Page 14 of 56 FirstFirst ... 410111213141516171824 ... LastLast
Results 131 to 140 of 553
  1. Collapse Details
    Senior Member
    Join Date
    May 2010
    Posts
    2,530
    Default
    Quote Originally Posted by Grug View Post
    Well if the chart hasn't been done properly, then it's completely pointless anyway (and shouldn't exist, since it would literally serve no purpose), so let's assume it has been done correctly. How are you reading the chart? The first chip is a reference chip to establish the white clip point. So you count from the second chip downwards, the steps between each solid line showing your DR. Once the lines meet the noise floor (i.e. you can't see a black gap between the line on the waveform and the noise floor below it) you've reached the limits of your shadow detail (unless you're willing (and able) to distribute your images with exceptionally noisey shadows.

    Attachment 139233

    Attachment 139232
    I'd like to see those charts run at 800ISO. No silicon sensor is near peak saturation or optimum SNR at 6400.


    Reply With Quote
     

  2. Collapse Details
    Senior Member
    Join Date
    Feb 2009
    Location
    Long Island
    Posts
    7,490
    Default
    Quote Originally Posted by Razz16mm View Post
    The Canons are beefy. the BP-975 is 55Wh+ so two would outlast a 98Wh mini. Camera power consumption is around 27-30Wh. One smaller BP-955 runs my old GL-2 for 4 hours or so depending on roll time.
    There is a separate DC in plug for external power that accepts higher voltage batts plus several third party adapters for the battery mounts planned so lots of flexibility. The camera is not designed to power external accessories given its primary design purpose. Those will need their own batteries or larger V-mounts with aux power ports.
    Remains to be seen if I can afford a Komodo for a serious hobby cam, but the nice thing about it compared to other cameras is I don't need anything but the body, a couple of lens adapters and some media. Already have everything else in the kit.
    That 98Wh mini was just an example for the cost and size.

    A regular V-mount could have 200-300Wh for the same price if size/weight didn't matter.


    Reply With Quote
     

  3. Collapse Details
    Senior Member James0b57's Avatar
    Join Date
    Dec 2008
    Posts
    5,281
    Default
    Quote Originally Posted by Razz16mm View Post
    I'd like to see those charts run at 800ISO. No silicon sensor is near peak saturation or optimum SNR at 6400.
    They are raw cameras that are only ever at base iso.

    6400 is just pushing in post. They likely exposed for native iso and pushed them in post. Other than the Gemini, Red cameras only actually have one iso setting, the rest is metadata.
    Last edited by James0b57; 05-30-2020 at 10:37 AM.


    Reply With Quote
     

  4. Collapse Details
    Senior Member ahalpert's Avatar
    Join Date
    Apr 2011
    Location
    NYC
    Posts
    1,040
    Default
    Quote Originally Posted by James0b57 View Post
    They are raw cameras, so they are only ever at base iso.

    6400 is just pushing in post. They likely exposed for native iso and pushed them in post. Other than the Gemini, Red cameras only actually have one iso setting, the rest is metadata.
    What does that mean in effect? That as you push them, the noise floor rises and the stops above the floor get compressed into a smaller range at the top of the output brightness but the processing prevents the top stops from all blowing out and thus the original DR is preserved plus some noisy stops at the bottom that become brighter?


    Reply With Quote
     

  5. Collapse Details
    Senior Member James0b57's Avatar
    Join Date
    Dec 2008
    Posts
    5,281
    Default
    Quote Originally Posted by ahalpert View Post
    What does that mean in effect? That as you push them, the noise floor rises and the stops above the floor get compressed into a smaller range at the top of the output brightness but the processing prevents the top stops from all blowing out and thus the original DR is preserved plus some noisy stops at the bottom that become brighter?
    I think you described it better than i could have!

    Will just post this inaccurate graph here, perhaps for the lurkers, if there are any anymore, as at least it shows how many cinema cameras do iso. Canon seems to add gain along the way, so, i believe Canon is a type of “true” iso, and this graph does not apply to Canon C. Also, dual iso has become popular, and those cameras have two “base” iso settings. But within each base iso, they shift iso using pushing and pulling.

    However, Alexa and most Red’s tend to do iso like this:

    Iso800 - cleaner shadows, “thicker” mids, but less highlight roll off.
    Iso1600 - noiser shadows, “thinner” mid tone tonality, longer highlight roll off

    The clipping points aren’t changing, just where the operator chooses to land the exposure. Iso 1600 boosts the brightness of the mids and allows for more highlight rolloff. Or pulling down to iso 400 cleans up the shadows but leaves less elbow room in the highlights making a “video” look up top sometimes.


    There is a good interview with Rob Maclachlan who DP’d the ‘Red Wedding’ episode for the HBO series “Game of Thrones”. In the interview, he says that they shot the wedding by candlelight at iso1600, which not only gave them an extra stop to work with, but extended the highlight range keeping the candle flames from clipping.



    When lowering the iso, the operator is using less of the shadows and working their back against the highlight clipping point. Whereas raising the iso, the operator it lifting the shadows and exploring the murky waters of noise, which tends to have more wiggle room for a stop or two. But highlights are that impassable barrier, there is not wiggle room with highlight clipping on digital, unless you have a lot of dynamic range for a film like infinite roll off. Wally Pfister over exposed the film in shooting the Joker interrogation scene in the Dark Knight by like 5 stops and they pulled it back down.
    Last edited by James0b57; 05-30-2020 at 03:23 PM.


