Thread: Canon EOS C70

Reply to Thread
Page 4 of 9 FirstFirst 12345678 ... LastLast
Results 31 to 40 of 84
  1. #31  
    Senior Member Alex Lubensky's Avatar
    Join Date
    Nov 2016
    Location
    Kiev, Ukraine
    Posts
    257
    Yup, it's exactly the same latitude as the latest C300. There are plenty of real-world tests out there (for example this one from Potato Jet on minute 4:30). Nothing close to Alexa (they claim two stops more btw). Yup, cleaner shadows (noise reduction?), which might make sense for non-lit scenes with 120fps.

    Btw 10 bit is 1024 color/tonal values, while 16bit is 65,536 color/tonal values. Literally, it's a 64 times difference. 10bit can look good enough out of the box, but dipping into commercial color grading - na-ha.

    While this camera totally makes sense for event/cheap tv productions, it totally lacks crew capabilities and codecs for narrative/commercial productions. Different markets. Each has it's own marketplace.
    Last edited by Alex Lubensky; 09-29-2020 at 02:43 PM.
    Reply With Quote  
     

  2. #32  
    Quote Originally Posted by Alex Lubensky View Post
    I don’t think it’s better in DR than Komodo. It’s a +4 highlight stops camera compared to +6 Komodo. Highlight latitude is way more important in feature world. Also no SDI, no dtap etc etc etc. clearly made for one man band. Totally different use case scenario
    It's about total dynamic range. If you think it has 2 stops less than Komodo in the highlights, then add 2 stops of ND with one button, bro.

    Now that the cameras are clipping at the same place, what's the rest of the image like? What's the color / detail / noise character like in the highlights, midtones and shadows?

    I agree lack of SDI isn't great. Though a lot of the time I'm viewing a feed from a Teradek these days. With a decent rig locking it down, HDMI can work. For D-tap, just get a battery with D-tap.

    If the sensor were equal then sure I'd go for the more pro camera, but I wouldn't want to sacrifice picture quality until I get to at least Alexa range. I mean you can rent an Alexa Mini for $500 per weekend or 3-day week in LA. So yeah between Alexa Mini and LF on the high end and the C70 and the a7s3 and S1H on the low end, I'm STRUGGLING to see why I'd rent, let alone buy, a Komodo.

    Quote Originally Posted by Alex Lubensky View Post
    Yup, it's exactly the same latitude as the latest C300. There are plenty of real-world tests out there (for example this one from Potato Jet on minute 4:30). Nothing close to Alexa (they claim two stops more btw).
    I guess this just shows how differently people see things :)

    In that Potato Jet video, to me the C300 mk3 has a TON more latitude in the shadows than the FX9. No way that is just "noise reduction", there are clearly extra usable stops there from the Alexa-like dual gain architecture.

    My DP buddy owns an Alexa Mini, a C300 mk2 and now a C300 mk3 and does unbiased side by side tests. He has no dog in the fight, he owns all of 'em. C300 mk3 / C70 sensor is impressive.

    Quote Originally Posted by Alex Lubensky View Post
    Yup, cleaner shadows (noise reduction?), which might make sense for non-lit scenes with 120fps.
    Huh? I often need latitude when there's plenty of light. Day exteriors for example! Surely you don't believe you only need latitude for scenes without light? Weird statement. I've been to Ukraine, it's not overcast all the time there, you have beautiful sunny days and countryside and cities which need nice shadow detail :)

    Quote Originally Posted by Alex Lubensky View Post
    Btw 10 bit is 1024 color/tonal values, while 16bit is 65,536 color/tonal values. Literally, it's a 64 times difference. 10bit can look good enough out of the box, but dipping into commercial color grading - na-ha.
    Don't listen to silly RED marketing (I think Sony do it too)! You're not getting a clean 16bits of data out. The last couple bits of a linear 16bit ADC on current-gen sensors are noise and the top bits aren't going evenly to the stops at all.

    I don't think you understand linear vs log and the way bits from the ADC get distributed across stops.

    10bits of tonal values on a good log curve is great. Bear in mind the C70 is probably feeding that 10-bit log signal from dual 14-bit or maybe even 16-bit ADCs. It's not using a 10bit ADC, that would be super weird.

    I've worked on tons of high-end shows that were 10bit log Cineon DPX 2K files. Like The Avengers for example. Looked great.

