Want to learn a few editing shortcuts to cut post-production time?
Look no further. In this section we have highlighted a few areas to help you out with any of your issues and concerns while editing on an Avid system.

Overview:
Keying
Secondary Colour Correction
Colour correction
What is a Vectorscope?
What is the Y-Waveform?
What is the White Zone?
Fixing that Image!
Useful functions under the HSL
Keyframes
Trim to fill

Keying

Keying is what we call it when we select a colour range and make it transparent.
In the film post-production world, keying is the effect that allows blue/green screen to work - by telling the program that we want the green section of the image's colour range to be transparent instead of green, we can superimpose it on a background.

This is lots of fun. It is a technique that has been in use for decades (with varying degrees of success) and recently has allowed films like Sin City to be filmed entirely on indoor sets with totally 3D generated backdrops, letting filmmakers place the actors in any kind of unreal, over-the-top background they want.
The way that we access this effect is by entering the "Key" section under the effects menu. Here we have a few options - Chroma Key, Luma Key, Matte Key and RGB Keyer. Today for the sake of brevity, it'll just be looking at Chroma Key, because it's the best example of what I'm trying to show.

When you open up the chroma key effect window, you'll be presented with a range of interesting options. The one that we're most interested in however is the one that looks least like an option. That's the small coloured box in the top right of the menu. If you move your mouse over this, you can see that it becomes an eye dropper.



With this tool, we can click on the coloured box, and then drag the cursor to wherever we like on our source or record windows and select a colour to exclude. Now if you take a quick look at my timeline here:



You can see I've imported a still image of a blue-to-white-to blue gradient, which I've put on V2 over V1 which contains some footage of some cars on a starting grid.
Once the effect is applied, and we've dragged that point in the effect editor over to a colour we've chosen (in this case I'm using the darkest blue) we can start to play with the "Gain" option.
In this case, Gain repesents the degrees of colour change which the effect will accept as the range of colours it wants to make transparent. The next 3 pictures are of what happens to the image with the gain at 0%, 50% and 100%.



So, in picture one we have excluded no colours, although by picture two we've gotten rid of all of the darker blues and a lot of the lighter ones. By the third, we've excluded everything blue, white, and even the black bars at the top and bottom of the image - However, if I'd had some Magenta in there we'd still be able to see it, because it's just too far away from blue in the colour spectrum for it to be removed.

This is the main lesson - if you're going to try and key things out, then make sure there are no other instances of it or anything like it in your shot! Unless you've got access to some top quality lighting equipment and the skills to set it up, you're going to end up with varying gradients even in your "solid colour" background and that means that it's just not worth trying. I've seen people try by just hanging a blanket in the background.
It can be a very powerful effect when applied well, but it's often altogether too easy to do it badly, so be careful!! Keying out hair is really difficult.

Secondary Colour Correction

Secondary Colour Correction is the effect that gave you Sin City, Pleasantville and Schindlers List. For those of you who haven't seen any of those: It's about choosing a single range of colours and working with them separately. I've been asked about this a lot of times - "How can we make this bit in colour and have everything else in black and white?"

Well, theoretically you would select that bright blue t-shirt and then invert that selection and then pull down the saturation on the image, but this isn't a feature that's available in Xpress Pro or Media Composer - it's a really advanced function but if you really, really want to do it then there are workarounds, and a lot of dedicated effects programs will do it too (Avid Effects for example) so get your Quicktime Reference codecs ready . The most important thing when you're doing this however is to set up your shot right.
This means - if you want everything but the blue shirt to be black and white, make sure that there are no other blue things in the shot - they'll be changed too. If you can, do it the way Sin City did it and have fully CG backgrounds, that way they're a separate layer anyway and you can work with them seperately (this is a rare luxury though).

Remember also that the colour selection process is tricky, especially with clothes and things because the light will be different across the folds and the resulting colour spectrum you'll be selecting could be quite broad, so if you've got a pink carpet lit from one side and a red balloon, it could well be that in changing the carpet you'll get into the lighter spectrum of the balloon and change that too.

