F5: F55/C300/Alexa Low light test

Interesting that the colour (particularly the skin-tone of the clapper loader) seems to improve markedly as the ISO climbs on the F55, I wonder if that's simply a matter of the higher ISO pulling more accurate colour information due to a higher exposure hitting the photosites?
 
I commented on the post that it might have been interesting to also test 4k on the F55 down-rezzed to HD to observe the different in color noise, etc.. I think they'd find some differences there.
 
Well the ipad does seem to have pretty low capability in terms of what it can display according to this: (although the ipad3 is nearly not as bad as the ipad2)

 
I'm still deciding but not sure I agree that using an ipad3 could be all that bad.

Reading more about it, the ipad3 has one of the widest color gamut from other competitor tablets not to mention exceeds even some computer desktop displays.

I think most people would be surprised to know that their computer monitors probably don't reach 100% of RGB and that an ipad3 is probably higher than the monitor they use.

I recently bought a Datacolor Spyder 4 pro and started measuring the RGB output of various monitors that I use, and so far the Apple LED 27" cinema display is the only one that can get a 100% rating of RGB. All my other monitors vary in the 90's range.
 
Last edited:
You can be as snarky as you want, bu a proper chip chart for both color and brightness should be a reflective device not transmissive, or at least transmissive with a known color source. So a cardboard chip chart (just a few bucks) and an inkie are a lot more accurate than the flourescent backlight of an iPad. If you do want to go expensive then you can always get a DCS backlit transparency setup like we have in our calibration test room, but something tells me you'd rather just hit me up with some attitude.

Cheers.
 
Snarky is when you trivialize other peoples work out of hand. I learned quite a lot about the cameras from the test.

I've found you can learn a great deal from virtually any type of chart that you can refer to later when viewing what you shot.

I use flowers quite a bit.

Cheers.
 
Admit One said:
Snarky is when you trivialize other peoples work out of hand. I learned quite a lot about the cameras from the test.

I've found you can learn a great deal from virtually any type of chart that you can refer to later when viewing what you shot.

I use flowers quite a bit.

Cheers.
Click to expand...

Usually Mitch is very helpful and informative. Since he usually has the support of Abel Cine behind him I think he's used to having the gear and charts so I'll give it a pass. Hey, some of us don't have color charts. I don't know that paint chips from the hardware store that I would use would be any more accurate.

No matter what is used if it's consistently seen by each of the cameras you can learn something from it. Changes in color from camera to camera if nothing else. I do get his point about a cell phone probably not producing an accurate color chart, but... there's always someone (usually well intentioned) that will not like some aspects of any test. I go with the thought that you can learn something from any test.

No test is perfect. This one has at least tried to cover several bases. I mean, jez, the monitors we're viewing the test with are probably not accurate either.

I remember years ago I had a really nice Sony Pro (tube) monitor. Beautiful images on that monitor. However, any TV set we played the video looked different. So you could set your colors to the one monitor that no one else but a few broadcasters have, or set them to look good on as many TV sets as possible. What do you choose?
 
Last edited:
Thanks Mitch for my first laugh of the day.
The test may not have been perfectly scientific, but I appreciate the time and effort to share.
I found it interesting anyway.
 
Hi All,
Thanks for your interest in my test and for your comments. It is, of course, a valid point that the iPad is a source of light. This is one of the areas where the test is less than optimal. A sudden change in our shooting schedule meant the night earmarked for the test was no longer feasible and in addition a change in the weather forecast meant the test could either be done immediately or not at all. Unfortunately, the ensuing rush left no time to organise getting a Colour Chart. I opted to use the iPad as opposed to no colour chart at all.I have also to point out that this is the first camera test where I have circulated my results and would like to clarify the background. First and foremost, this is by no means to be compared to a meticulous, well-planned studio test undertaken in a controlled environment as, for example, the Zacuto camera shootouts. The test I did came about as a result of a night scene where we had to shoot a chase throughout a small town. Our budget did not extend to traditional lighting solutions involving cranes, additional lights and crew. We had a practical problem which I had to solve and solve fast. My solution was to test the cameras to establish whether we could achieve useable footage both from a technical and aesthetic point of view as well as in accordance with the script and director's wishes. We found the results so interesting I decided to share them despite some obvious shortcomings.

