Jump to content

Tom Yanowitz

Basic Member
 View Profile  See their activity

  • Posts

    76

  • Joined

  • Last visited

 Content Type 

Everything posted by Tom Yanowitz

  1. Tom Yanowitz
    (you mean that the 5D has this with odd and even lines instead of for each pixel ?) Well, Full Well Capacity is still more important than these electrical trickeries I think, as it influences both sensitivity and dynamic range.
  2. Tom Yanowitz
    By the way if someone finds something that is or might be inaccurate, please tell me and I'll make an update. Why do you think Arri doesn't share more that kind of infos ? I think either they think we don't care about the in-depth stuff, or because it makes EI 800 looks bad. They have this popular "dynamic range" image where we see how you get +5/-9 stops at 200, +7/-7 at 800 etc etc. This is sort of true, but they forget to mention that every stop isn't created equal, in terms of how many values each has going for him. A highlight stop might have hundreds of values for it (arriraw), where a low-light stop ony has a couple. Out of curiosity, I tried to see at what ISO would a middle gray have as many values below and above him : it's ISO 64. And in terms of stops, it depends where you decided that the dynamic range ends. The Arriraw has a high-contrast part of the curve that is 8 stops long so if you want high contrast low lights in your file, you'd go for ISO 64 or 80 with a +4/-4 middle gray. By the way if Art Adams passes by on the topic : Thank you, I learned a lot with your Alexa articles back in the day, and I hope the stuff I shared will be useful to you as well.
  3. Tom Yanowitz
    Hi guys! After some research, I made a little chart one every rich man's favorite camera, the Alexa, which digs into how it really works. So I thought I might as well share it. The content holds true for every Alexa I believe, since they all have pretty much the same sensors and algorithms (even the Alexa 65 I assume, who consists of three Alexa sensors merged). On the right of the table are the only available curves to my knowledge, the LogC ones. Every other curves isn't, but should be available, so here they are. Where did I find this stuff ? A combination of gathering bits of informations here and there, and the ability to think like an Arri R&D guy I guess (they always seem to make the most logical choices in sensor architecture). CLICK and enjoy for zoomable chart.
  4. Tom Yanowitz
    What's up with all the Indian DPs and their stupid topics lately ?
  5. Tom Yanowitz
    One quick way to picturing the "trade off" is shooting an actor with some pratical lights in an overall very dark room. Let's say you can compensate the fstop to get a constant +1 reading on the skin tones as you change the ISO. ISO 200 : you'll get details in the shadows but the practcals will be burnt. ISO 800 (for example) : you'll get less details in the shadows, but the practicals will have details. The skin tones will be cleaner at 200 (not sure we can spot the difference though). (I couldnt seem to edit my previous post anymore)
  6. Tom Yanowitz
    So a couple of things. - An image sensor doesn't have a base sensitivity. (The only "real parameters" it has are the size/surface of its photosites and its bitrate) - Middle grey is more or less randomly picked by the manufacturers and is pretty much useless in the digital world. - Placing your optic fstop according to the incident reading is getting old (unless you really don't have any color grading planned and shoot 709, which can be considered as in-camera grading) - I don't think "ISO" (as the group of people in charge) has done a thing yet for digital. Some say the Alexa native ISO is 800, some say 200, some say 640 some say 1000. Everybody's wrong, it would be too easy a thing to have a magic number that gives a better image thant the other. But for every digital camera, if you expose the old way (incident reading, caucasian skins at +1 etc etc..) then the lower the ISO the cleaner the image. Why ? because you allow more light to hit the sensor. The more light the better. That said, the more you expose, the closer your highlights get to the clipping point of the sensor, so you have to check for blow out highlights more. It's a trade off, low and high ISO both have their good and bad. So there's no one definitive answer, that's why it's better to choose your ISO/EI almost case by case depending on the scene, and it's more interesting that way (compared to lazily sticking 24/7 with what the manufacturer told you was good). I hope I don't offend any of the great DP's out there but when's something is not well understood you might as well point it out.
  7. Tom Yanowitz
    What is love ?
  8. Tom Yanowitz
    So for me your graph shows how we wish sensors would work ideally. But didn't we settle some posts ago that any mapping like this one comes post-capture, and the only possible graph for the actual capture is a straight line from (0,0) to the clipping point ? (by the way with what program do you draw graphs like thise one, I need one :) )
