Canon C300 Mark II Lab Test – Dynamic Range 2 Stops Less Than Expected

September 24th, 2015
Canon C300 Mark II Lab Test - Dynamic Range 2 Stops Less Than Expected

canon-c300-mark-ii-review-featured

The Canon C300 Mark II is here and comes with a lot of high expectations by customers and a hefty price tag of $16,000. Many are willing to put down the investment as they anticipate superb image quality in return and a new flagship cinema camera that has been several years in development by Canon. In our test lab we took a closer look at the C300 Mark II’s performance and found it less powerful than expected.

[UPDATE:] Canon responded to our article.
Read: How Canon Measured 15 Stops of Dynamic Range on the C300 Mark II

At cinema5D we conduct camera reviews and comparisons. As manufacturers are not limited in the way they advertise their camera’s performance we strive to put the numbers in perspective with unified tests on the latest cameras on the market.

Being a shooter myself I know it’s actually really hard to objectively pinpoint the performance of a camera. In this Canon C300 Mark II lab test I want to give you some insights into our findings about the new Canon C300 Mark II and show you how we tested.

canon-c300-mark-ii-review-03

The Canon C300 Mark II’s reputation

The Canon C300 Mark II has raised a lot of eyebrows since the announcement in April. The first Cinema EOS camera, the C300 has been very popular and remains the go-to option for many TV, film and independent productions. With the C300 Mark II Canon introduced numerous advancements, most notably 12-bit 2K as well as 4K 10-bit internal recording and improvements in terms of sensor technology and image processing. According to Canon:

“a 15-stop dynamic range is provided by a new photodiode design that simultaneously lowers the noise floor”.

Considering that the defacto cinema standard Arri ALEXA was announced with 14 stops of dynamic range the Canon C300 Mark II comes with a bold claim. The information surrounding the new Cinema EOS camera seems to suggest that Canon’s new sensor could actually compete with the image quality of the good old Arri ALEV III sensor that is used in all 9 versions of the ALEXA Arri has released over the years.

As with any new camera on the test bench I was curious, but eventually what I found during the lab test wasn’t what I was hoping for.

canon-c300-mark-ii-review-05

Testing the Dynamic Range

So I looked at the dynamic range. At cinema5D we measure this using a DSC labs XYLA-21, an LED-backlit transmissive chart that displays 21 stops of dynamic range. Each vertical bar represents one stop of light. This way it’s very easy to judge dynamic range just with your eyes. At the end we evaluate the recordings with a software by IMATEST that spits out a dynamic range value. There’s some more science behind it, but I’ll spare you the details.

As we are recording each ISO value with each camera using the identical very sharp Zeiss 50mm CP2 T/2.1 makro lens with interchangeable mount we can compare all cameras to each other. In our database we have about 20 cameras on record so far.

In our tests and according to our workflow it turned out the camera actually has 12.3 stops (measured) of usable dynamic range.

Usable dynamic range. What is that? That means within this range you have picture information that you can use. Anything beyond a certain “Signal-to-Noise” ratio is so noisy we think it’s unusable. I must say 15 stops of dynamic range is something I cannot find in the Canon C300 Mark II. There’s always a chance there’s some hidden setting in the menu to unleash the HDR potential of the sensor, but I couldn’t find it.

canon-c300-mark-ii-review-08

Reading comments of other camera enthusiasts it seems that many expect the new C300 sensor to hold up to the performance of the ALEXA cameras. What I see is that the C300 Mark II is still far away from matching this reference standard and almost on par with the Sony FS7 (measured at 12.4 stops).

An additional surprise was that I saw a lot of noise in the dark areas on the Canon C300 Mark II even at its base ISO of 800. While the camera has a very clean image in the brighter areas and has a really neutral tone with minimal color noise, there is a point in the dark areas where the noise kicks in strongly. When I did some test shots of natural subjects I realized that images shot on ISO 800 that are just 1 stop underexposed can quickly become a little too noisy for my taste especially in comparison to a camera like the Arri AMIRA (our reference camera). That is rather the opposite of what I expected after all the talk about the Canon C300 Mark II sensor and processing being so new and advanced. Ultimately one has to admit the 5 year old Arri ALEXA sensor is doing better…

Looking at the Test Charts

Below you can see 3 different cameras compared with the method explained above. The Canon C300 Mark II, Arri ALEXA and Sony FS7. You can see the usable dynamic range highlighted in red. The two fields (stops of light) to the left are overexposed, the range to the right is too noisy (underexposed).

dynamic-range-C300-mark-ii-vs-fs7-vs-alexa_B

It is interesting to see how close the Sony FS7 and Canon C300 Mark II are in terms of dynamic range performance, while the Arri ALEXA clearly has a wider dynamic range and much better noise performance. Furthermore you can see that the ALEXA has a much nicer highlight rolloff and a very nice “looking” noise overall even in the far blacks. The Canon C300 Mark II has very little color noise, which is nice, but at one point in the darks, the noise is getting very severe. You can see this much better in the following image. These are the same shots with gamma lifted equally so we can better see what is happening in the shadow areas, or what would happen when we push the image in post:

dynamic-range-C300-mark-ii-vs-fs7-vs-alexa-gammalift_B

Camera Settings:

  • Canon C300 Mark II: C Log 2 – Cine Gamut | ISO 800 (native) | F/4.0,5 | 4K (downscaled to 2K) 10bit
  • Arri ALEXA: Log C | ISO 800 (native) | F/2.8,5 | 2K, 12bit
  • Sony FS7: Slog 3 (EI) | ISO 2000 (native) | F/5.6 | 4K (downscaled to 2K) 10bit

I am aware that some people might have questions about our tests. Here are a few answered in advance:

  • Why did I compare 4K to 2K? I decided to shoot the Canon C300 Mark II in 4k (not 2K) for this test as I felt the image looked better when downscaled to 2K. I am aware Canon says they achieve best image quality at 2K in 12 bit. I did not see any improvement for the tests conducted. 12 bit will not give you a higher dynamic range, only better gradations.
  • Why didn’t I shoot the Sony FS7 at ISO 800 also? The Sony FS7 has a base ISO of 2000 and performs best at this speed. I did not see any improvement in dynamic range shooting at lower ISO’s. Same goes for the Canon by the way.

C300-mark-ii-horizontal-stripOne more thing I’d like to mention is that strip of light that is visible on the Canon C300 Mark II recording. In the very dark areas you can see a faint horizontal strip across the image coming from the opverexposed fields on the left. I have no explanation for that phenomenon.

High ISO’s?

The C300 has a reputation of producing nice lowlight images and the Canon C300 Mark II is said to be even stronger in that regard, with an available ISO range of up to 102,400 it sounds promising. In my initial tests however I felt that the camera also didn’t perform as good as I expected. Personally I thought I would probably not go beyond ISO 6400 for most projects, which is more or less in the ballpark of the Sony FS7’s lowlight capabilities. Other DP’s might of course go much higher.

High ISO Canon C300 Mark II vs Sony FS7

When I compared the Canon C300 Mark II and the Sony FS7 with the chart I found that indeed they look quite similar in terms of usable range. The Canon I would say performs only slightly better at the same ISO speeds. You should know though that ISO values and the way they affect the image brightness are often varying on different cameras. Thus I had to open up the aperture about 1 stop more for the Sony to get the same exposure, meaning the Sony FS7 is actually about 1 stop less light sensitive at this ISO speed.

lowlight-C300-mark-ii-vs-a7s_B

Camera Settings:

  • Canon C300 Mark II: C Log 2 – Cine Gamut | ISO 12,800 | 1/840th | F/4.0,3 | 4K (downscaled to 2K) 10bit
  • Sony FS7: Slog 3 | ISO 12,500 | 1/1000th | F/2.8 | 4K (downscaled to 2K) 10bit

canon-c300-mark-ii-review-01

Conclusion

Even though the Canon C300 Mark II seems to be quite on par with the Sony Fs7 in terms of dynamic range, I must also say the image of the C300 Mark II is more neutral and the noise is less saturated and mushy in comparison to the Sony FS7. So while unfortunately there is a lot of noise in the shadows, at least it doesn’t look so bad. Also the C300 Mark II seems to be about 1 stop more light sensitive than the FS7 at high ISO speeds.

I must also mention: Curious about the 120fps (crop) mode in 2K I found that it’s actually very soft (think 720p) making it much less usable. Personally I expected much more from the Canon C300 Mark II.

Of course there are other aspects to consider about this camera, many of which are not related to the tests discussed in this article. A big point is that the internal codecs are much stronger than on the Sony FS7 and thus it can retain details much better. Also the ergonomics are different and can be a buying argument if you have gotten used to the first Canon C300. There is built-in Genlock and a few other features professionals might enjoy. Check out this comparison of specs between the C300 Mark II’s main contenders for a basic overview.

