GH2 Philip Bloom Christmas Shootout

[Rick Burnett]

Knowing the noise floor of your camera has many more implications than just shooting controlled lighting. When pushing up into the ISO values, it also controls how far you can push your footage. Having worked with a 7D since its release, I can tell you, noise gets to be a real problem at times in situations where I have a lot of dynamic changes in lighting during a scene. Those can be hard to shoot.

I've worked with an HVX200, 7D, AF100 and now an FS100. If everyone thought that low light didn't matter, they'd be still shooting on a HVX200 with a S35 adapter. The truth is, it does matter. And I'm not talking artistically, I'm talking about having to add lots of light to a scene just to get within a sensors sweet spot. It takes a lot of time, power considerations, and cost for equipment (whether rented or not) when using a less sensitive camera. There is nothing wrong with this, they've been doing this for years with the likes of film and such, but on the other hand, there is nothing wrong or unprofessional about not wanting to do this either.

With the FS100, I can light scenes with small lights tucked away behind props in ways I'd never imagined before. I've used flashlights to get some really great accent lighting. I've also been able to shoot at night AND at night at 1/120 and still keep the noise much lower in scenes where actors are in and out of situations I could have NEVER done with the HVX200 without a huge expense.

THIS is why low-light is important to me. Given I know the 7D, AF100 and FS100 like the back of my hand, and seeing them in tests like this, it allows me to get an idea of the performance of the other cameras at least a little bit. When I first saw FS100 low light footage, I didn't believe it. Then I bought the camera, now I know how well it performs.

However, this is ONE aspect of the camera. People are really getting hung up on ONE aspect. Low-light is not the end all be all either of the camera. For instance, while I think the GH2 does really well on many tests, it still does horrible on the rolling shutter test. This affects me, especially when I use my 85mm lens. This is why I do not use a GH2. For others, they may not care, or shoot wider, or always lock down their shots. Even with the AF100/FS100 which have practically the same exact rolling shutter, it becomes a problem at times. (And they are not that far from the rolling shutter performance of the original Red One, not the MX version).

Also, what these tests don't show you is things like the ND filter in the AF100 (and F3 or C300). Once you use ND in your camera like this, you will miss it TREMENDOUSLY. EVERY time I use my FS100 I miss the ND. Every time I used the AF100 or FS100, I missed the communication between my lens and the body on the 7D.

Take each test for what they show you. Noise performance typically scales between the ISO values (although the AF100 is strange in that if you've used it, the noise stays strangely consistent over a wide range). With this test, from cameras you know, at least you can get a relative idea of how they performed between each other.

 

[roei z]

^^^^ this

 

[magnifico]

... With this test, from cameras you know, at least you can get a relative idea of how they performed between each other.

Hi,
if you dont know any of these cams you cannot do a precise comparison for low light. Why? Because it is an extreme low light test, a NO light test in witch no camera gives a decent image. If only one camera gave a decent image, ok we can say something. Of course you can do some extrapolation here, well well. But as far as gh2 is concerned it is difficult to do so. Why? Because it won the resolution test in part 1.
That means if you want to shoot outside with light, you get the best cam here among the others. And with the hack and GOP1 you obtain a good fluidity for motion. That means you can borrow gh2 and just get the best footage for your broadcast film.
gh2 is not a war beast concerning sensor, that means we can guess that process, software and processor inside it are very good. The PB test doesn't give the limits for all the cameras in low light. Hay I dont blame Philip here, I just try to say what I am thinking ok? How many stops is gh2 behind (after denoising) c300, 7D or fs100 ? Because for sure there is a threshold where gh2 and fs100 are on par concerning low/average light.
If you have answer please give it to me (hm us) as I am very interested...

 

[stoneinapond]

...But as far as gh2 is concerned it is difficult to do so. Why? Because it won the resolution test in part 1.
That means if you want to shoot outside with light, you get the best cam here among the others...

I really don't think that is an accurate statement. While the GH2 does deliver great footage out-of-camera, that is not the whole story. If you are looking to do professional color correction, then some of the other admittedly dull looking images from the other cameras will be your friend before the GH2.