    Bruce Allen
    www.bruceallen.tv
    Reply With Quote  
     

  3. #33  
    Senior Member Andy Roberts's Avatar
    Join Date
    Dec 2008
    Location
    Southeast US
    Posts
    1,069
    We are definitely at the point where the difference between some of these cameras is so minuscule in regard to pixel peeping, that debates over which camera is better is tiresome. Pick the features and capabilities you need for your style and workflow and go shoot.
    Reply With Quote  
     

  4. #34  
    Quote Originally Posted by Andy Roberts View Post
    We are definitely at the point where the difference between some of these cameras is so minuscule in regard to pixel peeping, that debates over which camera is better is tiresome. Pick the features and capabilities you need for your style and workflow and go shoot.
    Exactly.
    Reply With Quote  
     

  5. #35  
    Quote Originally Posted by Andy Roberts View Post
    We are definitely at the point where the difference between some of these cameras is so minuscule in regard to pixel peeping, that debates over which camera is better is tiresome. Pick the features and capabilities you need for your style and workflow and go shoot.
    I agree with you in terms of resolution. Almost everything is good enough.

    In terms of latitude and color, it's not always noticeable in RED's best cameras but it's not exactly pixel peeping to see the difference when stuff is clipped or noisy in the shadows or weird color! You can do great work if you have a top colorist but it's just a tougher starting point than Alexa and you can see most people never quite grade it out.

    It just comes across as cheap and digital feeling, even if in a 720p video. Maybe I'm sensitive to it after having worked with RED since 2007 but I have noticed other directors, commercial clients, etc don't like it either.

    I do hope that we stop pixel peeping and start looking at overall image.

    I am very happy that the horrible "grid of red balls" that messed up the flares in RED is gone in Komodo at least!

    Bruce Allen
    www.bruceallen.tv
    Reply With Quote  
     

  6. #36  
    Senior Member
    Join Date
    Oct 2009
    Posts
    1,183
    Quote Originally Posted by Alex Lubensky View Post
    Yup, it's exactly the same latitude as the latest C300. There are plenty of real-world tests out there (for example this one from Potato Jet on minute 4:30). Nothing close to Alexa (they claim two stops more btw). Yup, cleaner shadows (noise reduction?), which might make sense for non-lit scenes with 120fps.

    Btw 10 bit is 1024 color/tonal values, while 16bit is 65,536 color/tonal values. Literally, it's a 64 times difference. 10bit can look good enough out of the box, but dipping into commercial color grading - na-ha.

    While this camera totally makes sense for event/cheap tv productions, it totally lacks crew capabilities and codecs for narrative/commercial productions. Different markets. Each has it's own marketplace.
    I don't entirely disagree. I'm not going to buy this camera, but that doesn't entirely correlate with my own experience or with other direct comparisons:

    https://vimeo.com/402727266

    The spec for Canon Log 2 is +6.3 stops of highlight dynamic range at base 800 ISO, it's either +7.3 or +7.8 stops on the Alexa depending on what generation of sensor/firmware. So yes, that's a +1.5 stop (roughly two stop) difference, but +6.3 stops is still really good and similar to F55 or Red at 800 ISO. I do think Gemini etc. has better dynamic range than a C200, but the DGO here might bridge the gap...

    Fair enough regarding 10 bit vs 16 bit, but it's logarithmic encoding vs linear encoding, which is entirely different. Arri raw is 12 bit log, for instance, 14+ stops distributed among 12 bits. By contrast, an entire 1/2 (8 bits) of a 16 bit linear package is taken up by one stop of dynamic range... and both Arri raw and red raw have no issues with tonality. 10 bit log recording is pretty robust.

    Also that video isn't a comparison against an Alexa. And it holds to +4, but caucasian skin tones are +1 stops over 18% gray and a white card I think is at least +2 stops over 18% gray. So if the image holds to four stops over on an incident meter, that might very well be equivalent to +6 stops over on a spot meter. I haven't seen footage from the Komodo, but unless it's much better than any other Red camera, the highlight dynamic range should be comparable to the C70. (Roughly +6 stops over at base, meaning caucasian skin, which is about a stop over 18% gray, is starting to clip when lit to +5.)

    This camera isn't for me, either, and the Alexa probably does have 1-2 stops more highlight dynamic range (than, frankly, anything other than film, not just this camera). But I've worked with the C200 and Canon Log 2 and the +6.3 spec is accurate and competitive with every camera other than film or Alexa. This camera isn't for me either, and for some of the reasons you mention, but highlight dynamic range is not one of them UNLESS you're comparing against Alexa or film.
    Last edited by Matt W.; 09-28-2020 at 12:10 PM.
    Reply With Quote  
     

  7. #37  
    Senior Member Alex Lubensky's Avatar
    Join Date
    Nov 2016
    Location
    Kiev, Ukraine
    Posts
    257
    Quote Originally Posted by Bruce Allen View Post
    It's about total dynamic range. If you think it has 2 stops less than Komodo in the highlights, then add 2 stops of ND with one button, bro.

    Now that the cameras are clipping at the same place, what's the rest of the image like? What's the color / detail / noise character like in the highlights, midtones and shadows?