Personally I think the FS7 and C300 Mark II can compete well in terms of overall performance, but the Sony FS7’s 240fps and other external recording options make it very competitive in regards to the Canon C300 Mark II that requires a much bigger investment. Eventually it is up to every user to decide which features and aspects of each camera’s performance is important to them.

We’ll conduct some more tests at cinema5D and keep you up to date. Subscribe to our newsletter to stay informed (no spam guaranteed), or like us on facebook if you enjoyed this article.

We hope you liked this Canon C300 Mark II lab test. Please do share your own observations and thoughts about this article in the comments.

Note: We contacted Canon to get a response about our findings. A Senior Canon Representative involved in the development of the camera came back to us and let us know that they will wait to assimilate more technical evaluations before commenting.

avp-logoThanks to AV Professional for lending us their camera.

 

Tests conducted to the best of our knowledge. Errors and omissions excepted.

198
Leave a reply

guest
1
1
- 7
Filter:
all
Sort by:
latest
 Jürgen Gutowski
Jürgen Gutowski
Member
November 5th, 2016

Danke Sebastian!

Dein Test hat mich vor einem ärgerlichen Fehlkauf bewahrt. Ich habe mich eh schon gewundert, das die Kamera bereits um 1/3 im Preis gefallen ist.

Member
March 11th, 2016

Cinematographer Geoff Boyle confirms, that is was a very good decision for us to buy the Canon C300 Mark II. to be our 4 K cinema camera.
https://vimeo.com/158017639

Member
February 22nd, 2016

Sorry for the assumption, but that’s just what it looks like without the information you’ve now provided. If the numbers you’ve quoted are correct, there is something wrong with your camera..

Member
February 22nd, 2016
Reply to  Barry Goyette

thats what I thought. had canon look at it and they say it matches their reference camera in the shop.. in which case I say both cameras have a problem :(

@ patrik Z ; as for sharpness being at +2, thats nothing on a scale that goes to 50, and coring at +5 takes any edges issues away. however I have NEVER seen that problem with either mk1 or 2. a small adjustment like that will not drive the noise through the roof either. sharpening is much maligned I find, but if you expect to actually get resolution from the camera factory defaults on most cameras are well placed. I can rarely say I’ve ever used it in a negative value because the image just gets too soft. there is some math behind it that every camera needs to so some sharpening to hit its specs.

I also followed canon’s recommendation of NR @ 2. realistically the sharpening + NR cancel each other out. Also FWIW I have seen some recent charts shot with NR at 0,2 and 12. While at 12 it does flatten out the finest detail, I’d hardly call it as dramatic as you are making it out to be – and thats on charts. The in camera NR is good and if it cLog requires moving it to 3 or 4, so be it if thats what it takes to get rid of the pattern noise.

probably what cLog wants is the selective NR engaged… and Arri apparently doing exactly this to keep the blacks / shadows looking clean.

Patrick Zadrobilek
Member
February 22nd, 2016
Reply to  Steve Oakley

Steve,

are you really talking about the C300 Mark II?
Because sharpness range is from -10 to +10, not 50.
I’ve done sharpness tests and as soon as you hit sharpness -7 you see increasing “edge hardening” and it gets worst on every step. Best artifact free sharpness is from -10 to -8.
I will write about this in my Review coming out today, so keep an eye at http://gadgetflux.net

Member
March 2nd, 2016

I just looked. sharpness range is indeed 0-50, NR off,1-12, coring 0-50. so the minor adjustment to sharpening here is small.

Patrick Zadrobilek
Member
March 2nd, 2016
Reply to  Steve Oakley

To clarify: in which menu point do you see your sharpen settings?
I’m talking about image sharpness in picture profiles at Menu->CP->OtherSettings->Sharpness->Level.

Member
March 2nd, 2016

Yup same place. Ran it all the way up to check

Member
February 19th, 2016

Who do you believe, Alan Roberts or Sebastian Wober?

Jim Martin
Jim Martin
Member
February 19th, 2016

That’s easy, Alan…..impeccable reputation….

Member
February 19th, 2016

I’ll believe the person without the website that gets paid by the click.

But really, can’t they both be right? Alan Roberts is certainly an esteemed reviewer with more experience in his left pinky than Sebastian, but Sebastian’s review was really faulty in it’s “angle” and possibly his lack of understanding at how Arri cooks the books (sensors that show more noise in the highlights generally don’t show less noise in the shadows unless there is some serious work being done — noise reduction and loss of detail, followed by the adding in of some “pretty” noise). His tests didn’t compare detail in the shadows, only noise, and so we’re left with an incomplete conclusion that unfortunately, Sebastian really wanted to insist was very complete. (that’s that angle part). Otherwise, his results are generally in line with his previous testing, and that’s all fine. For him, most of these cameras are over-rated, (except the holy grail Arri of course), and as long as he’s consistent, it’s all good. His site got more play and PR over this than it’s probably ever had, the only downside being Sebastian having to deal with knuckleheads like me :-) and possibly some cold shoulders over at Canon at the next few trade shows.

Member
February 19th, 2016
Reply to  Barry Goyette

I really wanted to like this camera, as an owner of two C300 PL’s, but I see too many compromises out of the gate for the price point. I was disappointed that ProRes wasn’t an internal record option, the form factor is a pain, lots of little bugs in operation, color space doesn’t match Mark I(looks more like Sony magenta), didn’t see signal to noise improvement, more dynamic range–but only in shadow detail. Changing lens mounts still seems to be expensive and difficult, high frame rates not competitive with Sony, crop is not ideal quality or operationally.

No question the resolution even in HD is superior with the Mark II, as well as the codec. More ND options is nice. AF options seem to be a good feature for some. 4K, of course. Just feels like Canon is playing catch to Sony at a premium price.

At this point, I’m more excited about the Varicam LT, except for the potential difficulty marketing anything with a Panasonic name plate on it.

Member
February 20th, 2016
Reply to  Jeff Regan

Hey Jeff, as an original C300 owner, I can really say that this camera is a marked improvement in so many ways. Up to 12,800 this camera is distinctly cleaner than the original camera, and beyond that it’s about the same. Canon’s codec is really superior as a shooting codec to pro-res, and the dynamic range extension goes both ways..the highlights on this camera look nothing like the oversaturated highlights of the C300. The AF features will truly change the way many of us work. I don’t know what bugs you’re talking about (canon not making magnify work in during doesn’t qualify as a bug, more like a difference of opinion). The color on this camera is superior in every way to the original. The magenta you’ve seen (I’ve seen in too) is more of a white balance and rec 709 LUT issue. Once you set up the camera properly it simply isn’t there. And I’m getting a completely different response to my Arri L7Cs and Area 48 LEDs with this camera, which is a great thing, because the C300 always struggled with them. Lastly, the Cfast combined with the new codecs and better LCD and viewfinder (in particular the EVF is amazingly good) have made this camera much more self contained. No need for a recorder/monitor in most situations, although I will probably add a SmHD 702 sometime soon.

The Varicam LT looks like a great camera, pretty much designed to compete in the same market as the C300 II. The lowlight footage I saw at NAB last year was quite impressive. The best move they made was putting an EF lens mount on it, that should help it on the marketing side.

James Manson
James Manson
Member
February 19th, 2016
Reply to  Barry Goyette

I still shake my head at the damage the reputation of the C300MKII suffered due to Sebastian Wober’s flawed test. I witnessed orders being cancelled because of an online article that has now been shown to be incomplete. Bad news is good news for websites. Not much love for Canon in these here parts and the article certainly had an ‘Angle’

James Manson
James Manson
Member
February 21st, 2016

I have looked at the Alexa in comparison to the C300MKII and the final image from the Canon C300MKII matches the Alexa incredibly well. It is a camera that can actually compete with the Alexa which is twice the price but all we hear is that the C300MKII is too expensive. Testing or no testing, the C300MKII is a monster of a camera that has been unfairly tarnished by your article that basically went viral. It was syndicated across practically every camera website there is. I’m not sure why you are unable to see the damage that was done to the reputation of the C300MKII by this.

Seriously, what is that you are actually risking?

Member
February 21st, 2016
Reply to  James Manson

I wouldn’t overestimate the effect of this article on C300 II reputation or sales. The reality is the price puts it out of range of the bulk of those who might have been influenced by the article (it was obviously written with FS7 fans in mind). There were and are certainly many other respected voices speaking out with a completely opposite opinion from Cinema5d. The reputation of the camera will come from talented people putting it to work on projects they care about, not from some overhyped step wedges.

James Manson
James Manson
Member
February 21st, 2016

I came to the conclusion that the C300 MKII is a monster camera by using it along with many other cameras filming many episodes for two television shows both lifestyle based in many, varied conditions, some brutally harsh. Auto focus that you completely discount makes the camera a monster for me alone. I am basing all my comments on my own experience using the camera since early November on an almost daily basis not because I work for Canon which I do not.