If you're very, very keen and you have a really good shot that you've set up carefully then there is a way…
If you have the exact same clips on V1 and V2, in exactly the same place, then you can adjust the entire of V1 and then Chroma Key through V2 so that for those colours you want to change, you are looking at V1.
So - You've got a clip of a green field with a blue sky, and you want to adjust the saturation of the blue sky without changing the saturation of the grass. You put the same clip on V1 and V2, edit the saturation of V1 for the whole image, and then key out the blue on V2.

Doing it this way is tricky and if the shot moves you'll have to involve 100% PiPs which crop based on shot movement. It's really hard and I recommend you don't do it, but I know many people would like to know the answer.
If you have access to a Media Composer, take a look at the AniMatte effect. - it's very clever.

Colour correction

There's than enough to talk about on the subject of colour correction for a whole blog of it's own, so I'm going to tackle various different ways of using the colour correction tools in a more simple, effect-like fashion. I'm going to introduce a few of the measurement tools included in the software too.
The actual task we're going to be completing is a white-balance correction, without using the "colour correction" effect - it works OK, but this is better. White balance, for those who don't know, is a constant consideration when you're filming. A very distilled version of it's definition is: "The colours that your camera sees". A (much) more detailed explanation can be found here, although you'll get by if you just read the section entitled "Film Photography".
Regardless, if you haven't set your camera up right, you can end up with some nasty looking images like this:



As you can see, it's blue. Very blue. However it is not beyond saving. Fortunately, there is an even amount of too-much-blue in the whole image. This means that if we have some reference points, we can tell the system how much blue should be removed.
Our salvation lies with Luma. Luma is the part of the image's signal which controls light strength and therefore the greyscale. Because we know what black, white and grey should look like, we can find parts of the image which should be those shades and tell the system.
Then comes the clever part - the system will see that there is X amount of blue in the luma, work out how much blue as an average of the three (B, W & G) needs to be removed and take it out of the entire image.

Before I tell you exactly how to do this, I'm going to give you a quick run-down of some of the tools within Colour Correction. If you go into Colour Correction mode you'll get a window which looks similar to this:



If it doesn't look exactly like this, don’t worry - you can change what's being shown to you very easily by clicking on the titles - for today, set yours up like mine with a Vectorscope, a "current", a Y-Waveform and the Hue Offsets under HSL.

What is a Vectorscope?

A Vectorscope is a clever thing indeed, although will usually look quite nasty. All those green dots represent each and every colour used in your image. If you scrub around a bit on your "Current" window, you'll see how they update. There is reference text on the vectorscope to tell you what region is what colour (it says YL, R, MG, B, CY & G, for Yellow, Red, Magenta, Blue, Cyan and Green) but the colours are organised in the same way as the Hue Offset colour wheels, so it's easy to just look at those and see for now.
In my example, we can see that almost all the colours are heavily concentrated around the blue and cyan areas of the colour wheel. In order to get a balanced picture we need to move these up towards the yellow and red area.

What is the Y-Waveform?

The Y-Waveform does things a little differently. It shows us our Luma in marvellous graphical representation, just as the Vectorscope deals with Chroma - however - because there are fewer "dimensions" of colour within Luma, it's better to show it in a different way. Here, the light strength is shown, but based on a column for column basis, so that the left-most column of pixels in the Y-waveform represents the left-most column of pixels in the image. In this example you can see how the white section on the back of the police car causes a spike in the levels on the Y-waveform to the far right.

A key thing to remember about these two is that aside from giving you an easy-to-interpret mathematical interpretation of your image, they also provide useful warnings. TV, for example, is actually very bad a reproducing colours (especially orange - useful to know) and there are only so many shades, or degrees of difference that can be displayed - a normal PC's colour space is RGB 255, although TV won't do more than RGB 235 so you need to be careful. If you've got colours outside the inner ring of boxes on your vectorscope, or you've got blacks or whites in your "White Zone" on the Y-Waveform then you'd better think about fixing your image.

What is the White Zone?

The "White Zone" is a warning zone set at the top and bottom of your Y-Waveform with the "Clip High" and "Clip Low" functions within the HSL - Controls tab. The defaults are usually quite good, but if you're working within strange formats or putting out to strange formats then you might want to change them. To experience the "White Zone", boost your image's brightness (also from within the HSL - Controls tab) to the top and see what the Y-Waveform does.

