Sony Levels Filter - Glenn Chan's Video

Richard Jones wrote on 8/27/2008, 4:00 AM
Can anyone explain the difference between Input and Output in the Levels Filter and advise when and why to use one rather than the other?

In Chapter 11 (Repairing Exposure Issues) of the Vasst instruction video on Colour (Disc 4), Glenn Chann deals with an over exposed image by using Output End to correct the highlights and Input Start to correct the dark tones: for an under exposed image he uses Input End to restore the highlights and Input Start to deal with the dark tones. But it's not clear why he chooses to use one rather than the other. What is more, his interesting article about levels on the Vasst site is hepful but does not seem to cover this point.


The Help file indicates that the Input slider should be used to adjust the luminosity of the colours in the event as it is captured coming into Vegas but this just can't be right as the event has already been captured and been placed on the timeline! It also states that the Output sliders should be used to adjust the luminosity of the colours in the event as it is fed into the preview monitor but I don't think this can be right either as any adjustments made with the Input slider also appear in the preview.

Playing with Levels I have found that the two Input sliders allow you to take the higlights and shadows to 100 and 0 (or wherever you want them on the Waveform) with each one working independently of the other. Each of the Output sliders moves the dark and highlight tones in the same way although this time they also seem to compress rather than stretch their overall distribution as the other end of the scale moves up and down at the same time.

Puzzling! Which to use, when to use it and why? Can we stay with Input alone and ignore Output? If not, how do we decide whether and when to use them both?

Regards to all,

Richard

Comments

TheHappyFriar wrote on 8/27/2008, 5:41 AM
"long" discussion over @ jetdv's forums: here.

basically. it boiled down to use it when you think it needs it, like CC.
Richard Jones wrote on 8/27/2008, 6:05 AM
Thanks Happy Friar . That's my point really. As I said in the jetdv discussion, I am really looking for something more than intuition. There must be a reason for having the Input and Output sliders and they must have their own uses otherwise why have them?

Richard
farss wrote on 8/27/2008, 6:23 AM
The Levels FX provides a linear transfer function, think of it as a straight line if you're into graphs. The input start is the x1 value. Input end is the x2 value. Output start is the y1 value and output end is the y2 value.

So if your original video has values from 0.1 to 0.6 you'd likely set that as your input start and end. Assuming you wanted to keep the low values (black) where they were then output end would be also 0.1. If you wanted to shift your highlights up you might make output end 0.9.

Now in that scenario of you changed input end to 0.5 any values over 0.5 will be clipped.

Pretty much the Level FX is the same as the Colour Curves except you are only using a linear curve.

Hope this helps, sometimes getting some graph paper and a pencil and ruler really helps.

Bob.
JohnnyRoy wrote on 8/27/2008, 7:51 AM
Actually it's pretty simple to see if you always use the waveform monitor (part of video scopes) while you do the correcting. The input sliders boost the inputs and the output sliders cut output. One is a boost, the other is a cut (or rather a hard limit). The rule-of-thumb is: if your video is inside the graph 0-100 and you want to boost it to fill the graph, adjust the inputs. If you video is out of range and you need to cut it to bring it within the 0-100, adjust the outputs then if you still want more spread adjust the inputs.

As Bob said, these sliders set the clip values so if you use the output to cut the whites, no amount of boost from the inputs will bring it over the output limit. You can see this behavior very clearly when using the waveform monitor.

~jr
GlennChan wrote on 8/27/2008, 11:34 AM
The sliders do different things. An easy way to understand it would be to add a black to white gradient onto the timeline and look at the waveform monitor video scopes (as jr points out).
rmack350 wrote on 8/27/2008, 11:42 AM
Scopes, Scopes, Scopes.

The Vegas filters don't give much feedback so you really need to get in the habit of opening the scopes window when you use any sort of correction filter. Can't stress that enough!

JR's description is very good, I think.

Rob
Richard Jones wrote on 8/28/2008, 4:25 AM
Thank you guys. This is really most helpful . What Bob and Jonny said fell nicely into place when I used the Black to White Gradient as suggested by Glenn.

One final & probably silly question if you don't mind. What exactly is "Input" and "Output" in this connection & as used by Bob & Jonny in their replies?