Member
February 20th, 2016

Sebastian,

As usual, you have a hard time seeing the forest for the trees. I was generally supporting your review, saying that it’s ok in Cinema 5d’s methodology for the camera to have tested the way it did.

However, your angle, and your placing of the Alexa on a pedestal doesn’t assist your review. Testing DR with an singular test like this DOESN’T tell the whole story. A step wedge can’t tell you how much detail is being retained or lost due to noise reduction. Arri’s sensor shows too much noise at 50% of peak white to be considered Tier 1 by EBU, yet it’s noise profile below 10% seams to go in the opposite direction of what it should. No other camera exhibits this. (EBU is considering rewriting its specification because of how the Alexa performs).

From Newsshooter– “At closer to peak white the Canon is actually quite a bit less noisy than the Amira (although film is actually noisy in the whites which is where the negative is at its densest). At the lower end of the curve, at below 10% peak white, things get noisier with the C300 MkII. The ARRI does something clever here which makes these lower levels less noisy.”

The ARRI produces a beautiful image, but it isn’t the engineering benchmark that you want it to be. For instance, it is only a 3k sensor and it’s measured resolution maxes out at 20% fewer lines than the C300. Generally higher resolution sensors exhibit higher noise, yet the C300 exhibits the same or less noise than the Arri everywhere except in the area below 10% where Arri is cooking the books.

Your article, it’s headline, your twitter posts, your methods for detailing your results (pushing the step wedge image in post so you can “see” whats going on in the shadows), …all of it was slanted. Inflammatory. This camera easily has 2-3 stops more DR than the original C300, yet your review indicates about a half stop. Your twitter posts suggested that it was actually the same sensor as the original.

I’ve been shooting with the C300 II now for months. I’ve never seen noise on this camera that was objectionable at the ISO’s we’re discussing. Yet somehow your article would make one believe that the noise on this camera was extraordinary. Actually, the noise levels and DR in Clog 2 are on par or better than what we see from the F65, F55, Red Epic, Varicam, et al. All cameras that are significantly more expensive. You article, in its weird way of comparing it to both an $70k camera and the $7k FS7, argued that it was simply too expensive, ignoring the fact that it performs similarly to many cameras at 2-3 times it’s price.

Sebastian…Buddy…sorry. I have to disagree that you simply don’t care how much DR a camera has. If you didn’t you would have simply posted the step wedges, and shut the hell up. Instead you editorialized the process in a number of ways. Which is the only reason you’re getting beat up over it.

Member
February 21st, 2016

—I think you have just read a lot of stuff on the internet and have little knowledge about proper camera tests. Bye—

….writes the man who published his “not-faulty, complete and unquestionable” test of the C300 mark II within hours of it becoming available.

….writes the reviewer who’s only “real world” mention the camera at the time of the review includes shock that a camera with a cineon type log curve would exhibit noise when underexposed (uh…maybe you should sit down with Sony Guru Alistair Chapman and have him explain that one to you).

…..writes the “creator of proper camera tests” who has tested so few cameras that he didn’t recognize CMOS smear until several commenters explained it to him.

Yes…Sebastian….I do read a lot of stuff on the internet. It’s a cool place to learn things. I also have worked professionally as a photographer/DP every day for the past 26 years. I’ve owned and used more cameras than you’ve had birthdays. I’ve never met a perfect camera and the C300 mark II is certainly no different in that respect. So “offending me and my camera” is really something you’d have a hard time doing….except when you assume you know anything about me or my experience…which by the way…seems to be a nasty tendency of yours when engaging people who disagree with you. (You should get that checked.)

Member
February 21st, 2016
Reply to  Barry Goyette

Please see image on this link for mk2 noise – clear horizontal and vertical pattern visible. graded image http://stevenoakley.com/images/c300mk2_noise.png

Member
February 22nd, 2016
Reply to  Steve Oakley

I’ll see your noise, and raise you.

4k C300II. Canon Log with Canon LUT applied in FCPX. No noise reduction. Graded and sharpened (which would have emphasized any noise if it were there.) Screenshot is not full resolution, but there is no visible noise in the 4k file.

https://www.flickr.com/photos/barrygoyette/24813260609/in/dateposted-public/

As Sebastian said, what you’re showing is not typical of this camera, unless you’ve underexposed the image or are shooting at 25,600 ISO or above.

Steve — As has been said many times here and other smart places, log gammas are NOT designed to be pushed in post…EVER. Log gamma’s (especially CLog2) are designed to be pulled. So the question “how far was it pushed?” isn’t really relevant. If you pushed at all, then it was underexposed. Are you using your waveforms when you shoot? CLog2 at base ISO has a clear clipping point at 92 IRE. Bring your exposure up so your highlights kiss that line and then back it off a smidge.

Member
February 22nd, 2016
Reply to  Barry Goyette

@ Barry, this wasn’t “pushed” by any means. why are you assuming that ? Shot you see shows blacks pulled down a bit from original raw shot. Waveform is peaks @ 80, most of the image at 40-20 with some of it down it 0. Another words a decent exposure.

settings where : Clog2, Rec 2020, 4K, Matrix neutral, sharpness +2, coring +5, NR 2,Color Matrix Gain 10+. G-R -2, G-B -2. As for ISO, best guess is 1000. Shot you see shows blacks pulled down a bit. Waveform is peaks @ 80, most of the image at 40-20 with some of it down it 0. Another words a decent exposure. The noise isn’t in the stop or so right above black, its 2-3 stops above black. Had into canon for them to look at and they told me to turn on the NR to 2 – which I had on this shot.

what it comes down to is this : cLog2 should ONLY be used in the most extreme of circumstances like interior shots against windows where you are trying to hold onto the exterior detail. other than that, cLog 1 or one of the cine gammas will be a better choice to keep the camera out of the noise floor.

Patrick Zadrobilek
Member
February 22nd, 2016
Reply to  Steve Oakley

With Canon Log 2 I would never ever set any other sharpness than -10, because it introduces all kinds of artifacts like double edges, raises noise dramatically and in camera noise reduction reduces resolution very fast. I will have my review ready soon were I test a lot of things on gadgetflux.at

CLog2 is beautifull when exposed correctly without any major noise in it when graded correctly, at least not much more noise than WideDR. The older clog btw, is much more noise free if you want to go for a perfectly “Alexa” clean image.

Member
February 19th, 2016

yAs an update since I had my Mk2 @ canon. They officially recommend setting noise reduction to 2 in cLog2. Looking at roberts test on noise reduction, you could easily go to 3 w/o any impact of significance on the image. If this fixes it, canon should consider updating the preset in the next firmware update. If Arri is basically doing the same – doing noise reduction in the lower parts of the image, so be it. FWIW the C300 Mk2 has selective noise reduction feature. you could enable that and set it up to only kick in below say, 20 IRE and solve the problem.

Member
February 19th, 2016

Fyi:

“Canon has announced that the EOS C300 Mark II digital cinema camera has been independently tested in accordance with European Broadcast Union (EBU) recommendations for both HD and UHD content acquisition…1

The report, by broadcast industry expert Alan Roberts, recognises image capture of up to 15 stops of dynamic range using Canon Log2 and qualifies the EOS C300 Mark II for R.118 Tier 1 in HD and Tier 2 in UHD2.”

Member
January 31st, 2016

I have been filming with the Canon C300 for the past few years. I have just picked up a brand new C300 MkII and I am getting a terrible high level of noise… at 800 ISO, the picture is just unusable. CPS Europe AGREED that my footage is getting far too much noise and actually their technician was pretty embarrassed… but in the end, he just blamed it on using Wide DR Gamma. They asked me to do more tests which I have done… and I am still getting the most awful noise. I did more tests shooting C-log, BT709 Preset, with CP setting… the footage is just bad… Let’s speak frankly, the footage is just terribly crap. So far it’s 18k in the bin and still no help from Canon… Not to mention that the picture is also soft… If anyone had similar issues?

James Manson
James Manson
Member
February 1st, 2016
Reply to  Olivier Sarbil

Soft? C300MKII? Lens, operator or defective camera. The C300MKII produces amazingly detailed sharp images.