Fixing that Image!

You'll see on your three colour wheels (the ones in the HSL - Hue Offset tab) that there are small text labels on them - Shd, Mid and Hit. These are for Shadows, Midtones and Highlights, and if you were to play around with them, you'd see that those are exactly the areas that they effect.
What we're interested in now (If you did try, you'll get something of the idea of exactly how hard manual colour correction is...) are the small buttons with eye-droppers on them to the bottom-right of each of the wheels. If you hover your cursor over them you'll get a text that appears saying "Remove colour cast from Shadows/Midtones/Highlights".

Click on the first one, then move up to your badly white-balanced image and click on an area that should be black (or at least very dark). So that you know, I've clicked on the area under the rear wheel arch, because it's very very dark.
Then take the next button and click somewhere that should be a middle grey tone. I use a patch of the stone in the building behind the car. Finally do the same for a white area. Once you've done these three then you should have a much better looking image, because you've said what the Luma channel should look like, and the program is then able to remove an equal amount of blue from the entire image. Here's one I prepared earlier:



As you can see, the colours in the Vectorscope are now located much more centrally and the image looks a lot better. So that you can see, here's the finished product in my record monitor with the original clip loaded into the source monitor:

Useful functions under the HSL

There are four very useful functions under the HSL - Controls tab - Hue, Saturation,  Brightness and Contrast. This is what they do:
Hue - Rotates the pixels on the vectorscope around the hub - give it a try and see what happens.
Saturation - Amplifies the Vectorscope, makes blues blues, reds redder etc.
Brightness - Moves the Y-Waveform up and down.
Contrast - Increases the difference between colours by making them stronger relative to each other.
Here's a trick - take an ordinary clip that looks a bit washed out, load it up in Colour Correction mode, go into the Controls tab, boost the Brightness, Saturation and Contrast by between 5 and 10 percent each and see what you think. It's one of those things that makes film and TV look different to what your Dad shoots on his Handycam - they do this to everything.

Keyframes

There are a lot of ways of doing things – and doing things in Xpress Pro is no exception - there are often a lot of ways of getting the same effect. The trick with the effects is to look closely at everything that they can do. The example I'm going to use is a simple Picture in Picture effect (PiP from now on), and show how there's a bit more to it than you might think.
As everyone reading this is at different levels of ability, I'm going to explain the whole keyframe thing from the ground up.

A Keyframe is something which defines something's state at a certain frame. Every effect has a minimum of 2 keyframes, one at the beginning and one at the end. The really important thing is that you can add more of them wherever you like and set attributes at any point. What this actually means practically is that you can control aspects of the effect over time without having to define changes at every point because the program will look at the keyframes and work out the difference between them and then make the changes over every other frame for you.

So: We've got a PiP effect, and we've gone into the effect editor and clicked on that little pink button in the corner called "promote to advanced keyframes". This is scary, because the contents of the effect window as we know it disappears and is replaced with titles and a pile of small arrows. Clicking on the arrows gives you graphs like this:



These graphs are very useful. In this case, the horizontal axis of the graph represents the amount of time the effect lasts (in this case it's the same as the length of the clip we've applied the PiP effect to, of course transitions work slightly differently, as they aren't restricted to the length of the clip they're applied to) and the vertical axis represents the intensity of the particular aspect of the effect. In the example I've opened the "Foreground" graph, because (to bring things back to what I was talking about at the beginning) we want our PiP to fade in and out, rather than just appearing suddenly. One way to do this would have been to add disolves to both ends of the PiP clip, but it's untidy and more processor-hungry to do it that way as you'd be previewing 3 effects at once (well, you'd only be previewing 2 at once, but you'd be buffering all 3, slowing your system down), but because we have the "foreground" option, we can go in and do this to the graph:



