Guess the camera

Can you guess the cameras?

Recently I did some testing with my cameras:
* BMCC (Blackmagic Cinema Camera) in ProRes HQ Film 1080p 4:2:2 10bit (Tamron 17-50 2.8)
* Canon XF305 in 1080p 50MBit 4:2:2 8bit (built-in 18x lens 1.6-2.8) with a custom picture profile
* URSA Mini 4.6k in ProRes HQ Film 2160p 4:2:2 10bit, downscaled to 1080p in post (Tokina Cine 50-135 T3)

I threw the footage into Premiere Pro's timeline and used RGB curves to match the 3 cameras as close as possible. Can you guess the camera for each screenshot?

Klick the previews for a full size view:


Camera A = ?


Camera B = ?


Camera C = ?

Cam A = Canon
Cam B = BMCC
Cam C = 4.6k

Yes?

Cam A = Canon
Cam B = BMCC
Cam C = 4.6k

Yes?

My guess as well !

EDIT: mmh, now I think C is Canon because it's a little bit warmer, and A 4.6k.

Hard to tell, great idea for the little game ^^

A = BMCC

B = Canon

C = Ursa

Hey Robert,
Do we win a prize if we guess correctly?

A = Canon XF-305
B = URSA Mini 4.6K
C = BMCC 2.5K

OK Robert, so what is the correct answer?
Cheers

Timbutt got it right

A = Canon XF-305
B = URSA Mini 4.6K
C = BMCC 2.5K

Because of its 3 cmos design the Canon is almost immune to moire. Funny that the URSA Mini showed the strongest artifacts.

Timbutt got it right

A = Canon XF-305
B = URSA Mini 4.6K
C = BMCC 2.5K

Sweet! What do I win? Please tell me it's a trip to Hawaii... or the Bahamas.

If you ever come to Austria, I'll invite you for lunch.

At which resolution was each filmed?

Edit: Oops. Read the OP.

So then next question. Was the moire in the screen at the top of the frame present in the 4.6k image before downscaling?

If you ever come to Austria, I'll invite you for lunch.

Sweet! Free lunch! I accept. I've always wanted to visit Austria. Salzburg especially so that I could visit Mozart's home.

Tim, Salzburg is a 3 hours drive from Graz.

@ Jonesy
The moire was present in the UHD footage too.

I'm surprised that the Ursa performed so bad. That was the least favorite of the three. I enjoyed the first and last as they were more filmic, less videotesque.

I disagree Thomas, the UM (middle shot) has cleaner whites that the Canon, and less contrast than the BMCC, and looks more like film in its latitude and color saturation (which looks more natural to me over the other two), the blue is "cleaner" looking. Filmatic can mean many things, and has more to do with lens choice, lighting, framing, and exposure than it does sensor technology. We are getting to the point with many excellent sensors available, the difference is going to be down to "how" it is used, and the talent of the person shooting/directing, etc.
Cheers

I suppose that "filmic" more has to do with which types of movies you grew up with. I was conditioned by the 1970's and 80's look of film. Film emulsion today looks more "digital" than it used to, to the point that certain directors such as Darren Aronofsky nowadays prefer Super 16 as it more resembles Super 35 of the 1980's. Also, one may analyse a picture apart but it always boils down to (at least for me) the emotional response, the spontaneous reaction when watching a picture for the first time.

Perhaps "bad" was a strong word. To be honest, i had a somewhat hard time to distinguish between the tree, but the URSA was the picture that stood out the most and diverged from what I associate with film, although it still looks quite o.k. But the moire is terrible, you have to agree on that one.

I prefer the last of the pictures, the image of the classic Blackmagic Cinema Camera.

Very interesting test. I'd be curious to see it repeated on a closeup face. To my eye, one of the things that sets the Ursa Mini apart (when shooting raw or ProRes444) is the subtle skin tones.

The amount of moire on the UM46 is troubling to me. How about anyone else? Shouldn't it have less? Perhaps in rare cases the additional resolution is a handicap? I don't know. I wonder if shooting at 4.6K would have eliminated some of it?

Another observation, perhaps the OP could chime in, but it appears to me the Canon and BMCC have been lifted to match the UM46. Although there is still a hair more in the shadows of the UM46, they are nearly the same. However, I believe I may see some noise in the shadows on the other 2. If these we moving images, that noise may be going crazy and the UM may really show its strength in that category. Thoughts?

Tim, Salzburg is a 3 hours drive from Graz.

@ Jonesy
The moire was present in the UHD footage too.

Isn't Arnold Schwarzenegger from Graz?