    Reply With Quote
     

  6. Collapse Details
    Senior Member James0b57's Avatar
    Join Date
    Dec 2008
    Posts
    5,281
    Default
    An extended highlight roll off mimics a more analogue feel, which i feel is more similar to how our eyes see things. There is a dampening as analogue reaches the limits in audio and visual mediums that digital doesn’t quite have until there is more headroom than needed. So, we may only “need” 12 stops of DR, but we need an extra 6 stops on top of that for better roll off in high contrast situations.


    Reply With Quote
     

  7. Collapse Details
    Default
    So, we may only “need” 12 stops of DR, but we need an extra 6 stops on top of that for better roll off
    18 stops is an awful lot of dynamic range. An f-stop represents twice as much light. So 11 stops is not 10% more than 10 stops. It's 100% more. 18 stops is 2⁶ as much light as 12 stops --- that's 64 times. The image sensor's photodiodes must have 64 times less noise or 64 times the well capacity (or a mix thereof).

    in high contrast situations
    High-contrast situations are 8-12 stops, when I and others have measured them.

    • The difference between a pure white fabric and the blackest dye is 5 stops.
    • The difference between direct light and shadow varies. It could be 2-5 stops, if you remain outdoors.

    So if you snap a pic of a scene where someone in black is under a tree, and someone in white out in the open, you're talking 8 stops.

    In this article, Dynamic Range of an Image, Roger N. Clark measured 11. He has white things in direct sun and black things under a patio umbrella. I could imagine darker shade. Then again, this is pretty extreme, since his house is behind him or to the right, blocking any light from that direction.

    ---

    Hollywood sets typically are lit to around 7 stops. When they shoot outside, they bring lights, partly to lower the dynamic range. You might see lights shining up into trees from underneath.

    ---

    Or take slide film. It had a dynamic range of 7-8 stops, yet still looked nice. Search Google Images for "velvia", for example.

    ---

    Your screen has a dynamic range of 6-8 stops. Movie theater screens are 9-10 stops --- and that was the ideal goal, according to some Academy test instructions.

    Real-world scenes are 7-12 stops, but you must always "compress" the dynamic range to the 6-8 stops of the viewer's screen. That affords you some of the leeway you were talking about for highlight roll-off.

    ---

    I feel that most people decide how much dynamic range they want in camera like this:

    1. Find an image you like.
    2. Find out which camera took that image.
    3. Look up that camera's dynamic range.

    This is unscientific. It's subject to a lot of variables. There are many other things that can affect the look, even if you're considering only highlight roll-off:

    • Was the image recorded compressed or raw? The Canon 5D looks drastically different depending on whether you use its stock image processing and codec, or if you bypass the whole shebang and record raw with Magic Lantern. Its raw imagery looks great, even though it has "only" 11.7 stops.
    • Who graded it, and how? Again, from looking at Fujifilm Velvia, I believe that most cameras could look good if I had access to the raw image and graded it myself.

    ---

    If a camera has 12-14 stops, and you only ever use 7-8 of them for given scene, where do the other stops go? I think most of them are going to ISO variability.

    I used to wonder how a single sensor could have multiple ISOs. If you crank up the amplifier, you crank up noise right along with signal. So you're not changing the dynamic range, just the midpoint.

    A camera with a dynamic range of 13 stops, shooting scenes of 8 stops, would have 5 more stops to play with --- the difference between ISO 800 and 25600.
    Last edited by combatentropy; 05-30-2020 at 11:48 PM.


    Reply With Quote
     

  8. Collapse Details
    Senior Member ahalpert's Avatar
    Join Date
    Apr 2011
    Location
    NYC
    Posts
    1,040
    Default
    Dunno exactly how many stops I need. All I know is I run into situations where I can't keep all the detail I want.

    The biggest problem for me is having incident light sources in the frame. The sun, clouds, practicals. And also, if you expose something at the edge of your contrast range, it usually won't have as rich a tone and color. So it's great to have more range than you need for simply registering detail.


    Reply With Quote
     

  9. Collapse Details
    Senior Member
    Join Date
    Feb 2009
    Location
    Long Island
    Posts
    7,490
    Default
    The best that you could hope for is a true 13.5 - 14 stops because anything over 15+ is questionable in mostly everything else out there. Probably not in the budget anyway.

    So a solid 14 stops would be nice until more cameras advance and add an additional stop in the next 2-3 years.


    Reply With Quote
     

  10. Collapse Details
    Senior Member Jim Arthurs's Avatar
    Join Date
    Mar 2005
    Location
    Colorado Springs, CO
    Posts
    860
    Default
    Quote Originally Posted by NorBro View Post
    The best that you could hope for is a true 13.5 - 14 stops because anything over 15+ is questionable in mostly everything else out there.
    This old animation I created for Convergent Design is still useful in understanding sensor characteristics including a pretty cool visualization of a sensor's linear response...

    https://www.youtube.com/watch?v=aQJHqDGRoa8&t

    As an aside, I used to do back-light motion graphic animations ( you know... "Stranger Things" style titles but using the original multi-exposure optical techniques) in the 80's, using 5247/7247; a tungsten balanced 100iso stock. The "formula" for getting pure black and pure white was easy... 3.5 stops under middle grey was black (completely clear on the negative), and 3.5 stops over middle grey was "white"... you could push just a bit more into the highlights when exposing a lens flare effect, or a longer exposure to get a "white on white" effect, but that was hard to discern in the telecine bay of the day.
    Jim Arthurs


    2 out of 2 members found this post helpful.
    Reply With Quote
     

Page 14 of 56 FirstFirst ... 410111213141516171824 ... LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •