GH2 Philip Bloom Christmas Shootout

c3hammer said:
He didn't chose a zoom for the low light test. Did you even watch the video? Here's a frame from 19:27 of the low light setup and another from the skin tone test with the big zoom. Clearly not the same lens.


View attachment 45839 View attachment 45840
Click to expand...

I see your point, but I didn't see anything written or spoken about the lens used,
F stop and the test methodology was not clear for the low light test.

I am not doubting the end result, just trying to understand what he actually did to get the end result.

So, we really don't know what we're looking at still?
 
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.
 
Late said:
He tested high ISO capability, not low light capability. Low light capability depends on both the sensor and lens, not the sensor alone.
Click to expand...

NOt sure what you're trying to say. He didn't test ISO alone; he tested low light capability, which is a combination of factors engineered into the cams, including sensor size, codec, compression algorithms, iso, etc. ISO alone doesn't determine low light capability. Conversely, ISO capability is not measured solely by low light capability. Certainly high ISO settings affect quality of low light images, but I use ISO settings for far more than just to get an image in low light.
 
What I would say to anyone that doesn't agree with his test, do your own test how you want to then. I think Philip and others do a fantastic job of showing us all what we normally wouldn't see and quite frankly I am tired of people ripping it because it didn't necessarily show you the results you wanted with your camera. I have news for you - your camera is not perfect. Neither is mine. None of them are. Philip is VERY knowledgeable when it comes to camerasand how to perform tests, not that he doesn't make mistakes, but that's his thing. So if you think you can do a better job, go for it and you see what the response is. I for one appreciate that Philip went through the trouble to do this (and continues to do this despite nay-sayers), shooting, editing, etc. His time is valuable and he did something that benefits US, the user.
 
Kholi said:
They're all pretty gross in that regard.
Click to expand...

It's funny how far we've come in a couple of years. When I first used the Canon 7d with an f1.4 lens I was completely astonished. No videocamera before had given me such latitude in low light situations. A couple of years later and it's gross, heh.
 
Personally, I think low light tests are mostly pointless. No one shooting anything serious is going to shoot in those kinds of conditions. At least he put a disclaimer in the video that you wouldn't shoot like this. But then, if no one would shoot like this, why do the test? I would rather see a properly lit scene with dark areas in the frame and see how details in the dark areas look. I think the SCCE Great Camera Shootout was better for this type of test. I've shot with the GH1 and GH2 in low light and get better performance than this. You'd go with a fast lens wide open and not need to increase the ISO as much. Understandably, the other cameras performed better in this particular test, but I know the GH2 can do better when used properly so I'm not disappointed.
 
No, you would not use them in situations like this normally, but take the results of the F3 S-Log to Card (isn't it 8-bit 4:2:2 or is it 4:2:0?) and what we saw here: I know I can get something similar out of Epic/RED after the downconvert from 4K/4K to 1080/2K, and a denoise. But, being able to get that in camera is pretty sweet. SO this worse (awful) case scenario helped me to see that there's a potential for some very run n'gun material with the F3|S-log|Straight to Card setup, one that I had not seen before to be honest.

Doing these same kinds of things with the GH2 over the past month told me what I can and cannot do with the camera. It's a good thing, at least to me.
 
What is great about these tests is that under controlled circumstances these cameras can be used together for their respective strengths, and the end viewer will never be the wiser. I'm too lazy to find a link right now, but I found this great mash-up on vimeo that a camera store made that threw together clips from practically every camera in the store, and the only ones that really stood out were a compact point and shooter, and the go-pro fisheye. Everything else was pretty much interchangeable in 2 second edits, after output compression.

I'm as far from a pixel peeper as you can get. I still believe that dvd is high enough resolution for home viewing. A 4k future is ridiculous. Unhacked GH2 looks good enough to me in most normal situations, but I don't like the amount of jell-o in the image, same for the the nex 5n. What is the point of a tiny camera if you can't just pick it up and shoot. If the next generation can get this under control then I'm all in. I've said it before and I'll say it again, if I had the cash I'd get both of the 100's, sony and Panasonic, and use each to its strengths, and a set of manual lenses that can be shared alike. Barry Green proved that if you're careful about it in a proper lit scene, you could shoot 2 cameras and not know the difference. Take the Panny outside with the built in ND, keep the sony for nighttime city scapes and candle lit interiors, even at f4 with a chance of holding focus. Keep in perspective that these cameras give us the 35mm look for pennies. A dslr body is like one 10 minute roll of film and lab fees. $5k might get you and hour of raw footage. $20k for damn sure wouldn't get you a feature, you'd be lucky if you could manage 16mm. And that doesn't include renting a camera to run it through.

I can totally see the canon shooting a lot of low budget cable movies, the built in broadcast codec is the main selling point. If only they would mass produce the chip so it could be priced where we want it to be. The codec and processor are available now in the $3k xf-100, and that has a lens. They're trying to sell the same RED kool-aid that the image sensor alone is worth $10k+.

What needs to happen is enough good content is produced in AVCHD and a proper broadcast ready post pipeline, ProRes or whatever, and get it accepted as a broadcast standard, just like dv killed betacam. The high bit-rate hack is cool and all, but having a standard that can be shared with $100 consumer editors as well as pro level workstation is the key to the revolution. SD cards should soon be cheap enough to be saved as masters on important projects, or at least have enough of them to not need to recycle as often, risking precious footage.
 
I agree that it's good to know what the low light capabilities of a camera are. However, I would not choose a camera because of it's lowlight capabilities in a scenario like this because it's an unrealistic scenario 99% of the time. More important tests and comparisons with other cameras in areas of resolution & detail, codec quality, color correction capability, and color accuracy. I would choose my camera first, then test the limitations of it so I know what I can and cannot do with with and then be able to work around it. If the camera isn't capable of shooting well in a situation such as that in Phillip Bloom's test, then I would light it better and/or use a faster lens. So if I decided on a GH2 (which I have and even sold my AF100), I would realize that I spent a fraction of what the other cameras cost and the noise in such situations would be no surprise. I would use my faster lenses, such as my Voigtlander 25mm f0.95 and even portable battery-operated lights and still shoot at ISO 400 or 800 to keep noise down. This gets me the superior intraframe codec and resolution/detail of my hacked GH2 and less noise because it's actually lit properly.

I think it's a shame that a camera would be chosen because of it's low light performance. It just opens things up for bad lighting and cinematography. Did the FS100 expose in low light with little noise? Yes. Did it look good? No. Just because you can shoot in low light doesn't mean you'll end up with a good image. Lighting is always better.
 
Mike@AF said:
I think it's a shame that a camera would be chosen because of it's low light performance. It just opens things up for bad lighting and cinematography. Did the FS100 expose in low light with little noise? Yes. Did it look good? No. Just because you can shoot in low light doesn't mean you'll end up with a good image. Lighting is always better.
Click to expand...

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.
 
I think that Mr. Bloom was very upfront on the value of this test, and his testing methods. I think that a little perspective is in order when considering his results. OF COURSE, no professional filmmaker is going to shoot under the circumstances that were set up for the low light test. I think that the true value of the low light test is its showing us a situation where no professional filmmaker is going to shoot in. What we need to realize is that these situations where no professional filmmaker is going to shoot in have now become ridiculously extreme, because of these cameras. However, I think that all of us, no matter what camera we have chosen to shoot with, should take a step back, take a deep breath and be thankful for the powerful tools that we have available for us to use now. All of the cameras he tested would make much better images in situations where there is more light than just, "A sliver of light from the toilet", which was the case in that test. THAT IS GREAT. We have never had such a power set of tools available for making great moving images, at such low costs, EVER! Not a single camera Mr. Bloom tested is a bad camera. All of them have strengths and weakness' that creative professionals can evaluate and make their choice of camera to use based on their own individual needs. It is great times.
 
Last edited:
Lets see things as they are:

OK, the C300 and FS100 are the low light kings. If you can buy or rent, use them.

But, if you cannot buy or rent a camera like these, do not worry, use the camera you can afford and respect its limitations.

If your camera is not so great in low light, just add light and/or do a denoise in post.

GH2 is not the same of c300 and fs100 in low light, but it is not so bad, with a carefull denoise it can give a good result. Or just add light.