I love the GH2. I cannot afford, and at present don't want a different camera. I love the form factor and I love the out-of-camera resolution. But I'm also learning to color grade, and I already know for a fact that there is a point beyond which you cannot push the footage. And that for some is a very professional consideration.

 

[Rick Burnett]

Hi,
if you dont know any of these cams you cannot do a precise comparison for low light. Why? Because it is an extreme low light test, a NO light test in witch no camera gives a decent image. If only one camera gave a decent image, ok we can say something. Of course you can do some extrapolation here, well well. But as far as gh2 is concerned it is difficult to do so. Why? Because it won the resolution test in part 1.
That means if you want to shoot outside with light, you get the best cam here among the others. And with the hack and GOP1 you obtain a good fluidity for motion. That means you can borrow gh2 and just get the best footage for your broadcast film.
gh2 is not a war beast concerning sensor, that means we can guess that process, software and processor inside it are very good. The PB test doesn't give the limits for all the cameras in low light. Hay I dont blame Philip here, I just try to say what I am thinking ok? How many stops is gh2 behind (after denoising) c300, 7D or fs100 ? Because for sure there is a threshold where gh2 and fs100 are on par concerning low/average light.
If you have answer please give it to me (hm us) as I am very interested...

If you don't know any of these cams then really, it's going to be hard for you to really extrapolate anything useful out of these tests. They aren't intended to be the definite hands down answer all questions comparison that you seem to want. There is no way to test a camera to this level that could EVER satisfy people.

Look at it this way, if they set the cameras up to where they all shot at ISO400 and there was no noise to be seen, then what is the point of the test? You aren't showing anything. THE point of the test is to show the cameras shooting the same scenario in which you CAN see a discernible difference. It's pushed darker so the differences are easier to see.

And no, I don't believe the GH2 won the resolution test in part 1. It scored well, but I think the resolution on the C300 looked better to my eyes Second, again, you are focusing on ONE merit of the camera. If you need to move the camera around a lot shooting, then no, the GH2 might not perform well there. It just depends on how it works with all it's GOOD and BAD performance in the situation you need.

And here's the thing, you can't give a limit to where a camera is usable in low light because that limit is HIGHLY subjective. Go to the movies in the theater, watch films with night shots, you will notice a WIDE range of noise. Not only is the amount of noise one problem, the pattern of the noise, the color of the noise, the artifacts of the noise cleaning, it goes on and on.

Take it further into color grading and what happens to that noise becomes another problem as the noise is different.

You want a simple answer, there isn't one. Hands on with the cameras you want to know about is the only way you will really know. This test is for people who at least know one or two of the cameras being tested. It's easy to understand the results and appreciate them with that knowledge.

 

[magnifico]

@stoneinapond
thx for this information. For the first test of Philip Bloom I just focused in definition of the footage. And there, c300 and gh2 were my favourite. That appeared somehow objective for me. I didnt focus on colors since there were clouds etc.
In Personal View forum they started a project about grading skin tones. I am going to learn here about it.

@Rick Burnett
I followed the zacuto test 2011...and for me it was quite very obvious to select the best footages in the low light test. Considering details etc. But yes I think there are a lot of other factors may be a pro will stress and I wouldnt even notice. i am learning photo & video and I appreciate answers like yours. Thx.

Nota: the point is you said "ISO400". May be here limit will be subjective considering gh2 vs fs100. At ISO800 may be again they are on par there. At ISO1600 may be fs100 is better because showing here more detail and/or better color...That is what I wanted to see : the subjective analysis of the pro, the objective evidence of some new artifacts that disqualifies the next level.

 

[Rick Burnett]

@stoneinapond
thx for this information. For the first test of Philip Bloom I just focused in definition of the footage. And there, c300 and gh2 were my favourite. I didnt focus on colors since there were clouds etc.
In Personal View forum they started a project about grading skin tones. I am going to learn here about it.

@Rick Burnett
I followed the zacuto test 2011...and for me it was quite very obvious to select the best footages in the low light test. Considering details etc. But yes I think there are a lot of other factors may be a pro will stress and I wouldnt even notice. i am learning photo & video and I appreciate answers like yours. Thx.