Member
February 1st, 2016
Reply to  James Manson

Definitely not the operator, thanks ;)… the lenses are fine and work perfectly ont the Mark1. If you want to check out what I get form the Mark2 at the moment, please check out the link bellow.

the lens is a 70-200, ISO1000, Gamma Wide DR

I just upload a slow-mo testclip… check out the noise and the “sharpness”…

https://vimeo.com/153733071

PASSWORD: mark2test

CPS technician in France find it sharp, my camera dealer find the softness cinematic. ahahah

Member
February 2nd, 2016
Reply to  Olivier Sarbil

Hmmm that does look noisy for 1000. however, it also looks *really* under exposed too with not a lot of dynamic range. stretching that back out to a normal range could easily pull the noise out a lot. for such flat light I’d be shooting gamma 709 and 2020 or cine gamut color space. as example a 4K test here https://www.youtube.com/watch?v=_LefhSERi1I which is quite clean and sharp. I’d suggest a brighter exposure. the Mk2 likes more rather than less exposure

Member
February 2nd, 2016
Reply to  Steve Oakley

Hi Steve,

The shot needs a brighter exposure but it was shot to highlight some of the camera problems. At 1000 ISO, even under exposed and using the Wide DR gamma… I should not get such a high level of noise, that’s just a joke. Side by side with another Mark2, it takes 5mn to realise that camera is definitely wrong… You just can’t even get a sharp image! Yesterday Canon took back the camera and they finally admitted that the camera is defective (good for destruction). No matter the setting used, the quality of the picture was always terribly poor. Canon will exchange the camera and that’s the good news! But I can’t really hide my frustration… 18k camera and defective out of the box!

Dan Evans
Dan Evans
Guest
November 27th, 2015

For anyone interested, here’s some initial tests I did with Canon Log 2 and slow motion (2k).
https://vimeo.com/147038575

James Manson
James Manson
Member
February 1st, 2016
Reply to  Dan Evans

Thanks. Really nice.

 Emrah Cahit OZER
Emrah Cahit OZER
Member
November 26th, 2015

What do I get from the charts? Alexa 16 stops of DR?

Member
October 31st, 2015

Anyone know how to get C300m2 2k12bit 4444 footage into an editing program like Final cut x?

Member
October 31st, 2015

I don’t think 12 Bit 4444 is supported in FCPX yet.

Member
October 31st, 2015

thanks for the answer, I guess it’s Premier then?

Member
October 31st, 2015

I have no idea about Premier sorry.

Member
October 31st, 2015

I saw a video last week that said premiere is supporting all flavors of xfavc in the latest update. And canon is giving away a free year of premiere CC with purchase of the camera.

Patrick Zadrobilek
Member
November 1st, 2015

PremiereCC is supporting MXF 12bit 4:4:4 (there is no additional :4 which would be an alpha channel).

Joachim Hedén
Joachim Hedén
Member
October 22nd, 2015

Hello all.
I did a test that maybe should shed some light…
You can draw your own conclusions.

https://joachimhedenworkblog.wordpress.com/2015/10/22/some-dynamic-range-tests/

Kind regards,

Joachim

Member
October 22nd, 2015
Reply to  Joachim Hedén

Nice Test Joachim. Especially this part:

“You can draw your own conclusions.”

James Manson
James Manson
Member
January 30th, 2016
Reply to  Joachim Hedén

Thanks for a test done the way it should be.

Member
October 21st, 2015

So something no one has said : the C Log settings all have very high black levels. if you got out your old timey waveform monitor, or just looked at the camera’s you see the black levels floating at about 13 IRE. this floated black level means the codec will never crush blacks, but you can in camera. So if you REALLY want to BEST record what the camera is making, the stock LOG settings are basically rubbish.

I’ve been using either a custom setting using C Log as base, then adjusting the black levels down to 2-3 IRE. this expands the gradations recorded by the codec. this would also tend to reduce noise. While the Mk1 using 8bit codec, this is more critical, a 10bit codec is going to be more forgiving. Either way, the base C log settings are junk.

The other setting I use most often is simply a tweaked out Wide DR. Image reaquires the least amount of grading and looks good out of the camera. Its the same info as C log, it just looks better, take s a LOT less time to grade or mess around with. Even Wide DR can come out looking flat and Clog under the same shooting conditions would crush the image to basically 7bits.

The reality is shooting in sunlight, the DR of a 12 stop camera is NOT being pushed at all. in fact sunlight may actually be about a 8-10 stop DR in reality and the camera can capture easily everything from shadows to highlights.

Its far more shooting interiors with daylight that you run into more extremes of light that will start to push the DR of a camera…

the reality is, for what it would cost to shoot a dozen rolls of 35, you can own a camera that exceeds it. the pixel peeping here has gotten silly and it bears very little on what you can actually shoot in the real world. in fact the problem I find now is subtracting light from a scene with flags / solids / pieces of black rather than needing to always be adding tons of light.

On my last national spot shoot I used a silk to diffuse direct morning sun light, one 4X4 white board, and a 75W LED dimmed down as the tiniest amount of fill for the talents face. A far cry from what I would of used 10 years ago shooting on film. it would of been an entire circus of gear and crew.

So here is what I know: the Mk2 will be a better camera just for 4K over the Mk1, it’ll have more DR and if its 14 or 15 or 16 stops, it mostly doesn’t matter. my C300 paid for itself faster than any camera I’ve had before and its still making nice pix. Having the 4K option will be better. It will have fast ROI and be usable for 2-3 years which is more than you can say about a lot of cameras these days…

Jim Martin
Jim Martin
Member
October 21st, 2015
Reply to  Steve Oakley

Well said Steve!…..a real working pro with real working knowledge.

Member
October 21st, 2015
Reply to  Steve Oakley

Hey Steve, i feel the same concerning the diffrent profiles and black levels, but i didnt dig deep into the menu or touched the black levels – could go further into details? Do you have the time to create a video about it. The highly raised black levels seems to be overlooked by a lot of people.
Did you just compensate for the higher black with negative 13 or touched black gama aswell?

Member
October 22nd, 2015
Reply to  Florian Klaes

just adjust the master black level down while watching the waveform. I think it took around -5 or -7 to get the blacks down to about 2-3 IRE. Its a simple adjustment to make, and an important one on the Mk1 to get every last bit of gradation of out of the 8bit codec. I’d still do the same adjustment on the Mk2 as well. anything that improved the gradation recorded by the *codec* is a good thing.

FWIW at last NAB 2015 I was looking at Mk2 4K on 4K screen where they had dark blue / magenta wash on BG. perfectly smooth, no banding. 10bit helps a lot for shooting something like that, as does a tiny of of noise to break of the banding edges…even in 10bit.

The other thing no one is talking about is actual practical shooting. I’ve gone to ISO 4000 or 5000 for some shots. sure there was some noise, but nothing horrible. nothing anyone would turn their nose up at and say unacceptable. in fact quite the opposite, folks loved the low light shots very much. the noise is very fine, its not like the ugly globs you’d get on a old SD BVW400 or something similar… then of course folks slap film convert onto everything. go to the 3 perf 35mm setting and the grain added by FC (400-500ISO stock) is vastly more significant than what you get out of the Mk1 @ 4000-5000 ISO by a margin. So there is just some really silliness going on here…

Member
October 22nd, 2015
Reply to  Steve Oakley

Thanks! I find the noise feels more agressive and non-organic than on the C300-1 but is overlooked aswell it is a complete color free luma noise – i am waiting for a black friday sale on red giants denoiser. I should be easily to get rid of in higher ISO-sections.

Member
October 22nd, 2015
Reply to  Steve Oakley

Another thing i found in the manuel that isnt so common known is: Clog/2, base ISO: 800 – Wide DR, base ISO: 400.
Did you know that?

Member
October 22nd, 2015
Reply to  Florian Klaes

no. haven’t had a chance to look at the manual… speaking of which, where are people even getting Mk2’s ? no on seems to have them in stock right now…only pre-order

Member
October 21st, 2015

Does anyone in the US have this in stock?

Jim Martin
Jim Martin
Member
October 21st, 2015

We dealers will get more cameras at the beginning of next week….

Member
October 21st, 2015
Reply to  Jim Martin

Thanks Jim.

Patrick Zadrobilek
Member
October 20th, 2015

Hey Sebastian,

you should have tested clog instead of clog2. It’s far more noise free and also has a dynamic range of nearly 14 stops. Clog2 is a bit strange on the C300MkII on my tests, I agree on the noise at ISO 800, bit when going down to ISO 400 or ISO 250 the noise dissapears, but you loose about 1.5 stops of DR.

Patrick Zadrobilek
Member
October 21st, 2015

Hi again ;-)

That could be theoretically true, but it looks more like there is going on some gain boost via signal processing only on clog2, or else every Gamma settings would have been that noisy. So the conclusion is, to achive that “marketing 15 stops” we boost the sensitivity for clog2, but all not for all the other Gamma settings, which are clean. Check out the normal clog on the Mark II and you will achieve also about 14 stops of dynamic range – the true dynamic range of the C300 MkII, which is not bad at all and actually very competitive compared to other cameras in that price range ;-)

Member
October 21st, 2015

I think you’re wrong on this Sebastian. As you have stated previously in this thread, ALL manufacturers apply noise reduction as part of the processing of the image…so yes…all sensors are noisy…you have no way of knowing whether this sensor is noisier than others…you are only looking at results post processing. CLog2 appears to have less noise reduction than CLog, and BT.709 gamma has more noise reduction. All of these gammas have approximately the same number of stops, from numerous tests I’m calling it 14. Others have called it 15. You can call it what you want..but if you’ve tested CLog and found that it has the same noise characteristics as CLog2, I’d like to see that test.