As you can see, I've added two more keyframes (click where you want them, then right-click and select "Add Keyframe") to the automatic ones which are inserted at the beginning and end. Then I've dragged the first and last ones down to the 0 level. The effect will now make the image fade in nicely. To make it even nicer, you can right click on a key frame and change "Linear" to "Spline" which will smooth off the changes for you, or "Bezier" which will let you define the curvature. "Shelf" is worth a look, although I'm yet to encounter a situation where it would be useful (Cue lots of people saying they use it all the time).
So now we have opened the door to keyframes - have a play around and put together something like this and see what happens:



It will look strange, but it does give you an idea of the power that's hidden away behind that "advanced keyframes" button.
One important fact is that if you want to move a keyframe in time rather than just intensity at the point it was originally inserted, you need to hold down ALT - that's all, just don't think you need to make a new keyframe 4 frames to the left of the one you just put in because you didn't like it. Easier to just move them around a bit.
The real point of this tip was to make sure you have a close look at all of the different options within an effect - you can usually do a lot of extra things, even something as simple as an ordinary dissolve will let you control how strong it is throughout it's playback, you can even make it go in, out, in and out again should the fancy take you.

Basically - check that you can't do everything in one effect before you start stacking them on top of each other (called "nesting", which I will cover next time) because performance will be harmed and your timeline will get crowded.

Trim to fill

In Xpress Pro, you've got a section in your effects window called "Timewarp", and there are a load of different effects in here. Now, this is one of those times when there's a bit of a difference between Xpress Pro and Media Composer. I'll be covering the Xpress Pro way of doing things here (you can do much the same things in MC, although there's a slightly different way of doing things as there's a different workflow in there because of some increased functionality) because that way we don't exclude anybody.

There are a few different things in there (most of them with reasonably self-explanatory names), but the one I'm going to concentrate on is called "Trim to Fill".

Trim to Fill is useful because it lets you expand or contract a clip however much you like as precisely as you like. Basically, Trim to Fill makes the clip's behaviour in Trim Mode slightly different. Instead of cutting the clip, it changes the speed at which it plays back. So, if you drop a Trim to Fill effect on a 2 second clip, then enter Trim Mode and drag the clip out another 2 seconds, then you won't get 4 seconds of the source clip payed at 100% speed as you would in normal Trim mode, you'll get the same 2 seconds played at 50% of the normal speed.
Some useful applications of this include fitting things to music (dragging a clip out to fit a beat, for example), making short clips fit long narrative descriptions, adding dramatic speed changes - there are loads of reasons why you'd use this, and they don't look to bad either - there's a lot that goes on behind the scenes when you apply a Trim to Fill effect.

If you have a 2 second clip, and you want to make it last for 4 seconds, the obvious way of doing it is making each frame last twice as long. This is fine, but you have to remember that the human eye "refreshes" at just under 24 times per second (it's a bit more complicated than that, but basically - 24 frames looks fluid but for a lot of people 23 doesn't). If you play each frame twice, it means that there will (in a 24 fps project, which, for my convenience I have decided that we are working in right now) only be a visible 12 frames per second. This will look jerky, so what the Trim to Fill effect does is that it looks at a frame, and the following frame, and "draws" a frame in between that is a sort of "average" of the two. Then it does it again. If you've slowed something down a lot then it will start looking for the frame between frames between frames and drawing that one. I have added another little picture which will perhaps explain the process a bit more clearly:



In the picture, we can see the original clip, with a star moving across the picture from left to right. The second clip has extended the first by 100% by duplicating each frame. The third clip however, has extended the playback time by interpolating the frames and drawing new frames in between the originals. As you can imagine - the final clip is the much better option for doing this kind of effect as it prevents jerkiness.

One thing to be wary of though - if you start dropping the speed down to, for example, 10% of the original, the program is having to write 9 frames (or, between each full frame of information, there will be 5 generations of derivitive frames) which means you will, ineveitiably, lose some accuracy - although if you're slowing down that much you might want to consider re-shooting at a higher frame rate if it's at all possible.
Trim to Fill functions in exactly the same way in Media Composer, but you have a special window in many of the time effects which allows you to keyframe accelleration and set numerical speed percentage or time values and key-frame them, which is very nice but in my opinion Trim to Fill is a more than powerful enough tool in itself for normal applications.