  9. Tom Yanowitz
    For some reason I can't edit my previous post anymore. Ok so your example Ivon is 40,960 photons > 1023 as a start fair enough. So moving on : 40,960 photons mapped to 1023 (1024-1, max value in 10bit) 20.480 to 511 (512-1..) because the sensor being in linear, if you divide the photons by to you also divide the value by two. 10.240 to 255 (256-1..) 5.120 to 127 2.560 to 63 1.280 to 31 640 to 15 320 to 7 160 to 3 80 to 1 40 to 1 20 to 1 10 to 1 5 to 1 So it started to "clip" in the blacks at 80 photons, which is 9 stops under 40.960 photons. So theoretical-DR = 9 if Bitdepth = 10, doesn't matter the number of "photons" needed to make the sensor clip in the whites.
  10. Tom Yanowitz
    I just explained in my last post why DR is determined MAINLY by Bit Depth, that DR can in fact only be inferior to bit depth. (Capture bit depth mind you, not storage bit depth). They're just no way (for a camera with a linear ADC) that you can have a 14DR with a 10b ADC for example, either in theory or practice. And you, Carl and Ivon, keep arguing that they aren't linked in anyway. I don't get it. EDIT: I wrote this before Iron's last post
  11. Tom Yanowitz
    Yes I do think that's the case. It couldn't be 14.5 stops of DR without their weird cool "2x14bit = 16bit" sensor before creating 12bit raw or prores files. I'm curious to know the bitdepth of these cameras that claim 17, event 20+ stops of DR.
  12. Tom Yanowitz
    So I agree with most of what you said in this thread... except for the point 1) How could DR possibly be superior to the bit depth ? We both seem to agree on how the light information is then stored in different values. So going back to my "1B" graphs (the one that looks like a 2^x cuve) Say the sensor stores data in 10b Lets sayt its clipping point arrive when it receives 1023 photons. 1023 photons => mapped to the 1023 value (or 1111111111 as I didn't get how speaking in base 10 or 2 would change anything) 1022 photons => value 1022 ... 512 photons => value 512. One stop captured so far. Ok interesting info : every camera sensor has half of its values talken by the brighest stop of light. A stop that hardly anyone uses (expect the ETTR geeks like me), fearing blown out highlights. 256 photos => value 256. Two stops captured so far/ 128 => 128 => 3 stops 64 => 64 => 4 stops 32 => 32 => 5 stops. As we progress what we might call "low light stops" are getting fewer and fewer gray scale values for themselves. One reason why underexposing is a bad idea. 16 => 16 => 6. Ok only 16 values, at this point we enter the zone of the "probably unusable stops" (unless you're going for an image with deep blacks with no details) 8 => 8 => 7 stops 4 => 4 => 8 stops. What I might call the practical DR stops as the number of gray scales values per stops becomes ridiculous. 2 => 2 => 9 stops (theoretical stops, not anywhere near practical stops) 1 => 1 => 10 stops then eveything will be mapped top the 0 value. And now even the theoretical DR stops. That's why I maintain that DR < Bitdepth. And then what a log function will do (with some tweaking) is 1023 photons => value 1023 512 => 922 256 => 820 128 => 718 64 => 616 32 => 514 16 => 412 8 => 310 4 => 208 2 => 106 1 => 4 So we LOSE gray scales values in the highs (last stop before clipping being shrinked from 512 values to 102) while shadows are'upscaled' (as one would upscale from 720p to 1080p). As we can see in my 2B graph from the earlier post, Arri (and the others) don't go for the straight line (like in my example) but instead add a roll-off (I don't remember the word) to the shadows (stops furthest away from the clipping). That's probably because upsacling a stop that was ridiculously capture with 2 or 4 gray scale values (like my 8 and 9th stops in my example) probably looks awful when upscale to 102 values. Plus then don't map their "1023 photons" to the max value, and that allow them to have different possibility of mapping (creating the ISO/EI option in camera)
  13. Tom Yanowitz
    I'm quiet happy to have revived this topic people have this debate ! You guys have quiet a lot to say about it, it's great (though a bit confusing). Do we all agree on these 4 kinds of response curves for most professionnal cameras : 1A and 2A will have something linear (like lux) on the x-axis so basically that's how the camera and computers see things : 1/ y-axis = Coded values by the ADC pre-processing For the Alexa example, that'd be how the 16bit RAW file would like (if it was recorded) 2/ Curve post-LogC encoding (either 12bit raw or 10b or 12b prores). It's basically the values of 1A after going into a mathematical function involving their log. NOW 1B and 2B are the exact same curves, only we switch to a log scale on the x-axis (ex : stops of light hitting the sensor). Because the log scale is what we're used to (as humans). 1B/ 2B/
