DJ Joofa

To have good MTF you need sharp cut off filters. A standard technique in signal processing to simulate a filter with a sharp transition from pass band to stop band is to do filtering in a cascade of lower complexity filters. So Red can have a loose OLPF, and then oversample , (not upsample), i.e, 4K vs. 1080, etc., and then use a simpler low pass filter to get the desired steepness of the cut off response, to get to 1080. If done directly on 1080 data, then the corresponding low pass filter becomes more complex, and perhaps not always implementable. Therefore, as seen above, Red is naturally disposed to do a sharper 1080 than a direct 1080 camera, at least more easily than a 1080 camera.

The way we work with sensors, dynamic range is a point property, where as resolution is an area property, so inherently they are two different concepts, and hence, no issue of "having massive amount of resolution to mimick film" as far as DR is concerned. Just like color at a pixel is a point property, where as the texture is an area property around that pixel. Of course, one can always make relationships between dynamic range and resolution, since SNR, which is a rough correlate of DR (and which is implied by many people when they mean DR) can be traded with resolution, etc. The term "dynamic range of the sensor," should be taken to mean the dynamic range of a pixel, however, since in theory, all pixels have the same dynamic range, they may be used interchangeably. If sensor electronics warrant very little noise, then, resolution is mostly irrelevant to mimic the DR of film, as sensor electronics considerations at each pixels are more meaningful, more so in the case of CMOS, which has more processing elements on a single pixel than typical CCDs.

True. However, I am just reporting what I have read before as the people in Hollywood wanted to put a money tag on what should be considered an "independent" movie regardless of if it was made outside the studio system. I have tried to find that source again. But I can't relocate it.

I have read in the past that Hollywood has put an upper limit of just under $30 million for independent movies.

Many university libraries have them. Go to a good university library near you.

Don't know much about windowed 2K on Red. However, a comment on 4K->2K vis direct 2K: In signal processing terms, one can have a steeper low pass filter (sharper transition from pass to stop band) when using a light low pass filter, then oversampling (not upsampling), and then downsampling, as opposed to the application of a low pass filter directly. Therefore, Red should have an advantage when going from 4K->2K to cameras that offer direct 2K or close to 2K resolution (such as HD cameras).

You need to divide 2260 line pairs with the width of the sensor also to get lp/mm. Can't seem to remember Red's width, however, using 24 mm, it comes out to be 2260/24 = 94 lp/mm.

True, though stability has to be factored in. Plan 9 is a much more powerful operating system than Linux, and is from some of the same guys who developed Unix and gave us the wonderful C language. However, the market is held by Linux, which is a hodgepodge mess in relative comparison. Llinux just happened at the right instance and being free just took over (though I think Plan 9 is also free). Just goes to show that perfection is not the required thing -- just rise a little / marginally over the current bar and have a good price point, and there you go. However, some stability is important, and many times the customers prefer the trade off between stability and more features.

In addition to color, many industrial-strength keyers, such as Primate and KeyLight, use the luma information to accumulate the silhouette of the foreground vs. background. Several hacks are used on the luma information to get to that stage. Blue can have an advantage in outdoor shooting so that the key does not have to be perfect because people accept blue overtone on skin color and perhaps clothes, in outdoor setting. One may force white balance to use co-efficients that don't boost blue gain as much so as to reduce noise if that helps in background / foreground separation.

There are several variables in digital skin rendition and it is not always easy to control the process using standard color space transforms that are employed frequently in digital processing. However, according to some studies Fairchild color appearance model, which considers incomplete chromatic adaptation of cones to the white point, gives pleasant results.

I think you are mixing up linear and non-linear notions here. The way data is recorded on film, having a high DR is not always helpful as far SNR is concerned, because though SNR would remain more or less constant over a broad range of input variance, it would actually *decrease* after a certain value.

Don't know how do you standardize a number such as 4K. Additionally, standards such as H.264, MPEG-4, MPEG2, etc., define the decoder and not the encoder. Encoder is not specified as such by these compression standards.

I think the resolution on the print itself is higher than this. It is after projection the effect is around 800 vertical lines according to some studies.

Wrote the following on a different thread, but is applicable here also: Dynamic range is the number of variations you can have in your signal from pitch black to max white. "Tonal range" is an important parameter and that is typically the range a cinematographer has in mind that the image should have. They may not want to use the full range of variation that is offered by a camera for a few reasons and one is mentioned below. Since, by definition, the full range (dynamic range) of a camera is typically more than the range one wants to have in an image, the left over space or room is called latitude -- it is only useful in the post production after acquiring an image to shift the tonal range up or down for a more pleasing image. Please bear in mind that the normally the print film (the one that will somehow get to projection in a theater) has typically a smaller number of variations between black and white than the original camera negative. Another way to say the same thing is that the dynamic range of print film is less than camera negative. However, that is fine, because once a cinematographer has decided that which portion of the larger dynamic range offered by the camera negative to be printed to film (i.e., the tonal range), they should strive to keep that tonal range not exceed the capacity of the print film. There are several reasons that the print film variation (or dynamic range) is less than the camera negative, and one important one is that in a theater / cinema there is always some flare present that tries to reduce the contrast of projected film, hence, in order to compensate for that fact the print film has higher contrast than the camera negative, and that also makes its available range of variation less than camera negative.