Deal with the limitations and be happy.
 
Apefos Adapter said:
Deal with the limitations and be happy.
Click to expand...

What? No! We have to convince everyone that our camera is the best and that the test was flawed, and if Philip did a proper test, you would see! Why would you ever want to use lighting for video? :)

For the record, this is me being sarcastic.

Some people just don't understand. NOBODY'S CAMERA IS PERFECT. The first step is admitting it.

I'm glad you understand, Apefos.
 
My conclusion of the test are


Best bang for a buck the gh2. Exceptional value, with very high resolution, good dynamic range (better than what you see on the test) and ok/average low light. what would be really the next big thing for the hack gh2 would be a flatter profile or permanent idynamic range. The gh2 has more than the apparent DR with the hack (in the shadows) and those two above could bring some more.


Best middle upgrade is the fs100 with an external recorder like the Ninja. Already very good resolution, seems to have very good dynamic range with the glock picture profile and exceptional low light. I think with a cheap recorder this could be a very good upgrade to the hacked gh2 in someone has the money. It is said to have much less rolling shutter than the dslr. It has a hair less resolution, but with better DR and much better low light. In fact apart from built quality I don't find the c300 bringing anything that warrants the 3x price.


The best for me in this test if you have the money, is the sony F3. Very good resolution and exceptional low light and dynamic range. We have been seeing it in the test with its internal codec. But with an $ 1400 Samurai 10 bit slog at 220mbits, it should be even better and I still can't understand the positioning of the C300.
 
@Mike@AF +1
Hi,
Philip Bloom chose a Extreme (his word) light test. In fact it became a NOISE test. Ok it is fun ...It just killed small-photosites cams :-)
As I said in Vimeo forum:
" I know c300, fs100 are better cam than gh2 in low light condition. What I want to know for sure is : where is the limit in low light in which gh2 and fs100 (or 7D) are on par concerning details (with little moving of the subject) ? Let’s say gh2 >7D in average luminance concerning resolution. If we decrease little by little exposure we shoot, we gone reach a threshold where cams are on par concerning details. For that, we must consider post process denoising belonging to the whole process, because it is what we do in low light situation.

So where is the limit? 1600 iso for gh2 compared with c300 or 7D? As far as I am concerned it is the main thing I ask for myself.
No answer here ( in the Philip Shoot out test n2)because of no denoising and the lowest light limit . "

That is why I just ask the same question here ...
 
Danielvilliers said:
The best for me in this test if you have the money, is the sony F3. Very good resolution and exceptional low light and dynamic range. We have been seeing it in the test with its internal codec. But with an $ 1400 Samurai 10 bit slog at 220mbits, it should be even better and I still can't understand the positioning of the C300.
Click to expand...

I imagine the positioning of the C300 comes down to the onboard broadcast standard codec, which is missing from every other camera in the test.
 
Also, Alan Roberts completed his test of the C300. His conclusion:

This camera performs well. Resolution is very well maintained and is refreshingly alias-free, far more sothan other CMOS single-sensor cameras. Detail controls work well, and the factory settings are generallygood. Noise levels are similar to those of 2⁄3” cameras with 3 sensors, as is sensitivity. Dynamic range isunusually high, at least 12 stops. Operating the camera at significantly higher gain produces more noise, butnot dramatically so. The specified noise level of -54dB is achievable only at -6dB gain and with noisereduction switched on. Noise distribution is non-uniform, which gives the pictures a more film-likeappearance.
 
Kit Hannah said:
What? No! We have to convince everyone that our camera is the best and that the test was flawed, and if Philip did a proper test, you would see! Why would you ever want to use lighting for video? :)
Click to expand...

People have a tendency to value things more highly because we own them. It's called the endowment effect and it intensifies the longer we have used said equipment.
 
simonpwood said:
Also, Alan Roberts completed his test of the C300. His conclusion:
Noise levels are similar to those of 2⁄3” cameras with 3 sensors, as is sensitivity.
Click to expand...

Alan Roberts tends to say this about every s35mm sensor camera he has tested. To me it makes no sense. These cameras (F3, FS100, C300) are significantly more sensitive than 2/3" sensor ones.
 
You must log in or register to reply here.
Back
Top