Redlog

[King Cheung]

A member from here, luis ignacio barrague, just suggested me to use Redlog as gamma because it has better dynamic range. Then I tried it on my Panasonic LH2550 and Color. After testing with few shots, I am completely convinced.

Actually, what is this Redlog? I'm very curious.

Thanks in advance.

[Dan Hudgins]

You can find the gamma of red log by shooting a black, white, and 18% gray card.

Clip the image to have the white card and black card at the ends of the scale, and then measure the digital density of the gray card.

In Gamma 2.2 images the 18% gray card should read about 0.458 to 0.459. (0.0 being about 2% back and 1.0 being about 90% white)

If it were linear light the gray card would read about 0.18, so being "log" which is another type of "gamma" adjustment that makes the mid-tones lighter than they would be from linear gamma 1.0 sensor data, you end up with the gray card between 0.18 and 0.459, does anyone know the exact value?

I don't have a sample TIF that was exposed right to check that, anyway once you get that number you can use the gamma formula to find what gamma "red log" is, probably close to the DPX log of gamma 1.7?

I have been using RED LOG TIF (48bpp) for grading and it works well. Its better than linear light for grading since you don't need to rase the mid-tones as much to get to gamma 2.2 or higher for the finished result. Also RED LOG do not seem clipped too close, so you have some range at the top and bottom.

If you export linear light to 10bit DPX you loose some tones and may get banding in the graded images, but with RED LOG the mid-tones are higher and the dark tones are spaced more, so you "might" be all right with DPX, but I would use TIF for grading to reduce scaling errors since with 16bits you can pull the tones up and down more than once...

[King Cheung]

Dan, thanks a lot for your reply.

Though I must admit that I don't completely understand, I do have some ideas what it is about now.

Thanks again.

You can find the gamma of red log by shooting a black, white, and 18% gray card.

Clip the image to have the white card and black card at the ends of the scale, and then measure the digital density of the gray card.

In Gamma 2.2 images the gray card should read about 0.458 to 0.459.

If it were linear light the gray card would read about 0.18, so being "log" which is another type of "gamma" adjustment that makes the mid-tones lighter than they would be from linear gamma 1.0 sensor data, you end up with the gray card between 0.18 and 0.459, does anyone know the exact value?

I don't have a sample TIF that was exposed right to check that, anyway once you get that number you can use the gamma formula to find what gamma "red log" is, probably close to the DPX log of gamma 1.7?

I have been using RED LOG TIF (48bpp) for grading and it works well. Its better than linear light for grading since you don't need to rase the mid-tones as much to get to gamma 2.2 or higher for the finished result. Also RED LOG do not seem clipped too close, so you have some range at the top and bottom.

If you export linear light to 10bit DPX you loose some tones and mayl get banding in the graded images, but with RED LOG the mid-tones are higher and the dark tones are spaced more, so you "might" be allright with DPX, but I would use TIF for grading to reduce scaling errors since with 16bits you can pull the tones up and down more than once...