Member
October 21st, 2015

Actually, Sebastian. It’s your claim that we’re talking about…two posts back…. “the noise is the same” (clog as compared to Clog2). As far as you saying you’ve got the white paper on your side. I’m not finding anything in there that claims the noise characteristics are better/the same between the various gammas. All discussions of noise appear to be relative to chip design and signal processing, not gamma. Please enlighten me.

For proof of my claim…which is not a claim…just an observation….I’ll offer this. A BT.709 gamma video that has been online for weeks now. Plenty of areas there where we should be seeing that awful noise you go on and on about. (underneath the tires and shadows within the interior of the car). Except…it’s just not there. There’s a download of the original file if you care. https://vimeo.com/140570879

The noise you point out in the shadows of the C300 CLog2 exists to varying degrees in every camera that produces a Cineon type gamma curve, including your beloved FS7 and frankly even the vaunted Alexa, but certainly in the Dragon, the F65, Varicam etc. Manufacturers have a choice of leaving that noise alone (and preserving detail) or smearing it knowing that most of that detail will get crushed anyway.

Jim Martin
Jim Martin
Member
October 21st, 2015
Reply to  Barry Goyette

Barry, you are right on target with your posts…along with MR. Boyle (ASC) and others here in LA that have a lot more experience in testing, evaluating, and using cameras on a daily basis.

Kaster Troy
Kaster Troy
Member
October 21st, 2015

I can’t believe we’re still talking about this, it’s absolutely ridiculous. If anyone has a problem with Sebastian’s tests than go do some of your own. If you can’t do a test of your own which nobody here will, than please thank Sebastian for his time not only for writing this article but also for him wasting his time replying to these idiotic comments. I’ve never seen a more annoying nit-picky comment thread in all of my life.

Member
October 21st, 2015

Who said CLog gives more DR? Not me. (I don’t think the white paper says anything about it…but…wait a minute…aren’t you the guy who doesn’t trust anything Canon says about this camera….so how come it’s bible now?.). FWIW, Canon has shown a chart (at the EVS demo a few weeks back) showing that Clog and Clog2 have identical DR characteristics, with a different distribution of it’s 15 stops.

” if you believe a different log curve will magically change the noise behaviour…”

Magic?…no…noise reduction. All cameras have it. You’ve said it yourself. All I’ve said is that Clog and BT.709 gammas on this camera don’t exhibit the noise you’ve discussed in CLog2. I’ve even shown you a sample. How am I twisting anything? You’re the one who claims it’s the same. Got any evidence of that, buster?

Finally, You don’t understand my aim? Well Sebastian, I really don’t have an agenda until I see someone with podium spouting shite. Believe me, every word you’ve spoken has made me research this camera and others at increasingly deeper levels. Spending $16k on a camera is not something I do lightly. And as I wait in the cue for mine, I’ve appreciated what you’ve said, but from the beginning, really couldn’t figure out your angle on this camera. I still really don’t know what you’re on a bout. All I know is this blog, by producing the thinnest of analyses (a single step wedge test) has generated the most noise.

Do I “want the C300 II to be better than your “findings”? What a stupid question. If I’m planning on buying one, I want it to be the best camera for the money. Funny, those are almost the same words Geoff Boyle used to describe this camera. You know him…they guy who’s website doesn’t take advertising. The guy who shoots feature films. The guy who’s been testing cameras since you were in diapers.

Frankly, though, I’ve really had it with your constant name calling. Whenever someone disagrees with you, you start saying they are spreading misinformation. It’s pathetic and it really shows your immaturity in this business.

Bradley Stonesifer
Bradley Stonesifer
Member
October 14th, 2015

I’ve just done some side by side testing with the Arri amira and C300mkii. I’ll be sure to post the results in the coming days. In terms of image quality and dynamic range, one would be hard pressed to see the difference initially. My guestimate on my end is the Amira holds about 1/2 to 1stop more dynamic range in the highlights. The scene that i was metering was 7 1/2 stops over Key. Color sampling in 4k is also better in the amira due to the 4444 12bit vs the 422 10bit. One interesting note is that the CanonLog2 gamma curve has a few options in color space and even though there is more info in the cine gamut color space, the rec709 color space is the most similar to the Alexa/Amira Color science. I’d use the c300mkii as a gimbal drone camera any day. I’m currently testing the canon raw to the odyssey 7q+ and discovered that it is in fact capable of duel link out. You have to set the mon out setting to 4k raw priority to achieve this. Only time will tell if Canon and Convergent will be able to get us 4k60p, fingers crossed.

http://www.bradleystonesifer.com

Bradley Stonesifer
Bradley Stonesifer
Guest
October 14th, 2015

I’ve just done some side by side testing with the Arri amira and C300mkii. I’ll be sure to post the results in the coming days. In terms of image quality and dynamic range, one would be hard pressed to see the difference initially. My guestimate on my end is the Amira holds about 1/2 to 1stop more dynamic range in the highlights. The scene that i was metering was 7 1/2 stops over Key. Color sampling in 4k is also better in the amira due to the 4444 12bit vs the 422 10bit. One interesting note is that the CanonLog2 gamma curve has a few options in color space and even though there is more info in the cine gamut color space, the rec709 color space is the most similar to the Alexa/Amira Color science. I’d use the c300mkii as a gimbal drone camera any day. I’m currently testing the canon raw to the odyssey 7q+ and discovered that it is in fact capable of duel link out. You have to set the mon out setting to 4k raw priority to achieve this. Only time will tell if Canon and Convergent will be able to get us 4k60p, fingers crossed.

Anonymous
Anonymous
Guest
October 6th, 2015

So, if it’s a stop more sensitive, wouldn’t the DR be better if you measured using ISO 1600?

sanveer mehlwal
sanveer mehlwal
Member
October 5th, 2015

I can see 30 Stops of Dynamic Range. Not sure what the others are on.

Hahahahahahahahaha

Crimson Son
Crimson Son
Member
October 5th, 2015

FYI – when determining dynamic range it is BOTH subjective and objective.

Objective is you use scopes like a waveform to show if each stair step is discernible electronically.

Subjective, compare image qualities of the extreme ends (usually shadows as digital still tend to hard clip highlights) and noise patterns.

Not sure why there is waveform in this testing.

Member
October 2nd, 2015

I’m seeing 16-17 stops of DR. no idea where the claim of 12ish comes from… at all.

Member
September 28th, 2015

Seems Prokit.com have the Canon in stock, if you’re in the UK. When are Bhphoto getting them? It says late October on the website. I hope not!

Member
September 28th, 2015

This article is getting a slating on other forums! Silly isn’t it. Will this article tarnish Canon’s sales figures? In the short run, maybe. Canon aren’t in any rush to discredit these findings. Maybe it is all just subjective, and this article amounts to an equally subjective discussion as to whether the Beatles are better than Elvis. Beatles rule by the way!

Member
September 28th, 2015

No man. Elvis is the king.

Member
October 5th, 2015
Reply to  Barry Goyette

Lensprotogo are claiming the C300 Mark II is on par with the FS7 in terms of Dynamic range! Here: https://m.youtube.com/watch?v=C1uo7lQ87-A

Did Canon promote and push this camera on a spec that is subjective? It’s a little sly if they did. No comebacks, no proof, no response to this article.
Surely the price must be called into question, if the FS7 can compete in DR as well. The other features the C300 offers are great, but without the DR and slomo advantage, are they twice the price great? It’s all gone very quiet regarding this camera!

Jim Martin
Jim Martin
Member
October 5th, 2015

No, it has not gone quiet….a large amount of orders and people calling wanting to know when they will get their camera. Again,long time trusted sources (ASC level) who know cameras are solid on the 15 stops with no subjectivity…..

Member
October 5th, 2015
Reply to  Jim Martin

BGD –FWIW…canon has never responded to a blog post in the years I’ve been following cameras. I will say that a canon person told a room full of people “off the record” ( with a wink) that they were comfortable with the 15 stop characterization. The reality is this camera is going to take great images. From what I saw in the LP2Go review, (which, by the way, in the overall, he came down on the side of Canon, not the Fs7), the color on the C300 is superior, the low light is superior, the focus system is “groundbreaking” and superior, the usability is superior. The only things he gave to the Sony were price and frame-rate.