    I agree lack of SDI isn't great. Though a lot of the time I'm viewing a feed from a Teradek these days. With a decent rig locking it down, HDMI can work. For D-tap, just get a battery with D-tap.

    If the sensor were equal then sure I'd go for the more pro camera, but I wouldn't want to sacrifice picture quality until I get to at least Alexa range. I mean you can rent an Alexa Mini for $500 per weekend or 3-day week in LA. So yeah between Alexa Mini and LF on the high end and the C70 and the a7s3 and S1H on the low end, I'm STRUGGLING to see why I'd rent, let alone buy, a Komodo.



    I guess this just shows how differently people see things :)

    In that Potato Jet video, to me the C300 mk3 has a TON more latitude in the shadows than the FX9. No way that is just "noise reduction", there are clearly extra usable stops there from the Alexa-like dual gain architecture.

    My DP buddy owns an Alexa Mini, a C300 mk2 and now a C300 mk3 and does unbiased side by side tests. He has no dog in the fight, he owns all of 'em. C300 mk3 / C70 sensor is impressive.



    Huh? I often need latitude when there's plenty of light. Day exteriors for example! Surely you don't believe you only need latitude for scenes without light? Weird statement. I've been to Ukraine, it's not overcast all the time there, you have beautiful sunny days and countryside and cities which need nice shadow detail :)



    Don't listen to silly RED marketing (I think Sony do it too)! You're not getting a clean 16bits of data out. The last couple bits of a linear 16bit ADC on current-gen sensors are noise and the top bits aren't going evenly to the stops at all.

    I don't think you understand linear vs log and the way bits from the ADC get distributed across stops.

    10bits of tonal values on a good log curve is great. Bear in mind the C70 is probably feeding that 10-bit log signal from dual 14-bit or maybe even 16-bit ADCs. It's not using a 10bit ADC, that would be super weird.

    I've worked on tons of high-end shows that were 10bit log Cineon DPX 2K files. Like The Avengers for example. Looked great.

    Bruce Allen
    www.bruceallen.tv
    It does seem you have totally missed my point. Good camera for event/wedding/tv shows. Bad camera for crews for a lot of reasons. Totally different markets.

    I'm glad you've been to Ukraine, I've been to a lot of places too. Dynamic range is dynamic range no matter of a place, what I was telling - in narrative highlight latitude matters more than shadow one. Both matter, but if I had to choose - highlight one would matter more to me. It's not as simple as dropping a couple ND's and placing the highlight clip at the same point. You still have to deal with brightness>color>saturation relation on post, and you'll end up with various color shifts this way, especially if you go some extreme values like +3 stops. The middle-grey mapping this way is a weird decision in my opinion - I don't ever need to cope in shadows more than one stop. If you really need to - this means you totally messed up your lighting. I've mentioned the only use case which is suitable to me already before - slo-mo. Here I agree, lower stops are up to use.

    I don't care about marketing at all, I choose an instrument to work with based on my on-set experience. 100% of the work I'm doing is a crew job and a rental one, so I'm not really biased to any brand name out there. I've done some work on C300 MKI & II a couple years ago, and they are nice for some occasions, like concerts etc. The AF on Canons is a beast, but it's not a matter of a choise to me, where I have a full crew to work with.

    Besides, speaking of 16 bit range I was speaking of Arri Raw.
    Reply With Quote  
     

  8. #38  
    Senior Member Alex Lubensky's Avatar
    Join Date
    Nov 2016
    Location
    Kiev, Ukraine
    Posts
    257
    Quote Originally Posted by Matt W. View Post
    I don't entirely disagree. I'm not going to buy this camera, but that doesn't entirely correlate with my own experience or with other direct comparisons:

    https://vimeo.com/402727266

    The spec for Canon Log 2 is +6.3 stops of highlight dynamic range at base 800 ISO, it's either +7.3 or +7.8 stops on the Alexa depending on what generation of sensor/firmware. So yes, that's a +1.5 stop (roughly two stop) difference, but +6.3 stops is still really good and similar to F55 or Red at 800 ISO. I do think Gemini etc. has better dynamic range than a C200, but the DGO here might bridge the gap...

    Fair enough regarding 10 bit vs 16 bit, but it's logarithmic encoding vs linear encoding, which is entirely different. Arri raw is 12 bit log, for instance, 14+ stops distributed among 12 bits. By contrast, an entire 1/2 (8 bits) of a 16 bit linear package is taken up by one stop of dynamic range... and both Arri raw and red raw have no issues with tonality. 10 bit log recording is pretty robust.

    Also that video isn't a comparison against an Alexa. And it holds to +4, but caucasian skin tones are +1 stops over 18% gray and a white card I think is at least +2 stops over 18% gray. So if the image holds to four stops over on an incident meter, that might very well be equivalent to +6 stops over on a spot meter. I haven't seen footage from the Komodo, but unless it's much better than any other Red camera, the highlight dynamic range should be comparable to the C70. (Roughly +6 stops over at base, meaning caucasian skin, which is about a stop over 18% gray, is starting to clip when lit to +5.)

