Standardizing downstream LUTs.

[Paul Bruening]

I'm hoping Tyler, Robert or Dominic can help me on this question.

 

Since I am getting closer to getting this telecine/scan rig going, I have to consider my best approach to standardizing the system both upstream and downstream. Would I be well served by using Kodak's TAF/TEC strips as THE standard to my whole system? My thinking is: If I get each step set correctly starting at the capture point with TAF/TEC, then I can dependably set and inter-relate all devices and software downstream right out to film recording. Am I barking up the right tree?

 

Thanks for any help, fellas,

Paul

[Paul Bruening]

David? Anybody? Take a stab at it.

[Guest tylerhawes]

I suppose you could do that as a starting point, but in reality you can get better results by having a tweaked setup for each film that comes in, rather than a one-size fits all approach.

 

Also, I'm a Colorist, not an Engineer, so I tend to place my emphasis on subjective quality. For my purposes as a creative person, it doesn't matter to me if the scanner is calibrated to the recorder, so long as it gives me plenty of range and detail to manipulate in the DI suite, and that any color bias of the device is easily neutralized without a very complex curve.

[Paul Bruening]

Hey, Tyler,

 

Thanks for the input. My thinking is that standardizing will still leave room for artistic latitude. I don't want to nail down the process to only one look, necessarily. I'm, more, just worried that the system can get whacky, one machine to another, and get crazy results at final output. If more than one person is tweaking the gear to compensate for something subjective at their particular level or station, the system could go outrageously off kilter. I guess I just have normal engineering paranoia and want something I can lean on like the TAF/TEC system.

[K Borowski]

Forgive me for following you across threads Paul, but my thinking is that you are still overreliant on software to fix hardware problems. You know the old addage: "Garbage in, garbage out." It's simplistic and cliche, but there's a kernel of truth in there. Ultimately, you can do a great deal of work and hide all sorts of physical imperfections from scanning, but why bother? Ultimately you're wasting good time on a computer with something you can fix and prevent with a little bit of elbow grease before you rush forward with things.

 

Ultimately, I think a DSLR is a bad route to take as your method of digitization. You're taking another trip through a lens, with all of the corresponding generation loss, needlessly, and you're incurring a couple of hundred dollars in replacement costs for a shutter you don't even need to be using.

 

You probably want to look at modifying a high-end used Kodak still film scanner to your needs. You don't use another lens (at least per se), and you're using something that is actually designed for scanning. I don't think you need to worry so much about LUTs as you are. Film is an imperfect animal, and as such you're going to get color variance no matter how even your processing is. There are just too many variables to consider in processing and in film manufacture; it is not an exact science.

 

The most important thing you need to worry about is monitor calibration when you are dealing with digital files. Do you have a good CRT. I would NOT recommend using an LCD monitor of any sort for this sort of work, as even the best ones have color and saturation shifts depending on the angle at which you view them. CRTs are getting harder and harder to come by, so down the road we're all in for some fun times.

 

Make sure you don't limit yourself by using custom color spaces either. I don't know what the industry standard for ECN-2 stuff is, but make sure you find out what the color space is and use it. You basically need to find out what the lab or film-recorder service you're using uses for their color space and make sure to match their profiles with what you're using.

 

Keep in mind too that not everyone is good at color timing; it's a skill that takes years to master, not something you can pick up overnight. You'd probably be best off using a trained timer anyway. You're frankly not going to be able to match what a skilled timer can do unless you devote 4-5 years of your life to it every day. I've been doing a lot of color printing, color balancing for THREE YEARS now, and I still am nowhere near as good as someone at a lab. It's THAT HARD. And no, I'm not colorblind or color deficient.

 

Another tidbit of info, you can only spend about four hours a day doing color timing before you run in to what is known as "color fatigue". Your eyes start to play tricks on you. Color timing, be it on a computer or with a video analyzer, or looking at test prints/cinex strips is a serious piece of work that takes talent and skill to really master. Don't kid yourself into thinking that software is going to make it a breeze to pick up. The operator skill is paramount to good color balance, not the computer program, LUT, or monitor.

[Paul Bruening]

Okay, Karl,

 

I'm an idiot. You can back off, now.

[K Borowski]

Okay, Karl,

 

I'm an idiot. You can back off, now.

 

I'm saying this because I learned the hard way that you can't do it all yourself. If you want to follow down the same path and ignore my advice, that is up to you. Also it is really frustrating when one takes twenty minutes to write and revise a post and it goes in one ear and comes out the other. Instead of taking constructive criticism, try to look at it objectively. I am not out to belittle your endeavours.