Actually what you are describing would be the "tonal range" (for the lack of a better word) of an image, and latitude will be the compliment of this number. I.e., Dynamic range = tonal range + latitude. Where latitude (measured in stops) will be number of stops between pitch black and darkest black in the acquired image + number of stops from the "whitest" white acquired in the image to the max white the camera can acquire. Dynamic range is the number of variations you can have in your signal from pitch black to max white. "Tonal range" is an important parameter and that is typically the range a cinematographer has in mind that the image should have. They may not want to use the full range of variation that is offered by a camera for a few reasons and one is mentioned below. Since, by definition, the full range (dynamic range) of a camera is typically more than the range one wants to have in an image, the left over space or room is called latitude -- it is only useful in the post production after acquiring an image to shift the tonal range up or down for a more pleasing image. Please bear in mind that the normally the print film (the one that will somehow get to projection in a theater) has typically a smaller number of variations between black and white than the original camera negative. Another way to say the same thing is that the dynamic range of print film is less than camera negative. However, that is fine, because once a cinematographer has decided that which portion of the larger dynamic range offered by the camera negative to be printed to film (i.e., the tonal range), they should strive to keep that tonal range not exceed the capacity of the print film. There are several reasons that the print film variation (or dynamic range) is less than the camera negative, and one important one is that in a theater / cinema there is always some flare present that tries to reduce the contrast of projected film, hence, in order to compensate for that fact the print film has higher contrast than the camera negative, and that also makes its available range of variation less than camera negative.

No, at 4k it does not become irrelevant, however, please refer to the following SMPTE journal paper to find out that there is little benefit in going over 3K scans of typical film negatives: B. Hunt, et. al. "High resolution electronic imaging system for motion picture film", SMPTE Journal, March 1991. Stephen Williams (and some others on this forum) objected to me when I mentioned this before saying that study is over a decade old and film stocks have had an opportunity to improve. Fair enough. However, I went back to Glenn Kennel's book from 2007 (BTW, Kennel was also a co-author on the above mentioned paper), and he says: "Original negative films improved substantially in the last 15 years, now exhibiting higher speed and lower grain. However, the MTF has not improved much, and typical camera lenses are the same, and today's CCD film scanner and laser film recorders have similar MTF characteristics as the original Cineon scanners and recorders." Human visual acuity is not more than 60 cycles per degrees at the fovea (not to be confused with 60 cycles/mm or 60 lp/mm). In a theater, only one distance from the screen will have the best visual response and even that will not afford a resolution near 4K according to some studies.

Yes I agree with you regarding art. I have huge respect for Terrence Malick. When I saw "The Thin Red Line" I thought that this is the best movie I ever saw -- I saw it about 20 times in a theater and countlessly on VHS/DVD. Every time I would get an interesting new meaning. I don't know if I still qualify "The Thin Red Line" as the best movie I have ever seen, but it is certainly in the top 5. "Crouching Tiger Hidden Dragon" is also there in the top 5. I thought "Crouching Tiger Hidden Dragon" was also amazingly philosophical. There is so much Eastern philosophy in it that it is an awesome piece of art. While Malick's philosophy is secular in nature in "The Thin Red Line", the thought process in "Crouching Tiger Hidden Dragon" is very spiritual.

Don't know about Red's sensor. However, many sensors from some regular manufacturers do have blanking time embedded with active video in the signal out.

Some video compression schemes have been made around the notion of foveation. Please do an online search on foveation, compression, etc. Ref.: Sanghoon Lee; Pattichis, M.S.; Bovik, A.C., "Foveated video compression with optimal rate control" IEEE Transactions on Image Processing, Volume 10, Issue 7, Jul 2001 Page(s):977 - 992.

Keith, I have given you a reference in the relevant area. I can only do so much. If you want to keep your eyes closed, it is up to you. I can't argue with somebody who is not willing to listen.

I re-read my above statement and thought it was a little ambiguous because of the wording. I am not saying that Lubezki and Malick collaboration was disappointing in the "The New World"; the film was very beautiful as far as cinematography was concerned; what I am saying is that the film itself was big disappointment for me. As I mentioned before, Malick should have gone farther than "The Thin Redline" -- which I really liked. However, I think he just rehashed some of his existing philosophy and hardly provided any new insights into his own development as a philosopher and thinker. I thought that "The New World" was just beautiful images stitched together. Little new over "Thin Red Line" to remember and write home about as far as offering visibility into the director's world view for which he is well-known. I really think the movie was rushed through and hurriedly done. Additionally, IMHO Collin Farrell was a poor choice.

I think Lubezki and Malick collaborated on the "The New World" also, which IMHO was a huge disappointment, especially after Terrance Malick's "The Thin Red Line" was marvelously and amazingly done. I think Malick got carried away by the sentiments of "The Thin Red Line" and did not offer public any new insight into his philosophy. I hope that this time he offers something more.

Trust me, people are actively working on it. If you have not seen them mass produced then that does not mean that there are insurmountable obstacles to it. Several experimental non-linear sensors have been fabricated. Please refer to the following publication as an example where the authors claim that they implemented a 120 db DR !!! sensor. (Using the standard notion of 6db per bit, that would turn out to be 20 bits, or 20 stop, well into HDR range!) Hara, K. Kubo, H. Kimura, M. Murao, F. Komori, S., "A linear-logarithmic CMOS sensor with offset calibration using an injected charge signal", Solid-State Circuits Conference, 2005. Digest of Technical Papers. ISSCC. 2005 IEEE. Enough work is being done actively in this area. This technology is just maturing and sooner or later you will see the results.

This issue has been explored. It is an active research area. Please search online for nonlinear CMOS sensors. It is not terribly difficult to add an analog nonlinearity to sensors.

Yes surely, it is. Trying to understand it :blink: Thanks for sharing.