Thomas, I agree with you on the use of S16 to get a more creative Cinematic look.of the 60s and 70s film look. I think the smaller resolution of S16, andmthe BM s16'sensor has a,"film grain" like look to its noise pattern in the shadows. Looking at a still grab vs a motion sequence is not the same in judging how it will look projected.
Cheers

I'm surprised that the Ursa performed so bad. That was the least favorite of the three. I enjoyed the first and last as they were more filmic, less videotesque.

Tomas, I think your eyes have been tricked by the more apparent noise grain, while the URSA looks much more clean. So your brain associates the noise grain with filmic.

I have to say that I prefer the clean and sharp images of the Ursa (except for the moire in this sample). But I was surprised myself how close I could get those 3 cameras in the looks.

I disagree Thomas, the UM (middle shot) has cleaner whites that the Canon, and less contrast than the BMCC, and looks more like film in its latitude and color saturation (which looks more natural to me over the other two), the blue is "cleaner" looking.

Denny, the white wall of the house is not supposed to look clean, as it has a rough finish with glossy elements. The amount of contrast was just up to me and my grading. The same goes for color saturation. And all 3 cameras got individual in-post sharpening to match more closely (my Canon's picture profile has toned down sharpness).

Very interesting test. I'd be curious to see it repeated on a closeup face. To my eye, one of the things that sets the Ursa Mini apart (when shooting raw or ProRes444) is the subtle skin tones.

Jamie, I was planing to do those shots next. Stay tuned

The amount of moire on the UM46 is troubling to me. How about anyone else? Shouldn't it have less? Perhaps in rare cases the additional resolution is a handicap? I don't know. I wonder if shooting at 4.6K would have eliminated some of it?

Jonesy, this is the first time I found such extreme moire on the UM46k. Normally there is none to very little. The balcony where the moire jumps out is made of a perforated metal plate with lots of tiny holes. It seems I just nailed the resolution limit of the sensor and the de-bayering algorithm. At other focal lengths there is no moire.

And I am not shooting 4.6k atm because of the cross hatching pattern I have there. At 2160p cross hatching is masked out by the downscaling.

Another observation, perhaps the OP could chime in, but it appears to me the Canon and BMCC have been lifted to match the UM46. Although there is still a hair more in the shadows of the UM46, they are nearly the same. However, I believe I may see some noise in the shadows on the other 2. If these we moving images, that noise may be going crazy and the UM may really show its strength in that category. Thoughts?

All cameras have been lowered in the shadows not lifted. But the Canon and the BMCC have just more noise than the UM46k - and the UM46k has the advantage of the better sensor and 50% downscaling.

Isn't Arnold Schwarzenegger from Graz?

Indeed, he is.

Here are clip samples of the graded footage in DNxHD 185 10bit MXF for you to download:
https://we.tl/jSBceBuQ1K

Link will be active for the next 7 days (til 15.12.2016).

When comparing keep in mind:
BMCC and UM46k have been set to 11.25° shutter, ISO800 and F8 for avoiding the usage of an ND filter.
Canon XF305 was set to 180° shutter, gain -6dB and ND1 (1/4 = -2 stops) was active.

Tomas, I think your eyes have been tricked by the more apparent noise grain, while the URSA looks much more clean. So your brain associates the noise grain with filmic.

When digital film looks like (and evokes the feeling of) emulsion film, our eyes are always tricked Robert

The grain, yes, but also the more softer look compared to the ultra-sharp Ursa. That's why some crazy people, like myself, prefer the S 16 sensor range of cameras, such as the Ikonoskop A-Cam dII, the Digital Bolex D16, and the Blackmagic Pocket and Micro Cinema Cameras, to convey a classic emulsion film look.

Tomas, how do you like this short test scan I made of an old 16mm film sample, can you guess which film type had been used?

16mm_scan.mp4

Oh, and here are the ungraded samples of the 3 cameras:

Canon XF305 (only superwhites have been pulled backed):


URSA Mini 4.6k (could not nail the WB setting of the other two cameras):


Blackmagic Cinema Camera:

Robert, excellent test! Congrats, buddy.

All cameras have been lowered in the shadows not lifted. But the Canon and the BMCC have just more noise than the UM46k - and the UM46k has the advantage of the better sensor and 50% downscaling.

You mean even though UM4.6K has a better sensor and great downscaling, the other 2 cameras were still able to get the same result?

Tomas, how do you like this short test scan I made of an old 16mm film sample, can you guess which film type had been used?

How do I like it? I like it very much. But I cannot tell which type. Looks a bit like reversal but I dunno for shure.

Tomas, how do you like this short test scan I made of an old 16mm film sample, can you guess which film type had been used?

How do I like it? I like it very much. But I cannot tell which type. Looks a bit like reversal but I dunno for shure.