I got my first look at some mxf files off the C300 this weekend (lensRentals blog). I think they look stellar. The color is phenomenal, and the image is amazingly clean everywhere but the deepest shadows, (all cineon gamma camera show this, including the fs7).

Received another email from Geoff Boyle over the weekend in which he described the C300 II as “remarkably good”. Tall words from a guy who’s tested every camera that matters for more years than some bloggers have been alive.

Member
October 5th, 2015
Reply to  Barry Goyette

Cheers Barry. I look forward to seeing Geoff Boyle’s tests.

Member
October 5th, 2015

Not fair to say Mark II is 2x the cost of FS7. If time code and genlock are useful, a $2K adapter needs to be added to FS7, plus V-mount batts, V-mount charger. The Sony proprietary media costs more, and if doing a lot of 4K, many cards are needed. The author of this article stated the Canon codec was superior. Most likely color science is as well. The package price difference is only a few thousand more for the Canon, but the FS7 is a great value, no question.

The Radiant test showed the Mark II to have the best shadow detail and highlight detail was competitive with all but Arri models. The Sony E-mount is not a rigid mount. The autofocus modes are class leading on theMark II

Kaster Troy
Kaster Troy
Member
October 5th, 2015
Reply to  Jeff Regan

It definitely isn’t fair but the fact is the Mark II costs double the price of an FS7. Sony isn’t the only brand that makes memory cards for the FS7, Lexar makes them for half the price. If a person needs timecode then your out an extra $2k but not everyone needs that. The side-by-side tests will be fun to watch once the camera is released.

Member
October 5th, 2015
Reply to  Kaster Troy

Kastor, you’re right, non-Sony media is cheaper, but hopefully CFast cards will go down in price faster. If time code in or out is needed, it’s not just the $2000 for the adapter, but also a need for a 2nd type of battery system, which for V-lock batteries and chargers, adds up quickly. The white balance issue in CineEI mode is a problem, no LUTs can be app,I’d in playback, the build quality is not fantastic.

To me, the Mark II and the F5 are the real competitors, but obviously the FS7 has some very features vs. pricing. Just can’t get excited by that wobbly E-mount.

Member
October 16th, 2015
Reply to  Jeff Regan

E-mount has a great advantage. It can use optional speedbooster that gains one stop light and gives it full-frame look.

Member
September 27th, 2015

I’d just like to say thanks to the author for taking the time to carry out the tests. I think she should change the DR test to something we can all relate to. I look forward to her future tests on up and coming releases.

Member
September 27th, 2015

Sebastian — There’s a saying here in the US that “if you can’t take the heat, get out of the kitchen”. There’s not a respected scientific study published anywhere in the world that doesn’t undergo “peer review”. While you consider the details of your article irrelevant (just like you previously described Dual Pixel Autofocus as irrelevant to the value of the C300 II because you don’t use it — funny…in the room full of hollywood DP’s, operators and producers who sat through a seminar yesterday on the c300, the biggest “ah-ha” moment of the day was the demo of this feature, followed by several outbursts of “kiss my 1st AC goodbye”).

The fact is the devil IS in the details, and your test really doesn’t pass muster. First it starts with some of the skewed comments you made on twitter, here and other places that indicates that you aren’t really looking at the data to learn, but to prove what you already know. Your inflammatory article title is just click bait, because we soon learn that you’re not measuring Dynamic range, but rather “usable” range (also heard at yesterday’s seminar “what the hell does that mean?…why does C5d get to decide what usable is?”… Your twitter outburst suggesting that this is somehow the same sensor as the original C300 is so thoroughly off base (take a look at the Radiant test and see if you can see ANY similarity between the C300 II and the C500.)

A better title for your article. C300 Mark II Dynamic Range: Exactly as I suspected.

“There is no need to discuss or criticize our testing methods” —seriously…you just wrote that? Who are you…the f**king Wizard of Oz?

Take a close look at radiants test. Designed by someone who knows how to design a camera test (and like all camera tests, it has it’s faults). You’ll see that the Amira bests the the C300II by about a half stop in the highlights, with the Canon showing about a half stop more range in the shadows, (exactly BTW what you’d expect given the companies stated DR distributions) albeit with a fair amount of noise. Looking closer we see that the C300 shadows are shifted upwards a little more than the Arri, but more importantly..looking into the deepest shadows we see that Arri is simply applying a fair amount of noise reduction down in those shadows, with the detail either smeared or simply gone.

Why do I bring this other test up. Because your test has too many problems to be considered valid. Usable DR is so subjective it’s not something that can really be tested for. But you insist it’s scientific. OK…then why did you rate the Amira at 13.1 stops last year and 14 this year? Why did you rate the original c300 at 10-11 stops (indicating in the text that it was at the lower end of the range), and then later rate it at 11.4 (just barely below canon’s stated 11.7). Why in the past did you describe the 14.5 stop A7s (Usable 12 stops) as nearly as good as the Amira, and describe the C300 as disappointing in comparison at 12.3. Why have we never seen the charts for any of these tests….(except the A7s as compared to the GH4)?

I think I know why we haven’t seen those charts before…because they aren’t that important…to you it’s what imatest spits out…based on the S/N ratio you consider ‘usable. The thing is…that on this test…the way the Arri’s blacks are rendered, makes it appear that there is substantially less noise and more DR. So lets haul out the charts and show ’em to everybody. So wait…is it the imatest number? or the visual that matters. See everybody I talk to says that noise in a DR chart only matters when you can no longer see the steps. When we normalize the blacks in the arri chart and line it up with the C300…we get nearly the EXACT SAME STEPS (the arri has a marginally visible step at 16), and the noise is barely different, (which we know from the radiant test is being reduced in a greater amount in Arri’s processing–even your test shows that the Arri gets ‘magically quieter (and blurrier) in stops 12-15). So if this test is based on measuring a certain signal to noise of a flat density patch and one manufacturer is applying a greater level of noise reduction to that patch, you can’t see how that that little detail is…um….relevant?

The take away from my discussions with a number of experienced camera testers is that noise is really a separate issue here, especially in the deepest shadows where it can easily be graded out or reduced as Arri (and um….everybody else) does with little effect on important detail.

Back to the Radiant test. It is absolutely true that the Alexa family of cameras bested everyone in the under/over/DR test. It should…it’s the most expensive of the bunch. I didn’t see another camera in the group that clearly bested the C300 II. The Varicam came closest. While the C300 has more noise in the deep shadows than many of the other cameras…it becomes clear that it’s maintaining more detail in those same shadows (the only camera to maintain the fine ribbing in the back wall around stops 2.5-3.5) This indicates that the competitors are simply masking out noise (and there’s still plenty there) where canon isn’t…yet.

For myself, I pulled together a composite of all the “portraits” in the over/under test, so I could see what 10 stops of “exposure” range looks like. I found that the C300 was easily gradable from about 4 stops over to 2 stops under. That’s a really good range, with plenty of headroom at either end to roll off highlights and shadows…all of the noise and “smear” dropped away in each of those grades with little coaxing.

The takeaway of this is your test is relying too much on noise at tone levels that don’t really matter, and that every sensor on the market either displays, or masks.

The only achilles heel that I see is that sensor smear…again…it’s something that generally gets graded out (along with the noise) so as you said earlier (before backtracking) it’s really not that big a deal….unless you think underexposing by 3 or more stops is a great technique. (interesting that you mention that the c300 had this problem. I’ve been shooting with that camera since the day it launched, and I’ve never ever seen it before)…if it’s the same then, probably I won’t see much of a problem with the new camera either.

A few days ago, you said you wanted to hear what I think of your test. Well now you have. Lets keep it civil, my friend :-)

John Trusk
John Trusk
Guest
September 27th, 2015
Reply to  Barry Goyette

No need to be rude Barry…. we can disagree without being uncouth. They guys spent time doing these tests…. lets at least show them some appreciation for that.

Member
September 27th, 2015
Reply to  John Trusk

Actually, John, I’m being pretty polite. Sebastian is pretty much calling everyone out who’s questioned his tests. That’s not polite. He’s chastised me more than once on this forum and on other sites for simply asking questions or engaging in polite, spirited conversation. The only thing impolite I said was a reference to a child’s book in which man controls a city of drones with a mantra of “don’t peak behind the curtain”. Read his previous post and see if it’s not apropo. Otherwise, the rest of my comment is just riffing on a lot of what has gone on here (and in a few other places) the past few days, AND to point out a much more valuable test on another site.

I have been incredibly polite and have received a lot of crap from him for it. I tend to have a really good sense of humor about this stuff, but Sebastian’s attitude towards a certain segment of his commenters (the ones that disagree with his conclusions) gets a little tiring. If you go back and read all the comments on this and the original “is the c300 still competitive” thread, I think you’ll see what I mean. If you go back and read, you’ll see that I have already thanked him for conducting the test, and noted that it was valuable.