I can tell you, when I was starting out not too long ago, I watched hundreds of reviews, tests, comparisons, from every possible location I could find, and continue to do so even today. One thing is for certain, if you *really* want to know how one of these cameras works in your world, you've GOT to get your hands on one. Rent.

I use this example often, but I bought a D90 when they came out (Nikon) thinking it would be an amazing camera to shoot video with. The reviews at the time were great, the videos were awesome, and I was of course still learning a LOT about cameras. I get the camera only to find out, there is no way to set manual exposure. The rolling shutter is worse than any other camera I have ever used, and the line skipping causes artifacts every so many rows (I think 8 or 9). NO ONE at the time talked about these things, and of course, I didn't know to even ask about them. It was really a pain to shoot with.

Of course, I still have the D90 as it is a fantastic stills cameras

I say that here because people get hung up on one or two things thinking they will make the difference to all the other possible shortcomings. I was focusing on shallow DoF, which was great, but didn't even consider a million other things that were important.

With these tests, it is important to take that into consideration as well. There are SO MANY other factors on ALL these cameras that which is right for you is a tough decision for many people. These aren't meant to say camera A is better than camera B in any other context then just what is being tested, and even then, most reviewers will admit that their conclusions can be subjective in some cases when performance is very close.

I've used my 7D even today for shots where the FS100 does not fit, and I cannot modify the shooting location. I don't want to necessarily use it, but for me, the framing and location with the 7D justified the negatives of the image. It's all a matter of importance.

 

[magnifico]

The test of PB is fun; i am longing now for the next shootout of zacuto (the revenge...) with gh2 and other cams...

 

[Mike@AF]

You are assuming that everyone is buying these cameras for making movies.

A good deal of DSLR's are being used to shoot documentaries and even ENG work. In those situations you capture the action as it unfolds, often with no control over the environment. In such circumstances the cameras low light performance will be a deciding factor, along with everything else.

So now it's "let's shoot with no light so it looks crappy because this camera can shoot well with very little light"? It's like people are forgetting about what used to be done when cameras needed a lot of light or no one has heard of battery-powered on-camera lights. That would be much better than no lighting. Even docs should look decent.

Edit: Also, even if you couldn't use a light at all, you wouldn't use a f2.8 lens. With the GH2 you'd use a Voigtlander 25mm f0.95 lens and wouldn't have the exposure issue. You'd be able to stay at ISO 400 or 800 at worst and then get decent exposure and much less noise. Low light tests like this are just unrealistic and not of value in my opinion.

The skin tone test was good though.

 

[Rick Burnett]

So now it's "let's shoot with no light so it looks crappy because this camera can shoot well with very little light"? It's like people are forgetting about what used to be done when cameras needed a lot of light or no one has heard of battery-powered on-camera lights. That would be much better than no lighting. Even docs should look decent.

Why does it have to be an extreme viewpoint one way or the other? Low light doesn't have to be no light. Why not embrace technology that has advanced to make your job easier? Just because people used to use a huge amount of light doesn't mean that's how it has to be done OR that using less light means it will look crappy.

There is no advantage to using a ton of lights because you use a camera that is not very sensitive compared to using less lights on a much more sensitive camera (all other things equal). it adds cost in a ton of different ways with lights, power, hands to operate them, transportation. Further, depending on how much light you need, it also limits your freedoms on set in what kind of dynamic shots that you can make.

With a camera that is more sensitive, you can use lower power lights where you need them and get the same results. You can diffuse lights even further to spread them out. You can run them off of batteries, hidden behind props. It really allows you to look at doing things a bunch of different ways that you just didn't have those options before.

Doing something a harder way doesn't make it superior, it's just a different way.

I don't get this hostility people have about low light performance. If you can't understand the advantage of it that many of us see, then just move on to the other tests. For me, this matters. For me, this is where I want to see what the camera does. I understand lighting, I've spent a great deal of time looking at all the ways to light a scene. I put my efforts into the artistic merits of the scene lighting. A low-light sensor doesn't do anything to change the artistic needs of lighting and I can say that I've understood that from day one.

But, it's also not just about low light, it's about slow lenses as well. Or, deep focus on a large sensor. You can get better performance with a lower noise sensor at these settings. Again, it's about freedom.