See
In reality this was shot with a Canon XH-A1 1080i50 HDV - CCD sensor.
I have deinterlaced it, put a pseudo film frame around it, softend the outer edges, increased contrast and saturation, added 3 layers of artificial grain gradually to the brighter parts (like reversal film), changed fps to 24.

It is just our brains tricking us.

It is just our brains tricking us.

Actually, I knew that this was coming.

But, hey, what's the problem? Movies are about illusions, aren't they?

Timbutt got it right

A = Canon XF-305
B = URSA Mini 4.6K
C = BMCC 2.5K

Because of its 3 cmos design the Canon is almost immune to moire. Funny that the URSA Mini showed the strongest artifacts.

That's not "funny", that's just a matter of resolution... both other cams have too less resolution to differ the small details that cause the moire.

The Epic Weapon 8K Super 35 sensor must be presenting new challenges for RED as the photosites are smaller than they are used to. I imagine they may have OLPFs specific to the sensor.


Sent from my iPhone using Tapatalk

Olpfs have to be designed to match the sensors. That's always been the case.

The olpf does contribute significantly to the camera's look, since it alters the way that the sensor records fine detail. That is one way that camera makers differentiate themselves when using the same sensors.

...that's just a matter of resolution... both other cams have too less resolution to differ the small details that cause the moire.

Another case against to high resolution.

So, as promised I did some comparison shots with skin tones.

* BMCC (Blackmagic Cinema Camera) in ProRes HQ Film 1080p 4:2:2 10bit (Tamron 17-50 2.8)
* Canon XF305 in 1080p 50MBit 4:2:2 8bit (built-in 18x lens 1.6-2.8) with 2 custom picture profiles
* URSA Mini 4.6k in ProRes HQ Film 2160p 4:2:2 10bit, downscaled to 1080p in post (Tokina Cine 50-135 T3)

Lightsource is the 225W Dedolight DLED12.1D with 5550K as a back key light from the left. It is reflected into the white soft side of a Sunbounce reflector, which is used as the front fill light from the right.

Notes:
For the Canon XF305 I have used two different custom picture profiles.
CP1 is optimized for situations with enough light and -6dB, shadows are saturation muted and noise reduction is off. For this shot the camera was set to 360° shutter, F2.4, +6dB
CP3 came with the latest Canon firmware update and is a Canon Wide Gamma, which always uses +6dB and some low noise reduction. Camera was set to 360° shutter, F2.4, +6dB

The BMCC was set to 180° shutter, ISO800 and F2.8.

The UM46k was set to 180° shutter, ISO800 and F2.8.

The UM46k was the most sensitive of the three, the Canon was the least sensitive as you could guess from the 360° shutter.

Color correction was done without any LUTs, just saturation boosts, RGB curves and post sharpening.

Thank you for doing skin tone tests. It make two things quite clear, I think:

1. When viewing the image links full screen, the lower noise, finer detail and higher color quality/bit depth of the Ursa Mini is easy to see.

2. When viewing only the downscaled images that appeared directly in the thread, it is hard to tell them apart.

So, depending on your intended distribution, the differences may not matter. For example, last night I was at the house of a friend who is an Oscar voter and was shocked to see that the Academy still sends the voters DVD discs. The results are not pretty. The heavy compression turned the shadows into banded mush and all fine detail was buried under macroblocking. Couldn't tell whether a feature had been shot on 70mm film or an iPhone. It was painful to see some of the year's best films reduced to that.

Very nice job on the camera matching and the lighting. How do you like that Dedo LED?
I own the Rayzr7 300watt daylight LED and like it quite a lot. At 12Kg the Dedo is heavier and its CRI rating is slightly lower, but I'm sure it's built beautifully and likely is more even and controllable than the Rayzr7.

DVD? To play back on their 4K 240Hz widescreen TVs! Send in the clowns...


Sent from my iPad using Tapatalk

I don't know (dino), I still watch those large and bulky "Laser Disks", closest thing digital to film!
Cheers

Thank you for doing skin tone tests. It make two things quite clear, I think:

1. When viewing the image links full screen, the lower noise, finer detail and higher color quality/bit depth of the Ursa Mini is easy to see.

2. When viewing only the downscaled images that appeared directly in the thread, it is hard to tell them apart.

Fully agree. In HD and situations where the Canon has no problem with the DR it is very hard to see a big difference - but the Canon XF305 also has got a very good sensor and I carefully tuned the picture profiles.

In situations with huge contrast ratios, or with wide angle shots, or when you try to better separate foreground from background then the UM46k really shines. This is what this camera was made for.
And watching 2160p50 footage is really cool, although I wouldn't use 50 fps for cinematic results.

