Look: In-Camera or Post-Production?

Brandon Rice

Brandon Rice

Distributed Producer
This is a hotly debated topic I thought might be an interesting discussion. I also placed it in the DVX section as the 4.1.1 colorspace comes into play here.

Many believe in getting as close a finished image in the camera. They shoot with a low master ped, and may use gels to get different tones in the lighting of the scene.

Others believe in shooting a relatively flatter image, having the master ped a bit high, and colorizing and grading the image in post to the look they want.

I personally subscribe to the latter, but I am curious to what other people think, and why they do it... discuss....
 
well, this is all in my theory, which could also be a proven fact.

But the way i see it as though the recorded image, as it is a standard definition, 4:1:1 sample, toying with it even more in post will degrade the quality of the footage. The benefit me thinks of getting the look in camera is that you are recording the already adjusted image onto the tape and compression thereby preserving all of its initial quality. ?
 
Let me ask. Does toying with a still image in Photoshop degrade the quality or merely change the look of the image?

Also, isn't the footage going to be compressed for a DVD or the Web anyway?

All of these Q's aren't meant to be taken any way other than spurring on some conversation and talk about this.
 
im just saying, like if you were to shoot flatly with the dvx, master pedestal of +3 or so, it might be to the advantage of a user who wants the highest possible quality out of the 4:1:1 to crush those blacks in camera. I don't know if it's true but it seems to me that if the camera were to crush the blacks in the look before having to compress it to tape, then it would retain a higher quality of picture as opposed to taking the compressed image and darkening it. I just don't think you would get the same quality blacks.

but on the other hand you bring up a good point. When it is compressed to dvd or anyting viewable online, it might not be anything worth noticing, even if one were looking for it. But I'm not one to talk cause i've shot both types of ways, yet never compared the two for image sake.
 
I'm an in-camera guy. I don't get too funky with my stuff anyway, so all I do is a little touch-up work (contrast boost, adjusting any colors that I want to pop more, etc.)
 
Hm. Interesting thoughts. My personal thought on the matter is if you retain information in the shadows (shoot flatter with higher master ped (I have mine at -3 or 0 all the time)) then the image will look better even bringing the shadows down in post. I have never seen a noticable difference in a film in-post or in-camera.
 
Brandon Rice said:
My personal thought on the matter is if you retain information in the shadows (shoot flatter with higher master ped (I have mine at -3 or 0 all the time)) then the image will look better even bringing the shadows down in post.
Click to expand...

I usually sit in the -6 to -9 range. If I need more detail in the shadows, I'll light them more. There's nothing wrong with doing it either way, if you ask me. It's just a matter of preference. I just like to see what it's going to look like when I'm still on the set to make changes.
 
I assume that we are discussing video and not film so I will bias my response accordingly, but I think that you know my approach already.

In general with a heavily undersampled color space system, I prefer to get close in camera and keep the touch ups to a minimum in post (I guess being a DP and not a colorist this is my area of comfort).

My goal is to get as close as possible to the "look" that I want on set in an attempt to preserve the most usable information. If the final project is destine for very low resolution applications (highly compressed DVDs or web viewing) the argument for preservation of quality is somewhat lessened.

In some cases the director and I may not have gotten to the same point (in terms of look) or the director might not have any idea what they want. In these situations, I split the difference, but still light and shoot for what I envision as the final "look" of the project. Certainly getting an "in camera look" requires the tools to do so when shooting, but I like the results that I can achieve when I have the time to do it right.
 
But my continued question is why is there an assumption that getting a look "in-camera" somehow preserves the quality of the image?

I again go to my retorical question... if you colorize a photo in photoshop, does it degrade the image, or does it just change the look of it?
 
I'm with KHMuse once again. I try to get in-camera as much as possible. But that's only one perfectly valid workflow; shooting "clean" and creating your look in post is certainly completely valid too.
 
Any time you make a change to an image, you have to re-compress it for final output. Adding compression to compression will degrade the image.

Some codecs preserve quality well; others don't.
 
Hm. I see... but if you pull a FINE JPEG (compressed image) into PS, and add some color filters and then save... I am pretty sure it simply saves the same image, it doesn't compress it again....
 
Brandon Rice said:
But my continued question is why is there an assumption that getting a look "in-camera" somehow preserves the quality of the image?

I again go to my retorical question... if you colorize a photo in photoshop, does it degrade the image, or does it just change the look of it?
Click to expand...
Different things. PhotoShop is letting you work in an uncompressed space, and able to save it out as a PNG. And it depends on the source of the photoshop picture, you might be working in a true uncompressed 4:4:4 and not a color-subsampled, resolution-stretched, compression-artifacted source picture.

If you have to re-render your footage, you're sending it through another layer of compression. If you don't have to re-render (as in, if you kept it pristine and got the look you wanted in-camera) you could avoid that. DV is a lossy format, with low color sampling. HDV is an even more lossy format. Re-rendering or sub-rendering will introduce loss into the image. DVCPRO50, not so much. DVCPRO-HD, a bit more than DV, but less overall image degradation because the 4:2:2 color sampling preserves more color info.

So if you're never re-rendering except when you go to your final render to DVD, then no, you'll lose nothing. If you're rendering in-between, yes you're compromising your quality every time you do so. If you're rendering out a movie file to send to After Effects, you've compromised your quality to some degree by doing so (unless you use an uncompressed or lossless codec!)
 
So... by adding color grading in an NLE such as Avid, that changes the media? I don't see how that can be, as I create a Quicktime Reference file that pulls my footage direcly from the raw captured clips, then from there I create a DVD file.

Your Photoshop explanation makes sense. And I could see how Color Grading in After Effects and exporting out again could change the media.
 
OK, now I see where your question is coming from. Let me explain the difference as I see it.

Lets look at the worst case situation, namely 4:1:1 video. This differs a lot from say a digital still shot in RAW mode. Consider that for every 4 luminance samples (each of which can have a different value assigned to it) you only have 1 color value for the two color difference channels (Cb & Cr). If you adjust the color of the image, your luminance range remains the same (256 to 1 in an 8 bit system) but the step size (resolution) takes a hit due to the averaging caused by the undersampled color space.

In theory, a small shift will have less impact than a larger one, but there is still a hit to the overall resolution due to an increase in the step size caused by the averaging across the 4 samples.

Now if you are shooting 4:4:4 color space, then this aspect of the quality hit isn't much of an issue.

Now lets consider a totally monochromatic shot, one without any color information. Even here, getting the dynamic range correct in camera is going to result in greater resolution.

If you simply "light and shoot flat" you may not place your shadow detail and highlights at the extreme of the dynamic range of the system you are working with. Your minimum might be 1 stop or so above the bottom and your maximum might be one stop or so from the top. This means that you are reducing your system from an 8 bit one (with 256 levels) to a 6 bit one (now with only 64 levels). Sure you can expand this in post back to the full range, where your shadow detail bottoms out near 0 and your highlights tops out near 255, but you still have only 64 steps between this two extremes. Instead of being able to resolve steps 1, 2, 3, 4, 5, 6, 7, 8... you can only achieve 1, 4, 8, 16...

Hope that this gives you the information that you are seeking.
 
You must log in or register to reply here.
Back
Top