  14. Tom Yanowitz
    Well I think it's pretty hard to do otherwise. Apparently the Nikon D2X has some processing prior to the ADC (for white balance apparently) http://www.dpreview.com/glossary/camera-system/sensor-linearity The axiom camera might feature a non linear response as well apparently https://apertus.org/axiom-alpha-imagesensorcurious to see how that turns out. But basically ever camera that has "log encoding" has one to counter their linear response.
  15. Tom Yanowitz
    That might change in the future, but as of now every digital camera sensor reacts proportionnaly/lineary to the number of photons. So the curve of every camera as far as RAW data is concerned is like this : This one if you use a linear x-axis (like number of photons, lux etc..) Or like this if you wan a log x-axis (stops of ligth for example) So thats how RAW is stored and we cant do much about it (excpet non linear sensor, no yet available). All the encoding like logC comes from applying mathematical fuctions to this RAW data, and it's purely digital (nothing analog to it) For example if you want to see your file with the IE 800 logC curve : arri applies something like this : 0.25*log10(5.56*x+0.05)+0.39. And then we get something where every stop has roughly the same number of bit values than the others. So again the dynamic range is limited by the number of time you can divide by two starting from the clipping value: 4096, 2048, 1024 etc etc etc .. 1
  16. Tom Yanowitz
    (let me dig some old topics) I disagree and would go as far as to say that all camera sensors respond in a linear way to the photons. And so yes, any camera can't have more stops of dynamic range than its bitrate at the ADC level. exemple for 10bit : "C "(clipping stop) : 1023 (recorded bit value) C-1 : 512 C-2 : 256 C-3 : 128 C-4 : 64 C-5 : 32 C-6 : 16 C-7 : 8 C-8 : 4 C-9 : 2 C-10 : 1 The last stops being quiet unusable (only a couple code values for each stops), you can say DR is always inferior to the bitrate of the camera. Alexa hase 14 stops on 12bit ? No Alexa records two times at 14bit creating a 16bit file before being downscaled to 12bit. Theses numbers also explain "why the log format" On a log (stops of light) graph, the numbers above create a 2^x curve (each number being the double as the one before it) And log(2^x) = a*x so you're back with a linear curve witch makes the file easier to look at (that's why arriraw files are always displayed with a logC curve).
  17. Tom Yanowitz
    One of the best tools for exposing with Arri cameras for me is the False Color for extreme highs and shadows (a bit less interesting for midtones as you don't HAVE to expose skin tones in pink or whatever to get what you want). Where to put middle gray isn't as important as people think it is I think, especially if grading is part of the workflow (which it is if you record LogC) One thing to note is that waveforms, histograms and even false color adapt to the kind of display you're outputing so you have to look at those tools with the recorded mapping (log or rec709 or looks) displayed.
  18. Tom Yanowitz
    Thanks a lot everyone for your great advices! (I was convinced I had replied earlier but apparently not, so sorry for that). I'm currently looking at the prices for renting or buying bleached muslin for the two days, even that is hard on a low budget ! And as for the ISO setting of the Alexa, I was thinking it would be interesting to set it higher than 800 because : - less light needed - most of my set/costumes is high key/white, and I have shots with both the big source/ceiling and actors in the frame. So I might be able to get more "details" in the celing at a higher ISO, with actors at keylight, given the Alexa gives more highlight range the higher the ISO.
  19. Tom Yanowitz
    Footcandles ? No idea I must say, I'd like to achieve a high key look at T4 or 5.6 (2.8 if I really dont have enough light) with an evenly lit 40m² set.. Color temp well it doesn't really matter I think. I'll choose the color temp that makes my white walls as neutral as possible. Yes I'd like the ceiling to be evenly lit and do have a grid above the set, and we can adjust its height. The flooded 5Ks seems like a good idea, although i'm not sure at what angle I would place them. Maybe some smaller fresnels for some spots that the 5Ks would'nt cover. And the bleached muslin seems is a good idea too, I'll just have to see how expensive 40m² is.. Thanks ! Yes but you're talking about bouncing light right ? I have some wide shots where we see close to all the set so I need to diffuse from above rather than bounce for the ceiling.
  20. Tom Yanowitz