Thomas Ch. Weber, Cinematographer, Germany
 
My main observation from a very quick look at the results was to feel the F55 handled the sodium streetlighting far better than any other video camera I've seen - and better there than the C300 or the Alexa. Far more like I'd expect them to look to the eye. I don't like drawing too many conclusions from online postings, so I'd be interested to hear what Thomas thought about that?

I'm also a bit surprised that set to 4500K, the car lights etc seem a bit on the cold side? Would that be because they were LED headlights, not halogen? Difficult to be sure, but I think I'd find a colour balance of 5600K may give an overall more pleasing result?
 
harddrive said:
My main observation from a very quick look at the results was to feel the F55 handled the sodium streetlighting far better than any other video camera I've seen - and better there than the C300 or the Alexa. Far more like I'd expect them to look to the eye. I don't like drawing too many conclusions from online postings, so I'd be interested to hear what Thomas thought about that?

I'm also a bit surprised that set to 4500K, the car lights etc seem a bit on the cold side? Would that be because they were LED headlights, not halogen? Difficult to be sure, but I think I'd find a colour balance of 5600K may give an overall more pleasing result?
Click to expand...

Yes I noticed the sodium vapours looked pretty good and the harsh edge artifacting is gone that existed with the original firmware when the F55 first shipped.

These days there are a variety of color temperature headlights you can get for a car (kids think they are cool) sometimes on-purpose to the extreme "red" or "blue" side of the spectrum.

As for the color chart on the ipad, I think as long as the colors are produced consistently each time regardless of their color space it would still be pretty useful for getting clips from various takes or camera to match up better than without it. I'm curious to know why others would think differently?
 
I'm with Dennis on this. The ipad as color chart is useful in this instance as you are trying to match 3 cameras . At least you have the consistent (if wrong ) colors on the ipad to match the camera output. If you don't have power and you don't have a real chip chart with you would it be better to shoot nothing?
I'd much rather see real world tests like this than a bunch of charts. A lot can be learned from these real world tests.
Thomas, thank you for posting these.
 
harddrive said:
My main observation from a very quick look at the results was to feel the F55 handled the sodium streetlighting far better than any other video camera I've seen - and better there than the C300 or the Alexa. Far more like I'd expect them to look to the eye. I don't like drawing too many conclusions from online postings, so I'd be interested to hear what Thomas thought about that?

I'm also a bit surprised that set to 4500K, the car lights etc seem a bit on the cold side? Would that be because they were LED headlights, not halogen? Difficult to be sure, but I think I'd find a colour balance of 5600K may give an overall more pleasing result?
Click to expand...

Thanks for commenting on my tests. I think when it comes to discussing color reproduction there is no such thing as right or wrong, it's more a question of taste. For me the Alexa is the one with the best colors, especially for skin tones, but in this case also for the sodium vapor zebra crossing lights. They were indeed as orange as you can see on the Alexa shots and photographs from several different still cameras showed the same type of color as you can see in this album on flickr: http://www.flickr.com/photos/40003133@N08/sets/72157634732265838/. The F55 exaggerates the yellow a bit too much for my taste and I like the C300´s color even less as I feel it gave my pictures in this particular test (not in general) rather a plastic look compared to the other two cameras. We also shot another scene for a fashion show using both an Alexa and a F55 simultaneously. There was a good script-related reason for doing this but this short post isn't the place to go into details. I can only publish these shots after they ave been broadcast, unfortunately, but interesting thing is that the two cameras matched tremendously well. Nobody would guess that the shots came from to different cameras. However, here I also noticed that the F55 seemed to me to be a bit harder and the yellow stronger.


A lot of cars in Germany nowadays have LED headlights as was the case with several cars in the test. They didn't concern me in this case, however, and I ignored their color as no cars were involved in the actual shoot but I filmed them anyway as I was interested to see the performance of the dynamic range of the cameras. If I had had to film the scene with LED headlights as a major light source I maybe would have set the camera to 5600K or higher. For the street light scene I choose 4500K as a happy medium between all the different light sources involved: fluorescent, sodium vapor, and tungsten. At 5600K the tungsten would have become too warm for my taste.

cheers

 
This thread was both informative and funny at the same time, now old and from quite a long time ago.

Anyone since using an ipad2 as a color chart, or something similar? I'd love to start walking around a set with an iPad and having an assistant hold it up, it just seems futuristic and for sure 100% if the clients saw it, they would be impressed

:cheesy:
 
You must log in or register to reply here.
Back
Top