Question

1. How does the depth of field of RED compare to 35mm?

2. If there is any difference, can it be explained by the divergence in thickness of the 3 layers of the 35mm colour negative emulsion versus the flat CMOS chip of the RED?

The actual question should be: Is there a difference in the circle of confusion between 35mm negative and RED chip?
And, to clear any misunderstanding right away: both film emulsion and CMOS (and CCD) recording chips are analogue;
the latter category having the conversion to digital taking place behind the chip, in the A/D converter. By then the recorded (latent) image has already been shaped, with corresponding sharp and out-of-focus areas.
To complete the story, one should consider the eventual screening conditions as well. As we all know: large projection does emphasize blurriness.
But neither the viewing conditions nor the differences in quality of sharpness, i.e. definition of the media, are of any concern here.
Within a limited time and budget I wanted to be able to answer the questions as stated above.

Method

In order to being able to reproduce this test easily and precisely, I performed the so called “Belgian” lens testing method.
Actually originating in France this test can be recognized by the millimeter paper that is attached over the calibration mark.

Used: a 32UP at 75 cm of test-chart.
The camera is positioned straight in front of the test-wall.
Simple check with mirror to check perpendicularity: the cross-hair should be exactly over the center of the lens.

The top of the frame contains a strip with numbers from minus up to plus 10, that is moved in correspondence with every increment of focus.

The film camera just needs to run up to speed; on the RED a single frame suffices.
The developed negative is analyzed under a stereo-loupe (low-power microscope);
the RED r3d files are transferred trough RedAlert into 16-bit tiffs, viewed in Photoshop.
Both on film as in the files you want to look for squares where the lines can be clearly separated.
The number in the square is written down.

Screen grabs taken from RED files

The smaller the number, the sharper the image.

Data

Primarily 50D and RED where tested. After transferring the data into a spreadsheet, the following graph resulted:

This does look interesting but it seems to be standing up side down, and after some shifting, so that each curve is corrected for optimal sharpness to zero, plus adding the data from black-and-white negative:

The difference between 35mm color negative and the RED is clearly visible, especially towards infinity.
The difference with black and white is even more apparent, the drop-off on both the short side as the far side is very steep.

Conclusion

The outcome of this test has been too limited to be able to quantify the circle of confusion (CoC) for the 4K RED.
It seems that the dynamics in sharpness is clearly dictated by the thickness of the material.

This can be seen in the diagram above., with left objet O and on the right side image I; the Circle of Confusion C is determined by the thickness t and this again dictates distance T, which is the sum total of the depth of field.
The actual Circle of Confusion used depth of field calculations should be corrected to this.
In diagram and calculations I would recommend to use a CoC of slightly more than the average CoC of 35 and 16mm (approx. 0.018 – 0.020 mm).
And, when focussing the RED it is advised, in the case of doubt, to choose for the short side, due to considerable drop off at the far side.

In any event, I am going to watch black and white classics with different eyes again!


© Flip Bleekrode, Aug 10, 2009

Used material and resources:
Arri 435 ES
RED met buid 17
32 UP van ECR,
32 UP van CLF,
Kodak 50D,
Fuji B/W 71112 110-05 (80D/64T) supplied by Peter Roelofs, MPF, Cineco Lab, Photoshop, Sidney F. Ray “Applied Photographic Optics” 1988 Harold M. Merklinger “Ins & Outs of Focus” (http://www.trenholm.org/hmmerk/#TIAOOF) Paul van Walree “Depth of Field” (http://toothwalker.org/optics.html, also read the arctivle on bokeh, info about foto-chemie at www.cheresourses/photochem.shtml