This is topic Steadiness of dye transfer prints in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.


To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=002369

Posted by Christopher Seo (Member # 7) on 06-11-2001, 03:49 AM:
 
Some idle thoughts and reasoning:

(1) The dye transfer image is created through three separate printing passes of the various dye elements, CMY.

(2) For a print to be of acceptable quality, these elements must be aligned exactly to each other (within 0.0002 inch, according to http://www.technicolor.com/aboutus/about-innovations.html). Otherwise, fringing and fuzziness will result.

(3) The elements will be aligned to each other if they are properly registered on the film. This is accomplished (I think) with a pin-belt system; therefore the image is registered relative to the perforations of the film. (Obviously...)

(4) Therefore, in an acceptable print, since all three color elements are precisely aligned to each other, they must also be precisely registered to the perforations. And since each color element is steady, the overall image must also be very steady.

In other words, theoretically, a good dye transfer print should always be steady. As steady as the source material, anyway. Has anyone noticed this? Any corrections to my reasoning?


 


Posted by Josh Jones (Member # 447) on 06-11-2001, 09:22 AM:
 
In theory, unless yor movement is a POS

Josh


 


Posted by John Schulien (Member # 215) on 06-12-2001, 05:37 PM:
 
I've thought about this also. It's a good example of the difference between accuracy and precision. Accuracy has to do with measurement against a standard, while precision has to do with reproducability. A measurement or process can be accurate, but not precise, or vice versa.

The new IB printing machines use a short race-track shaped pinbelt that aligns the matrix with the receiver film, and presses them together. The machine then peels off the the two-film "sandwich" and passes it through a long "transfer box" which keeps the matrix in contact with the receiver film long enough for the dye to transfer.

I'm speculating that in order to get the best possible registration:

1) All three colors must be printed using the same machine, and

2) The pin-belt must be aligned to the same sprocket before each color is transferred.

Why? Imagine a tiny defect in the pin-belt that causes a slight misalignment of the matrix with the receiver. If the pin-belt were randomly aligned for each color, or the different colors were printed on different machines, then each time that defect came around on the pin-belt, it would create an error that would only affect one color, and produce fringing.

However, if the pin-belt were aligned to the same pin for each color, that same defect would be duplicated for each color, and there would be no fringing.

So it might be entirely possible to have an IB print with perfect color registration, but overall picture weave. For what it's worth, I've never seen anything like that, but still, I wonder if IB printing procedures call for the alignment of the pin-belt to the same sprocket, or even call for the same machine for all three transfers.

 


Posted by Christopher Seo (Member # 7) on 06-12-2001, 07:32 PM:
 
Ah, good points. I'd simply assumed that the dye transfer print was a product of randomly made passes through the printer, in which case I think my original arguments would hold-- the process would have to be both accurate (perfect color element alignment) and precise (perfect overall image registration).

But your arguments certainly make sense. It sounds like the number of registration pins involved in the belt number in the hundreds, and it would be very difficult or expensive to maintain them all to ensure the same precise tolerances. Making sure each sprocket hole ends up with the same pin in the same printer on each pass sounds like a good, practical idea. Although, if the lab can do that, you'd think they could also get lab splices in frame...
 






Powered by Infopop Corporation
UBB.classicTM 6.3.1.2