Kaster Troy
Kaster Troy
Member
September 28th, 2015
Reply to  Barry Goyette

Barry, do us all a favor and just zip it. You obviously have no respect for the time and effort Sebastian has put into doing these tests, all you do is complain and say how inaccurate they are. Go do some tests of your own if your not happy with the ones provided here. In your mind this is already the best camera ever invented and you’ve never even used it. Its like you take it personally when someone has something negative to say about this camera. For the price this camera is a joke!! Accept it and move on.

Member
September 28th, 2015
Reply to  Kaster Troy

Oh…Kaster…I was starting to like you.

For a guy who started the conversation with “Canon sucks bro, accept it and get over it, case closed lol”….. you’ve sure grown. Actually in my world, I look at a lot of sources positive and negative and use the lot of it to make my decision. It’s worked out for me. Apparently, despite all the shite that was spoken by geniuses about the original C300, it somehow managed to dominate a big section of the industry. I’ll wage you a gentleman’s bet the new one will do just as well.

I’m happy to head on down the road, as this place is an unpleasant place. The forums I belong to, people debate and discuss and then make plans to meet up for a beer. Around here, if you don’t believe in the Wizard, they toss you out on your can.

And seeing as you’re feeling so supportive of C5D right now…go take a long hard look at that FS7 again. Because there’s plenty of bad news in Sebastian’s test if you care to “take a look”.

And as far as I know, the best camera invented is the Alexa….can’t afford one. I’ll have to settle for the next best thing. (oh…and I have used it before).

Cheerio mate.

Kaster Troy
Kaster Troy
Member
September 28th, 2015
Reply to  Barry Goyette

Barry, you and I soulmates you know this! Apologize for coming across harsh, a lot of it is my dry sense of humor. I just hate seeing all this bickering over a DR test. I read the article and looked at the pictures and then moved on. I don’t know why people have to continually question Sebastian. He took a lot of time to do this and the test is what it is, simple as that. I think everyone should thank him and wait for the footage or test to come out. I don’t wanna sound like I’m a Canon hater, my first real cam was a 5DM2 then the 5DM3 which are awesome. I never had any interest in Sony until somehow I got my hands on an FS100, FS700 and a7s and they changed my life lol. Sony is killing it right now and the amount of camera you get for the money is just incredible. But there are definitely things on the FS5 and FS7 I absolutely can’t stand.

Member
September 28th, 2015
Reply to  Kaster Troy

:-D

You know, I’ve published camera tests (it was a long time ago in much simpler time)…so I know where Sebastians at with this thing, and I know what it’s like to have everyone call you on every detail. (like that other test over at radiant, which was a huge undertaking…they screwed up a couple of cameras (of course the FS7 is one), and so that’s all people are talking about in the comments..you can’t win.

I really have less of a quibble with the C5d test than it’s conclusions (a hallmark of most ASC supervised tests is that the tester rarely expresses his opinion, and lets the image speak for itself) and Sebastian’s general reaction to commenters that disagree.

Nice singin’ on “Goodnight”.

Kaster Troy
Kaster Troy
Member
September 28th, 2015
Reply to  Barry Goyette

Thank you sir, that’s very kind of you to say ;)

Member
October 12th, 2015
Reply to  Kaster Troy

Hey Kaster…thought you might be interested in a test that Geoff Boyle has been cooking up. Comparing the range of lower priced cameras from Sony to Bolex to Canon and many stops in between. He had generally “fair to midland” things to say about all of the cameras except for a few. Here’s the part that’s relative to this forum….

He starts with this…..”I don’t like the budget cameras, it’s not that they can’t produce perfectly usable pictures, they can. They just don’t produce stellar pictures and they are a pain in post.”

and then this….

“Sony FS7 it’s a Sony so it has a lot of menus :-) the pictures are fine, if you like Sony cameras you’ll love this. It may help to make me like Sony more!

F55, not really part of tis series but we missed it last time, a good camera, simple if you want it to be and great pictures, a truly professional camera. Worth the money.

C300-2 I’ve kept this to last because it is a clear winner, miles ahead of anything else in the tests, easy to use, lock it in cine mode, easy to read the rushes, easy to grade, just drop the appropriate supplied LUT on and marvel :-) This is what a camera should be.
They should have called it the C400 justas the C500-2 should be the C600. It’s that big a change. In such a short test we didn’t have time to play with the looks function but if they work as we’ll as everything else…

It’s very simple really, if you want great images and a pain free life use Alexa or Amira, F65 or F55, Dragon, Varicam, C300-2.

If you fancy a challenge then maybe include the KineMax.

Anything else is a Fiat 500…”

Just one man’s opinion, I know…

sanveer mehlwal
sanveer mehlwal
Member
September 27th, 2015
Reply to  Barry Goyette

Why don’t YOU conduct a test, which is more Scientific and Precise and let us all see the results.

Member
September 27th, 2015

I would…but then I’d have to fight with myself. :-)

GELAX STUDIO
GELAX STUDIO
Member
September 28th, 2015
Reply to  Barry Goyette

websites like dxomark and dpreviw also does camera tests,what if c5d does their tests base on the same standards of dxo or dpreview ?

sanveer mehlwal
sanveer mehlwal
Member
September 28th, 2015
Reply to  Barry Goyette

You would have to eat gumble pie, obviously.

Also, it’s easier for critics to call a film rubbish, rather than ever really make one themselves.

“The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts.”

I guess Bertrand Russel was thinking of you, when he said this.

Rastko Vukovic
Guest
September 28th, 2015
Reply to  Barry Goyette

Cinema5D.s test turned out to be spot on. https://vimeo.com/140441330 Now I knew that some people will be, like, hey you didn’t turn the log/gamma/hyper/camera up side down but the truth is this is THE test and canon IS lying and IS overpriced :) deal with it. Unless you work for canon. In that case go F yourself XD
peace
ps https://vimeo.com/140441330
pps https://vimeo.com/140441330
use your eyes

Member
September 28th, 2015

Hi Sebastian,

Thanks for the response to my comment.

I think what I meant relative to noise/DR is that by setting a very conservative number like you have, that results can often be skewed by increased noise reduction, and so while your test ends up being a better test of “image quality” than a less conservative noise number, it may be less accurate in determining actual dynamic range. As we see in the Radiant test, all of these cameras show a lot of noise in Log when pushed to their limits. Each one deals with that noise differently. Your test is totally valid, but does benefit cameras that reduce shadow noise (and detail) to a greater degree.

At the Canon seminar this weekend, they showed us a lot of clips that displayed how the C300 II was holding detail in the deep shadows on very high contrast scenes. This is something that your test doesn’t consider, and it’s an import an aspect of dynamic range that DP’s do consider.
There were a lot of experienced (old, like me) operators and DP’s in that room. Most seemed genuinely impressed, I also think that most would have considered that detail “usable”.

Again…thank you for your answer on this. I’m more than happy to agree to disagree.

Member
September 27th, 2015

Doesn’t look like LOG for the FS7 and A7S on the Radiant test?

Member
September 27th, 2015

Absolutely correct. That test, at least on the surface, appears flawed in that respect.

Member
September 27th, 2015

https://vimeo.com/140441330

What to you all think about this test? Does it leave us with any news concerning the dynamic range and soft slow motion picture?;)

C

Member
September 27th, 2015
Reply to  Conny Fridh

Looks promising, i think all this is subjevtive and each online evaluation is diffrent it would really help to see the C300-1 against the C300-2 with original footage for download of both and a statement by Canon about this.

Adrian Bacon
Adrian Bacon
Member
September 28th, 2015
Reply to  Conny Fridh

The Arri’s performed about what I’d expect to see on the Radiant test. Fairly good all around. The Red’s were about what I’d expect to see, also fairly good all around.

The Sony A7R2 had a ridiculous amount of clean shadow detail. Totally unreal. The downside though was by 1.5-2 stops over it was already completely blowing out highlights. This is in line with what I generally see with their stills cameras. They’re strongly biased to shadow detail at the expense of really ugly highlights. It’s just how Sony distributes their DR. Nothing wrong with that, though, something to be aware of and expose for accordingly.

The C300MkII was not as clean in the shadows as I was hoping, but with that being said, it looked to be mostly luminance noise and was not visually super ugly. In post it would clean up quite nicely and retain a fair amount of detail. I was actually hoping for more highlight range. It looked to be losing the highlights by 2.5-3 stops over. I really hoped it would have been more like 3-4 stops over. Nothing terribly wrong with that, just something to be aware of and expose for accordingly.

In all honesty, I thought the Sony FS7 didn’t look that great. The C300MkII looked way way way better at 5 stops under and had way more/better looking shadow detail there. Also, it started losing highlights by 2.5 stops over again. Again, nothing terribly wrong here, just something to be aware of and expose for accordingly.