There are a lot of amazing things that get discovered when people push equipment where it wasn't used before. I'm glad the person who discovered distorted guitar sounded great didn't listen to all the people that probably told them, no, you need to use it clean like everyone else before had amplified their sound.

 

[Mike@AF]

Why does it have to be an extreme viewpoint one way or the other? Low light doesn't have to be no light. Why not embrace technology that has advanced to make your job easier? Just because people used to use a huge amount of light doesn't mean that's how it has to be done OR that using less light means it will look crappy.

It's not about just using less light. What was demonstrated was basically using no light. I think it's great that the cameras are more sensitive than they used to be and I do embrace it. However, it seems to be getting used as an excuse to not light at all.

There is no advantage to using a ton of lights because you use a camera that is not very sensitive compared to using less lights on a much more sensitive camera (all other things equal). it adds cost in a ton of different ways with lights, power, hands to operate them, transportation. Further, depending on how much light you need, it also limits your freedoms on set in what kind of dynamic shots that you can make.

I never said anything about using a ton of lights. I mentioned a simple battery-powered on-camera light. That's not a ton of light. In fact, it's very little light, but is enough to get a much better exposure and less noise than the unrealistic test done in Phillip Bloom's shootout/test/comparison, especially if a fast lens is used than f/2.8.

I don't get this hostility people have about low light performance. If you can't understand the advantage of it that many of us see, then just move on to the other tests. For me, this matters. For me, this is where I want to see what the camera does. I understand lighting, I've spent a great deal of time looking at all the ways to light a scene. I put my efforts into the artistic merits of the scene lighting. A low-light sensor doesn't do anything to change the artistic needs of lighting and I can say that I've understood that from day one.

It's not hostility about low light performance. It's about poor lighting and the absence of poor lighting and a test/comparison that's unrealistic and really has no bearing. The Zacuto/SCCE 2011 shootout is a much more realistic example of how to perform a low light test in a shootout.

 

[David G. Smith]

The Zacuto/SCCE 2011 shootout is a much more realistic example of how to perform a low light test in a shootout.

Mr. Bloom said that very same thing explicitly in the video.

 

[TheDingo]

a test/comparison that's unrealistic and really has no bearing

I think Philip admits this right from the get go. His test is more of a "can I break the image these cameras produce by pushing them far beyond any realistic lighting situation" test, and it shows that under absolutely worst case scenarios the Canon C300 and Sony F3 / FS-100 can produce a more acceptable image. ( definitely not a great image, but it's a salvageable image if you had no other options )

Hopefully most viewers will realize that adding a few hundred dollars of inexpensive lights would produce a far better image with all of the cameras tested, than anything we are seeing in Philip's "no lighting" test.

 

[yoclay]

I think that it is not simply a question of using lights to boost a low light situation. I dispensed with that notion a long time ago. In reality lighting for me is always about sculpting a look. Truth be told, most shooting in low light is now happening so far beyond the old limits set by film that it is all gravy. That said, I know a lot of people are pushing this camera to 12,800 and that is great, but the blue channel noise is just outrageous and totally unacceptable by my standards. What concerns me the most remains the noise levels which are happening going into the shadows of many very normal, low ISO situations on the GH2, especially on flat surfaces with gradients. While I really like this little camera (reminds me of the HV20's popularity a few years ago) I have yet to see a test about that particular issue on a variety of cameras and I am certain that my perception of a quality image is also based on this kind of artifacting. Show me a test about how these cameras handle gradients at a variety of iso's and I will show you the camera which is my next purchase. Denoising/smearing in post is not a substitute for a clean output in my book.

 

[maarek]

It's about poor lighting and the absence of poor lighting and a test/comparison that's unrealistic and really has no bearing. The Zacuto/SCCE 2011 shootout is a much more realistic example of how to perform a low light test in a shootout.

Philip isn't testing lighting. He is testing extreme low light. As an extreme low light test, it was pretty good. We could FINALLY see that the C300 is not more sensitive then the F3 / FS100. No other test has even shown this yet. We could see that the GH2 was not as good on lowlight as some other cameras.