Regards,

Richard
farss wrote on 8/28/2008, 5:04 AM
Well if you used the horizontal black to white gradient, applied the levels FX and watched on the waveform monitor then the X axis is the Input Range and the Y axis is the output range.
Try adjusting the input start and end values and watch what happens to the waveform.
Try the same adjusting the output values.

There's one gotcha. The B to W gradient uses Computer RGB values. You can change the gradient to go from 16 to 235 to be the same as normal video. Hopefully then with a little experimenting you'll see why the values Glenn gives for input start are not 0.0.

With some experimenting you'll likely see one limitation of the levels FX. It's not overly flexible compared to the color curves FX. The CC FX lets you create a non linear function. Then you can say bring up the shadows without clipping anything.

Bob.

Richard Jones wrote on 8/28/2008, 7:04 AM
Thanks Bob,

I think I follow you but I'm still not too clear about the definitions of "Input" and "Output." Input from where and output to what? And what are we inputing and what are we outputting?

Regards,

Richard
farss wrote on 8/28/2008, 7:29 AM
What the terms refer to is the input and output range of values.

All FXs have an input and an output. The input is either your video or the preceding FX in the chain. The output feeds into the track above or if none then the video bus master. There's a flow chart in the manual that shows this in a diagram better than I can describe it in less than a few thousand words.

Think of it like your hifi amplifier, it's got an input and an ouput.
So this FX is taking a range of input values and outputing them to a different range.
If the input range is 0,1 and the output range is 0,1 what goes in comes out the same. Change the input range to say 0.25, 0.75. Then an input of 0.25 becomes 0.0 at the output and an input of 0.75 at the input becomes 1.0 at the output. Input values between 0.25 and 0.75 are scaled accordingly. Input values below 0.25 are also become 0.0 at the output and values above 0.75 at the input become 1.0 at the output i.e they're clipped.

Leaving those input values as is and changing the output to 0.25, 0.75 and input values between 0.25 and 0.75 will be unchanged i.e. the gain is unity.

Hope I'm helping and not confussing you. If you go back to your high school maths days and dig through the memory banks for linear equations you'll likely find something that'll help. Heck I just realised, after 40 years, finally I've found a use for all that stuff.

Bob.

rmack350 wrote on 8/28/2008, 8:36 AM
You're doing better than me. After 30 years, I just can't remember most of my high school math.

Bob makes a good point that color curves is a lot more flexible than Levels. Levels is just generally easier to grasp so I've developed the bad habit of using it every day in photoshop.

Richard, I wouldn't get too hung up on the idea of "input from what?" and "output to where?" They're terms to describe the levels tool and are analogous to knobs on real world tools in the analog realm but you really just need to know what each can be used for.

In short, "Input levels" allows you to amplify whites, attenuate blacks, and midtones (Gamma) can be moved around a bit. I use it to make whites whiter and blacks blacker.

"Output levels" allows you to boost blacks and attenuate whites. I don't use this too much but it'd work well when prepping a still photo for video if you need your whites to be less white and blacks less black.

Grazie wrote on 8/28/2008, 9:39 AM
I'm with Richard! Yioyu stick to your guns Pal!!! - Let's march on Madison and demand HERE and NOW!!!

Well, this is my take:

INPUT aka Line/Mic Level in Audio

OUTPUT aka what I send to tape in Audio

If I got TOO much coming in I clip. BUT, if I ain't got enough then I need to ALLOW more in. Using scopes gives me a "see" on the actual spread of video "noise".

Richard, keep plugging away - I am a BIG fan of yours. Truly!!!

Grazie
rmack350 wrote on 8/28/2008, 10:18 AM
Anyway, you're inputting to the filter and then outputting from the filter.

Rob
Richard Jones wrote on 9/1/2008, 4:03 AM
I spent a good part of the weekend testing and exploring this further and am very grateful to all of you for your help. I agree that Colour Curves are far less confusing and more effective as well and nowI have a much greater understanding thanks tn particular to:-

Bob for your clear and detailed (and patient) explanations - your comparison to a graph was useful although it's so long since I left school that I'd forgotten that things like linear equations even existed!

jr for your succinct but most informative analysis

Rob for your analogue comparison and helpful description of what happens

Grazie for being Grazie - I'm a fan of yours as well (mutual admiration society!)

Best Wishes and Thank You.

Richard