The URSA was meh. The Panasonic was better than I expected. They’d do just fine if lit properly on a TV set, but that’s really about it.

My take away, all the cameras had their strengths and weaknesses, a few of them I’d probably only use in certain scenarios that played to the strengths, otherwise, they were all fairly competent if used by a skilled user that knew how to get the most from them.

Member
September 26th, 2015

These two videos show that the C309 Mark 2 has very good DR:

https://vimeo.com/140441330

https://vimeo.com/137221219

Joseph Wilkins
Joseph Wilkins
Member
September 26th, 2015

Still no response from canon???

Member
September 26th, 2015

I’ll take 3 Ursa Minis and a shot of Jack for the same price, Alex.

Joachim Hedén
Joachim Hedén
Member
September 26th, 2015

Hi there,
Before your test, did you verify the chart with a spot meter? Did the light source perhaps dim, is there dust/dirt that could affect the levels?
I havn’t done any controlled environment tests, but my feeling after shooting with a production model 300 mk2 for a week, is that I’m holding highlights easier than with a Sony F5 wich I’ve shot with a lot in the past few years.
Also, about the term “useful stops” – it’s a very subjective term. Here’s how I think about “useful stops”: The useful stops are what’s available above middle gray, not below. What’s down there in the bottom 2 or 3 stops you’re not really “using” for anything – but you are always protecting your highlights from clipping. The Sony F5 for example has nominally 6 stops of lattitude above middle gray, but in general the F5 needs to be overexposed by 1 stop to look good. So, that leaves you with 5 “useful” stops above middle gray. Like I said – I haven’t done any controlled environment tests, but my feeling is that the 300mk2 is doing rather well in the real world “above-middle-gray-stop” -aplication of the term “useful stops”.
My 2 cents

Member
September 26th, 2015

Interesting that the smear is exactly the width of the first blown out chip.

Peter Hybæl
Peter Hybæl
Guest
September 25th, 2015

Kudos to all folks on the forums who have seen this coming based on the highlights in the example footage that was available for Mk II. They have brilliantly called it.. no 15 usable stops anywhere.. not even in the Alexa territory. The only excuse could be that this is a prototype C300 Mk II camera an not the final one.

Also haha it really shows that Arri does have some special sauce or some super clever tricks or hacks (I don’t believe in the strange physical low-con filter theory).. my best bet is that Ursa 4.6k will be 1 stop shorter than Alexa.

Einar Davíðsson
Einar Davíðsson
Member
September 25th, 2015
Reply to  Peter Hybæl

Cinema5D have stated here and elsewhere that this was a full production camera, not a prototype.

Kendal Miller
Guest
September 25th, 2015

Not good.

Clayton Burkhart
Member
September 25th, 2015

Sebastian, please see if you can do this same test with the new RED RAVEN when it comes out. I am entirely fed up with their claims that their sensor does 16.5 stops, when in fact I have noticed that it has never surpassed the DR of an Alexa, which as we know is a touch more than 14 stops.

Member
September 27th, 2015

Heres is an other interesting test. How do you feel this hold up to the idea of your evaluation of the C300 II?

Still no comment from Canon about the dynamic range and soft slow motion picture?

Kind regards
C
https://vimeo.com/140441330

Tobias Mennle
Tobias Mennle
Guest
September 25th, 2015

Sebastian, about the noise. Any log gamma needs a proper delogging with an S-curve before one can reasonably judge noise. You are dealing with 3 different log curves here, comparing Canon to Sony to Arri. You need to delog and colour correct. On my first test with the FS7 I had noise that reminded me of high speed S-16mm, but it completely vanished when I delogged. Most people just don´t seem to get it, and just start to colour correct log material to taste. You see it in Sonys first FS7 clip, and in Philipp Blooms C300 demo, too. So IMO delog both the C300 and the Arri and Sony footage and then judge the noise.

You write: “12 bit will not give you a higher dynamic range, only better gradations.” Again, you need to judge that in colour correction when manipulating the image. I bet a 12bit log will give you way more dynamic range than 10bit log. You really need to try it and pull up the shadows of both and then compare.

About slomo, Canon is so fair to describe the camera in their white papers, and there it is: At 100 frames per second the bottleneck in the processing is 1024×540 pixel. WOW. I know my FS7 150 frames and they look mediocre at best, but this sounds even worse. 540 Pixel…

I bought an early C300 MK1 against better knowledge, but this one seems even more underspecced.

Nevertheless I am sceptical about your test procedures and seriously doubt they do the camera justice.

Tobias Mennle
Tobias Mennle
Guest
September 25th, 2015
Reply to  Tobias Mennle

You write: “Looking at the test charts…You can see the usable dynamic range highlighted in red.” No, we can´t. We are looking at the outcome of three different log curves that need their specific delogging so we can judge the images reasonably. One simple point is, a proper delogging will make the shadows even darker, but also get rid of the noise. Part of the lights will get brighter. Etc.

More here: “Understanding log grading” by Mike Most is a brillant short explanation.
http://mikemost.com/?p=251

If you you for example check Shane Hurlbuts tests, of course he has always has perfectly graded images when he compares cameras, in whatever respect. We can totally forget about log images, we need the rec709 grades.

Please check your test procedure, you are not doing yourself a favour.

Member
September 25th, 2015
Reply to  Tobias Mennle

Tobias — just a correction on the Canon HFR. That 1024 x 540 “frame” that canon is referring to, is a color component frame. Meaning they are breaking down the 2k capture into 4 components (red, blue and two greens) with the pixel dimensions above. Resolution wise this is the same as any 2k sensor. Canon uses this novel process rather than a typical debayer. and it gives them a great 2K image when sampled downward from 4k. In this situation, the process is giving them largely the equivalent of a standard debayered 2k image, meaning something less than their stellar 4k downsampled 2k. It looks very much like what I see coming off the FS7 at 150 and 180 FPS, which is doing something internally to avoid processing the entire 4k image at that rate, probably binning. Two different approaches to largely the same problem.

Adrian Bacon
Adrian Bacon
Member
September 26th, 2015
Reply to  Tobias Mennle

Agree completely with this comment. The dynamic range is extremely subjective and this test does not take into account the post process. I’d really like to see what that chart looks like after it’s been properly graded.

Noise removal, contrast, saturation, tones/curves, sharpness, etc. all the post stuff we do to get it into a deliverable h.264 file.

Personally, I find that I’d rather see a rating of discernible stops, regardless of noise level because I know I’m going to be doing some amount of noise removal in post, and discernible stops gives me a ball park of what I have to work with in that regard.

With that being said, I’m seeing 17-18 stops of discernible range for the c300 Mk II on your chart. After post noise removal and grade, I can usually get a good looking image that is within 1-2 stops of discernible DR with either no visible noise, or dramatically reduced noise.

With that being said, I do believe canon has a fairly solid 15 stops and then some depending on how much post noise removal you’re willing to do.

We’ll see once people start posting graded real world footage. Philip Bloom has already posted some test footage and frankly it looks like there is a lot more DR than the first C300 (which he also has/had).

Member
September 25th, 2015

Canon must respond to the facts about dynamic range and The doft slowmotion crop image.

And they have to do It tight now. IF not I Will chancell My order of This Camera.

By The way, what usable dynamic range number did you get from the older C300? Also 12,3?

Thanks
Conny Fridh

Member
September 25th, 2015

Great,, please do IT soon!

GELAX STUDIO
GELAX STUDIO
Member
September 25th, 2015

Does the C300M2 shoot 120P FHD in 10bit or 8bit ?

Adrian Bacon
Adrian Bacon
Guest
September 25th, 2015

I want to see a comparison between the C300 and the C300 Mark II using the same test. That will tell you how much the dynamic range has increased by with the Mark II.

Comparing it to the Alexa and the FS7 is useful, but frankly, subject to your interpretation of what you consider to be usable dynamic range between the cameras.

Doing a C300 to C300 MkII comparison and using the same interpretation of useful dynamic range will show where the improvements are, and since the original C300 has been around for a while, how it performs is fairly well studied and documented, that will give a good interpretation of the improvements to the Mk. II.

Without that, it is a subjective test that is open to the interpretation of the tester.

Member
September 25th, 2015
Reply to  Adrian Bacon

Agreed 100%

Member
September 25th, 2015
Reply to  Adrian Bacon

Yes I also would like to see them do this test. Right now actually. Because to come with a claim like this, you need to back it up all the way.

GELAX STUDIO
GELAX STUDIO
Member
September 25th, 2015
Reply to  Conny Fridh

That is why cinema5D built their TEST LAB!
C300 have tested data,and u can check it and compare it to the C300M2~