    Hello everyone, I need your help. I'm currently a film school student. I will direct a very short film/exercise (one day of prep one day of shooting) in 40 days. So the mood and the set design will be a weird mix between Kubrick films (2001, Clockwork..) and THX 1138, kinda. Anyway we have this 4 by 10 meters set at school, that has no ceiling. But I have one obligation for this exercice/short : only use a 16mm lens (on the Alexa). 10 meter long room, 16mm.. I'd better need something as a ceiling. I'd want something white I could use as a diffusion so that the ceiling would be a giant soft source. So I have two questions for you experienced people : What would you consider for the ceiling (doesn't have to be realistic at all) given a very limited budget. What kind of light could be use above that fake ceiling, to create a high-key look overall (the set will be mostly pure light with some black lines to give some perspective, the costumes black and white as well) (we have tungstenes up to 4K, hmis up to 2.5k I think and kinos) I'll also have some diffused light coming from the walls (I'll remove the windows of the set and place 216 frames instead) Thanks everyone !
  21. Tom Yanowitz
    Thanks for all this information ! So basically, both my edit and grading files will look flat (mapped with a log curve) but the mid gray will have been placed to the correct level ? An other way of asking this is : will the middle gray have the same luma level in the digital file whether I shoot normally, or under/expose the negative, before I get to the grading. And by the way, what is generally that luma level for film scans ? I know that for digital cameras when we shoot log encoded raw, or compressed log, it's around 40%
  22. Tom Yanowitz
    Hello everyone, I re-up a topic that's 10 years old ! So, I have some 85 filters for a 35mm outdoor shoot at my school and decided to spot meter them. Found this : Filter, expected loss, measured loss : 85, 2/3, 2/3 (6 tenth on the meter) 85N3, 1+2/3, 1+2/3 (1 stop + 5 or 6 tenth) 85N6, 2+2/3, 2+1/2 (2 stops + 5 tenth) 85N9, 3+2/3, 3+1/3 (3 stops + 3 or 4 tenth) What do you is more probable in that sort of test : the filter is "off", or I failed at measuring them properly ? I can't remember the brand of the filters. Thanks guys !
  23. Tom Yanowitz
    Hi everybody, So at my school I'm going to have the luck to be the DP of a small short in 35mm. The worflow is this : Shoot 35mm the negative is developped and transfered in both mediocre HD for editing and decent 2K for grading and finishing. I know more or less how contrast works when you get a positive from the negative, that is : shooting the same negative at various EI + the positive give different looks (deeper blacks, more or less contrast in the shadows, less grain, more saturation etc etc) So I wonder how this applies when you scan the negative directly. I'll have a 500T, what if I rate it at 250 ? The guy from the lab told be he scans according to the middle gray chart (if we shoot one) What about things like blacks, whites, contrast, visible grain, saturation ? Thanks guys !
  24. Tom Yanowitz
    Thanks a lot Bruce and Dylan ! I will try this stuff out when I get a chance ! And then hopefully study Log curves too ! My end goal is more creative than technical. Art Adams talks about 'the sweet spot' of the dynamic range. Which implies (as for Zone System in film) that we should try to expose everything in the frame (or at least everything where we want details) in the zone of the response curve where the slope is the highest, instead of using on-set the full dynamic range from black clip to white clip. I think this also applies to Log C (at least of the blacks) I got interested in this both via Art Adams' posts and this zone system advice video. A lot of people say 'expose closest to the final look' but this video suggests we can actually better contrasty/low noise blacks by placing them in an unpleasant-looking way at -3 for the Alexa for example, and then getting them to the desired level in grading. What do you guys think about this way of doing ? Cause I've suggested the idea on a couples of projects at my school and it's considered 'too safe, without any risk taken' which I guess means anti-artistic.
  25. Tom Yanowitz
    Thanks David, Art Adams is a great read, I learned a lot through his posts ! I think maximum white in Log C depends on the EI. To quote Art : the MAXIMUM WHITE VALUE changes in LogC. As the ISO drops, the MAXIMUM WHITE VALUE lowers. At ISO 200 it’s 85%. At 1600 it’s 100%. At 800 it’s 95%. At 400 it’s 905. Why is that? I don’t know. That’s just the way LogC works. As for my Rec 709 tests I got my clipped blacks at 1.5%, whites at 100%, and my mid-grey... well that depends : The value of my grey card drops slightly depending on the EI : with a spot meter, usually 40.5%, less (38.5) when we reach ISO 3200. If I set my f-stop to the incident meter reading, the grey card falls roughly between 45 and 48%. So I guess that on set I should trust the incident meter more than the spot because of this 1/2 stop difference ? The reason I wanted to do it in Rec 709 as well is because.. some projects at school are set like that (rather unfortunately)
×
×
  • Create New...