It all makes sense and is correct. If you want to test ISO 800 and then look at the shadow noise - great - but that's a completely different thing.

 

[maarek]

However, it seems to be getting used as an excuse to not light at all.

You don't get it. The FS100 is so sensitive that you can be at ISO 100 longer than other cameras. You can use smaller lights, a more natural look easier. Blasting a house full of light is not natural. Now you can argue "who goes for that kind off look?" but it is getting more and more usual. The film Drive had excellent shots done with the Mark II in pretty natural lighting.

The tv-show Game of Thrones had very unnatural looking lighting, especially when watched in HD. Everything looked lit! I would wager a more artistic style with less light would've done wonders for it. Offcourse some people like that style. Some don't. Those who don't now have an option to make more interesting looks.

What you are basically saying is "I light everything with dozens of 4k lights and so should all".

http://thebuibrothers.com/blog/2012...ney-charters-drew-gardner-and-lan-bui-part-1/

 

[Rick Burnett]

To me extending the difference in the noise performance in extreme low light helps to understand the difference in the low light. It's not meant to be "here is where low light makes sense", it's "here is an apples to apples comparison on low light performance between these cameras". This way, you can compare how they differ. You want it easier to see the differences. Like I said, why would I want to see all these cameras performing really well at some low light situation where they weren't being stressed?

I understand the viewpoint of wanting to see where these cameras remain usable, but each camera has different limits. Clearly that is a lot of work and would still be wrought with subjective criticism. The next problem would be gamma curve differences, noise reduction settings, detail settings, and probably other things as well. This was a quick test to be able to rate the SNR between cameras in a way that is easy to see. For that, I am grateful. I've done camera tests. I know how much time they take. And I definitely know how ungrateful people are for the time that goes into them. Or, that people fail to see they are reading more into a test than was intended. I think Philip and crew did a great quick test that at least I can see a lot of information from. I think a lot of FS100 and F3 users will be happy to see how their cameras sit in this test.

If I can get my hands on a C300, I will do the rolling shutter measurements like I did on the FS100, AF100 and 7D. At least that test is not subjective!

 

[Apefos Adapter]

I am editing a video tutorial how to make gh2 low light image match the c300 and fs100 low light image quality.

I will upload the tutorial to vimeo as soon as i finish editing it.

 

[Mike@AF]

You don't get it. The FS100 is so sensitive that you can be at ISO 100 longer than other cameras. You can use smaller lights, a more natural look easier. Blasting a house full of light is not natural. Now you can argue "who goes for that kind off look?" but it is getting more and more usual. The film Drive had excellent shots done with the Mark II in pretty natural lighting.

The tv-show Game of Thrones had very unnatural looking lighting, especially when watched in HD. Everything looked lit! I would wager a more artistic style with less light would've done wonders for it. Offcourse some people like that style. Some don't. Those who don't now have an option to make more interesting looks.

What you are basically saying is "I light everything with dozens of 4k lights and so should all".

http://thebuibrothers.com/blog/2012...ney-charters-drew-gardner-and-lan-bui-part-1/

I did not say that at all. Go back and read what I said again. I said an on-camera light should be used. That's one light that would make that shot look better. My lighting setups are usually very simple with usually only 1-4 lights and fairly low power lights. I'd rather ND a window than put up 4K lights indoors to compensate. I don't think Game of Thrones looks "unnatural". It looks good for the period it's supposed to be set in.

 

[JoeJITSU]

You're looking at the cameras in the most Incompetent situation. I actually agree with the way he did this, and I can say I expected as much. If you want an extreme lowlight camera the Gh2 or Canon cams are not those. They're all pretty gross in that regard.

What you know now though is that this is the bottom floor for performance. A mere 1K of wattage in this specific setup would've cleaned all of that up drastically.

I don't even like the FS100's out of cam image at high ISOs.

F3 w/s-log will be something I keep my eyes on as an option now for strictly no light shooting. Also reminds mr that I haven't tried something similar to this with Epic. I'd imagine an RCP-X denies when converting to 1080/2K from 5K would look pretty straight.

I totally agree with you also how Phil shot this. But I must say...The 5D Mark II did look noticeably better than the other DSLR's. I'm think its because of its full size sensor.