So, depending on your intended distribution, the differences may not matter. For example, last night I was at the house of a friend who is an Oscar voter and was shocked to see that the Academy still sends the voters DVD discs. The results are not pretty. The heavy compression turned the shadows into banded mush and all fine detail was buried under macroblocking. Couldn't tell whether a feature had been shot on 70mm film or an iPhone. It was painful to see some of the year's best films reduced to that.

I can feel your pain. After being poisoned by UM46k's 4k it is hard to go back.

Very nice job on the camera matching and the lighting. How do you like that Dedo LED?

Thanks, Jamie. The Dedo LED is great in many aspects, color rendition is almost perfect and light distribution is a joy. The only disappointment is the output - from 225W I was expecting more bang, but Mr. Dedo already stated they were optimizing for color rendition - especially for skin color rendition and the output is not so high as he would like it. This is the next step Dedo is going to change.

Just for comparison:
brightness of spot set to 30° and in 3m
Dedo DLED12 (LED 220 W) Daylight: 2.888 lx
Vibesta Ryzr 7 (LED 331 W) Daylight: 5.819 lx
Arri T1 (Halogen 1.000 W) Tungsten: 3.989 lx
Arri D5 (HMI 575 W) Daylight: 7.700 lx
Arri D12 (HMI 1.200 W) Daylight: 15.306 lx

I own the Rayzr7 300watt daylight LED and like it quite a lot. At 12Kg the Dedo is heavier and its CRI rating is slightly lower, but I'm sure it's built beautifully and likely is more even and controllable than the Rayzr7.

The CRI alone is not everything, but the latest generation LEDs are really good nowadays.
How do you like the Ryzr7? I thought it has not been released yet - at least that's what their website said two weeks ago when I checked last.

The CRI alone is not everything, but the latest generation LEDs are really good nowadays.
How do you like the Ryzr7? I thought it has not been released yet - at least that's what their website said two weeks ago when I checked last.

Beyond testing it, I've only gotten to throw it into real production on a couple shoots, but I love it so far. It's not as easily shapable as the Dedo's I've used, but for throwing a lot of light from a distance or through a window it has worked well. The dimming adjustment is very fine grained which has been helpful for dialing in just the right amount of light without having to mess with scrims or ND gels. The color quality of the light looks absolutely lovely on skin tones, no spikes or tint shifts that I can see in the footage.

It does have a fan which is relatively quiet and didn't pose a problem for audio. My two younger colleagues (in their 20's) complained of an annoying very high pitch noise from it that I (at 40) and the other older people on set could not hear at all. I guess I can't pretend to be young anymore The microphones didn't pick it up either. I don't think it will be a problem, but maybe if I was filming young children or animals it might annoy them. Who knows? I'll have to test that next time my little nephew visits.

Thank you for the comparison chart. I've been assuming it was about the same output as a 500W HMI and it's good to have that roughly confirmed by your numbers.

BTW - If you haven't come across it already, I think you'll really appreciate this article by Steve Yedlin on how color data can be remapped to take advantage of a sensor's best attributes while shifting the image into a preferred aesthetic. In other words, similar to your approach above where you've achieved lovely results out of the XF305, rather than accepting only what it delivers out of the box.

http://www.yedlin.net/OnColorScience/

EDIT: Forgot to answer your question about Rayzr availability. It says "pre-order" on the site still, but they are shipping in limited quantities as they are able make them. I think they should say backorder, as that is more accurate. It made my order and it shipped about 8 weeks later. Bit of a pain in ass as I had to pay import duties to DHL that I didn't realize weren't included in the original purchase price. Also, they mistakenly had my unit set for European power specs, so I had to take the whole light apart to get to the voltage switch and then put it back together. The whole process took two trips to the hardware store and most of a day, but I eventually got it to work properly.

.

So, as promised I did some comparison shots with skin tones.

* BMCC (Blackmagic Cinema Camera) in ProRes HQ Film 1080p 4:2:2 10bit (Tamron 17-50 2.8)
* Canon XF305 in 1080p 50MBit 4:2:2 8bit (built-in 18x lens 1.6-2.8) with 2 custom picture profiles
* URSA Mini 4.6k in ProRes HQ Film 2160p 4:2:2 10bit, downscaled to 1080p in post (Tokina Cine 50-135 T3)

So amazed with the images BMCC produces... looking at these pictures only, I have absolutely zero reasons to buy the 4.6k over the 2.5k.

Same as Wayne, I was only confident about Miss 4.6K with whom I'm well acquainted. She's my true love upon which I bestow all my riches.


Sent from my iPad using Tapatalk

OK, Rick -- what's your wife feel about this
Cheers

Shhhh! She might hear you!


Sent from my iPad using Tapatalk

OK Rick, mums the word
Cheers