both cases?

I'm thinking of rendering out these initial edits bcause I have a reasonable amount of additional work to do on the films - addition of background to replace greenscreen, addition of text and multiple images, and effects, etc, all of which will gobble up computing power.

[Edward Troxel]

Actually it's "DV" which *IS* compressed. However, here's the good news:

Render DV to DV with NO effects - it's a straight copy so ZERO loss

Render DV to DV with effects - you will be VERY happy with the results. Vegas has probably the best DV codec out there. You can render MANY generations before seeing any significant loss.

[Don Bloom]

100% agree with Edward. I did an edit for someone recently and the tape had multiple video drop outs. I used Excalibur "fix bad frame" script so many time my fingers got sore but the point is I had to render the portion many times, like 6 or 7 to get all of the drops taken care of (at least as many as I could) and there was little if any generational loss.

[Seth Bloombaum]

You're getting good advice here!

In the terms outlined above, DV/AVI is one of the few codecs that can serve in the camera and as a digital intermediate. Several years back there was some testing done that showed 11 generations or more of DV was visually indistinguishable from the original.

I'm remembering (misremembering?) Edward might have been involved in that testing?

Originally Posted by Renton Maclachlan

...2. On several clips I've had to do very slight cropping as well because the edge of the green screen was showing. (Camera screen shows a smaller area than what is captured. :-( )

I understand that any cropping will lower resolution when rendered...

Don't crop for this problem, use a garbage matte. This matte is an important tool with greenscreen work!

You'll find it in the pan/crop window, down in the timeline area, click the "Mask" checkbox and make sure that header has the highlight, not the Position timeline header. From there, you draw some points (hard to describe, check the vegas help under pan/crop | create a bezier mask) , to make transparent large areas of your clip, including the garbage out to the sides of your greenscreen.

The advantage of this method is that no rescaling of the image is done, which means less work/faster rendering and no quality loss. To be fair, such loss is fairly minimal when done well, but why rescale unless you really want to?

[David Jimerson]

Be sure to do the garbage mask before you apply any FX, else you'll see a faint outline of the mask.

(This can be dealt with in another way, but that's harder to describe without a visual reference and applying the mask before FX is the simplest way to avoid it.)

[Renton Maclachlan]

Thanks Seth for the idea of the garbage matte. I don't know anything about this so will swot up on it. I've already cropped two clips so I guess I can leave them and if any others need something done I could use the GM...

[Renton Maclachlan]

Originally Posted by Seth Bloombaum

Don't crop for this problem, use a garbage matte. This matte is an important tool with greenscreen work!

You'll find it in the pan/crop window, down in the timeline area, click the "Mask" checkbox and make sure that header has the highlight, not the Position timeline header. From there, you draw some points (hard to describe, check the vegas help under pan/crop | create a bezier mask) , to make transparent large areas of your clip, including the garbage out to the sides of your greenscreen.

The advantage of this method is that no rescaling of the image is done, which means less work/faster rendering and no quality loss. To be fair, such loss is fairly minimal when done well, but why rescale unless you really want to?

Just reactivating this thread from back in July...and in particular am asking Seth if you would clarify this question of a garbage matte to remove some bad greenscreen stuff.

I have another clip which has the corner/edge of the greenscreen showing - [camera LCD does not show everything that is caught by the camera!!!]

I can't get a handle on how a garbage matte will achieve what is hoped to be achieved, that is a consistent green out to the edge of the shot. Last time I pan/cropped the image but you said that I shouldn't do that and rather should use a garbage matte. I didn't follow through on it all last time as I had solved the problem by cropping. But this time I want to use the GM as you suggested but don't understand the procedure...

[Mike Kujbida]

Renton, here are a few screen shots that I hope will clarify things for you.
I didn't light this but had to work with the footage afterwards :(
I used the Mask feature (3 masks in all) in Pan/Crop to get rid of the areas I didn't want to see.
Take note of the mask settings in the Path menu.
He didn't move around so I could set my masks very close to him.
This wasn't necessary but I wanted to show you that you can get away with a very small green screen (this one was 10 ft. x 12 ft.) and still get good results.
Considering what I had to work with, I was pleased with how it turned out.

[Seth Bloombaum]

Mike's illustration is excellent. Of course, more often, you're drawing 4 points to make a large box. But, as Mike shows, bezier masking tools allow you to make a shape of any complexity, and they can even be keyframed to follow an object or person. I did that on a project a couple months ago - with amazing results - but very time-consuming.

What is outside the box does not "turn green", for later chromakey, it becomes transparent. All that dark-green area of M's P/C window is actually transparent on the timeline.

Note also David J's post #8, above. It's an important point. Recent versions of Vegas allow you to specify not only the order of video fx application, but also, where in the chain pan/crop is applied. To change this order, open up the video fx for the clip, and drag the pan/crop button to the left to place it earlier in the order of execution.

[Renton Maclachlan]

Thanks guys...

So...rather than make the whole background green, which would then be extracted and made transparent with the chromakey tool, all outside the mask is non-selectively extracted to obtain the transparency needed, and then what's left of the green is extracted as a separate process via the chromakeyer?

Do you do two separate renders, first to get the result from the garbage matte and then that from the chromakeyer, or do them both at once.

If there is unevenness of lighting across the screen, this could aid in keying, couldn't it, by reducing the variation that needs to be selected by the keyer? I've got a pretty good screen but I've noticed that some times it doesn't all extract evenly/equally...

[Edward Troxel]

One render is fine. You can do both the chroma key and the pan/crop mask at the same time.

[Seth Bloombaum]

Originally Posted by Renton Maclachlan

So...rather than make the whole background green, which would then be extracted and made transparent with the chromakey tool, all outside the mask is non-selectively extracted to obtain the transparency needed, and then what's left of the green is extracted as a separate process via the chromakeyer?

That's right. As Edward points out, you typically would only render once, which would be your final composite, to include your foreground subject over your background images or clips.

Originally Posted by Renton Maclachlan

If there is unevenness of lighting across the screen, this could aid in keying, couldn't it, by reducing the variation that needs to be selected by the keyer? I've got a pretty good screen but I've noticed that some times it doesn't all extract evenly/equally...

Uneven screen lighting is the bane of chromakey compositing, without a doubt. It's always better to fix the lighting rather than fixing it in post. Granted, as you get better tools, and better with the tools, you'll be able to key some of those poorly-lit clips, but even then you'll be wishing the piece was shot better.