    This camera isn't for me, either, and the Alexa probably does have 1-2 stops more highlight dynamic range (than, frankly, anything other than film, not just this camera). But I've worked with the C200 and Canon Log 2 and the +6.3 spec is accurate and competitive with every camera other than film or Alexa. This camera isn't for me either, and for some of the reasons you mention, but highlight dynamic range is not one of them UNLESS you're comparing against Alexa or film.
    A can't claim or judge anything I haven't tested myself certainly. I'm judging over the tests I've seen over the web - none of those cameras are available at local rentals, except for FX9 (played around for a while, but didn't do a proper test). You're totally right about caucasian skin tone and the face highlight, but it's not better compared to FX9, and we all know FX9 DR isn't anything near Venice/Alexa/Gemini range. Again, I see no reason to pull back shadows on Canon, never really needed to, and even wanted to, so to me it's what you see - is what you get.

    I've done tests myself of Alexa Mini, Mini LF, Venice, Gemini, Helium, MX, Monstro, F55 in recent year or so. The IQ is there and it's pretty much 1-3 stops difference between those, you can totally work with any camera and be happy with the result.

    Also I've shot several projects with C300 mk I & II, fx7 mk I & II, EVA, Varicam LT. Each has it's strong sides and weak sides - I loved the c300 ergonomics, hated the menu, noise pattern, oversaturated blues. Eva has nice colors and the image to me was the best of those three (canon, sony, panasonic). Both Eva/Varicam had totally bad weight/balance and ergonomics. FX7 wasn't my favourite camera at all - I didn't like the colors out of the box, the ergonomics where not my favourite. Sony menu system is a mess. But, it's all about past cameras and the FX9 seem to have much better color out of the box, the best one out of this range of cameras to my eye.

    Literally it's now about apples to oranges - just pick a thing comfortable to your use case. I just can't agree with a claim that this sensor has a better DR compared to Komodo/Gemini. It clearly does not.
    Reply With Quote  
     

  9. #39  
    Senior Member rand thompson's Avatar
    Join Date
    Aug 2011
    Posts
    7,582
    Another article from Nofilmschool explaining the Dual Gain Output of the Canon C70

    Canon's Dual Gain Output Image Sensor Explained

    By Nofilmschool

    Article
    https://nofilmschool.com/canons-dual...nsor-explained
    Reply With Quote  
     

  10. #40  
    Quote Originally Posted by Alex Lubensky View Post
    I'm glad you've been to Ukraine, I've been to a lot of places too. Dynamic range is dynamic range no matter of a place, what I was telling - in narrative highlight latitude matters more than shadow one. Both matter, but if I had to choose - highlight one would matter more to me. It's not as simple as dropping a couple ND's and placing the highlight clip at the same point. You still have to deal with brightness>color>saturation relation on post, and you'll end up with various color shifts this way, especially if you go some extreme values like +3 stops. The middle-grey mapping this way is a weird decision in my opinion - I don't ever need to cope in shadows more than one stop. If you really need to - this means you totally messed up your lighting. I've mentioned the only use case which is suitable to me already before - slo-mo. Here I agree, lower stops are up to use.
    This paragraph shows you don't understand how the dual gain sensor on the C70 / C300 mk3 works (also you keep saying it's for slow-mo when dual gain actually turns OFF above 60fps if using the full sensor).

    The whole point of it is that compared to the C300 mk2 which was a bit noisy in the shadows, with thie new sensor you can expose 2 more stops under to protect the highlights BEFORE messing up color / noise / detail. Because 2 stops down, then it is using data from the second (optimized for higher gain and lower-noise) ADC pathway.

    This is a similar design to Arri's ALEV sensor. The difference is Canon mapped it by default so it's the same behavior as its predecessor to just lower noise. Whereas Arri mapped it by default so you have extra highlight detail.

    But it's EASY to map things the other way on either camera - on an Alexa if you want less noise, just rate your meter at lower ISO. On the C70, just rate your meter at higher ISO to protect highlights.

    With C70, in a studio where there is controllable dynamic range you would probably still expose as-is for the least noise.

    But if shooting outside, where there could be a wide dynamic range that you don't have full control over, you can safely go down 2 stops to protect skies / highlights more than you could with C300 mk2.

    Quote Originally Posted by Alex Lubensky View Post
    Besides, speaking of 16 bit range I was speaking of Arri Raw.
    No, you are confused. Arri Raw is not stored at 16bit.

    Bruce Allen
    www.bruceallen.tv
    Reply With Quote  
     

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