GH3 low light test/disappointed/thoughts

F

funkytwig

Active member
Did some low light tests and put at https://vimeo.com/56486521. I filmed around 20 seconds at 800, 1600, 3200 and 6400. Bad news is there is noise even at 800 (which seems worse than Canon 550D) but the noise is fairly low at 1600/3200. It does increase a bit at 6400 and this may be usable.

Very disappointed, I was hoping for more;(.

What do people think?

Ben
 
First, list what settings you used.

Such as color profile and settings
then, what codec did you shoot on? Did you shoot this in ALL-I thinking you would get better quality?
 
dustylense said:
First, list what settings you used.

Such as color profile and settings
then, what codec did you shoot on? Did you shoot this in ALL-I thinking you would get better quality?
Click to expand...

MOV FHD25p, Standard profile. Will do test again with All-Intra. Can see how All-Intra could help but is profile likely to affect image noise?

Ben
 
One: I don't see a noise problem, oh maybe a little in underexposed areas on the wood floor. But 6400 looks good to me.

Two: Many who see a noise problem may need to check their monitor settings. I mean the actual settings available from controls on the monitor. Most come out of the box with Sharpness, Contrast, and Brightness set way too high. There is a fine "balance" between getting maximum sharpness on the screen but without "accentuating" minor color artifacting.

I had one workstation that showed horrible noise on almost any video, the other two would show the same video just fine. When I thought to check the monitor settings Sharpness was at about 85%, Contrast almost as high, and Brightness set at about 80%. I pulled everything back to midscale and "tweaked" from there and use that workstation for some light video editing and some PhotoShop work.

I'm a sharpness nut, that's one reason I switched over to GH2s from Canon, but oversharpening on the monitor will create the appearance of noise where there is little or none.
 
Bruce Foreman said:
One: I don't see a noise problem, oh maybe a little in underexposed areas on the wood floor. But 6400 looks good to me.

...

I'm a sharpness nut, that's one reason I switched over to GH2s from Canon, but oversharpening on the monitor will create the appearance of noise where there is little or none.
Click to expand...

I think vimeo has smoothed things out a little but thanks for the advice about monitor settings.

What settings max ISO would you use?

Does the codec setting make a lot of difference in terms of noise?

Ben
 
I find the noise on the GH3 to be less than that on the GH2 in low light, it certainly looks more pleasing on skin tones, it's a small difference but still an improvement. If you want to film in low light there's plenty of fast glass to choose from, increasing at a lower price point, which will mean you won't necessarily have to use higher ISOs anyway. Add to that there are a number of noise removal options in post, so in practice, is it that much of an issue?
 
funkytwig said:
Does the codec setting make a lot of difference in terms of noise?

Ben
Click to expand...

The 72mb All I will have more noise than the 50mb IPB mode. 72mb is a little low bitrate for an all-I codec resulting in more noise. At least this has been my observations from personal experience with the GH3.
 
RatLabProductions said:
The 72mb All I will have more noise than the 50mb IPB mode. 72mb is a little low bitrate for an all-I codec resulting in more noise. At least this has been my observations from personal experience with the GH3.
Click to expand...

Interesting, so you are basically saying Panasonic have not really allowed a higher enough bit-rate for All-Intra so I guess it should be avoided. Have you any insight on if AVCHD or MP4 show more of less noise?

Also would the Non All-Intra MOV 25P be any more noisy than the 50P?

Regards,
Ben
 
funkytwig said:
Interesting, so you are basically saying Panasonic have not really allowed a higher enough bit-rate for All-Intra so I guess it should be avoided. Have you any insight on if AVCHD or MP4 show more of less noise?

Also would the Non All-Intra MOV 25P be any more noisy than the 50P?

Regards,
Ben
Click to expand...

Just look at the GH2 which had to have bit rates boosted to 150mb/s in order to get great quality iframe footage. Panasonic would never do this as in the end the GH series is a consumer product and a high bitrate would require only the highest end 95mb cards.

The All-I mode on the 5D3 suffers similar results and most people go with the IPB mode.

I like the 50mb IPB MOV mode best. I think it gives the best quality results. I rarely shoot other frame rates than 24p and 60p to be slowed to 24p so I can't speak to 50p (I'm NTSC land)
 
RatLabProductions said:
Just look at the GH2 which had to have bit rates boosted to 150mb/s in order to get great quality iframe footage. Panasonic would never do this as in the end the GH series is a consumer product and a high bitrate would require only the highest end 95mb cards.

The All-I mode on the 5D3 suffers similar results and most people go with the IPB mode.

I like the 50mb IPB MOV mode best. I think it gives the best quality results. I rarely shoot other frame rates than 24p and 60p to be slowed to 24p so I can't speak to 50p (I'm NTSC land)
Click to expand...

Is that 150MB/s for 1080/30p?

In terms of recording formats is there much difference in quality between MOV and AVCHD (Sony use AVCHD for broadcast so I guess it is good).

Ben
 
funkytwig said:
Is that 150MB/s for 1080/30p?

In terms of recording formats is there much difference in quality between MOV and AVCHD (Sony use AVCHD for broadcast so I guess it is good).

Ben
Click to expand...

150 Mbps (megabits per second) and greater all Intra hacks for the GH2 are usually for 1080p @ 24 fps. (24 fps is the best image quality for that camera).

MOV is just another container that can be used for the MPEG-4 H.264 codec. AVCHD is also a container (designed by Panasonic and Sony) for that same codec. However, some non linear editors and slower computers have a rough time real-time editing H.264 because it's so computer intensive and really a final delivery codec, not an intermediate codec like MOV or ProRes or Cineform, etc. Shooting in the 50 Mbps MOV mode of the GH3 can save you from having to transcode your footage before dropping it into your NLE's timeline.

Tests seem to show that the currently unhacked GH3 looks best at that 50 Mbps long-GOP MOV mode. We're anxiously waiting for a hack for the GH3 since its compression software is a bit more advanced than the GH2, so even a higher bitrate/all-Intra hack could potentially look better than a super-charged GH2... along with the ability to do 1080p @ 60 fps.

Nothing can be done about the increased moire and aliasing due to its use of a Sony sensor with worse pixel binning than the GH2's in-house Panasonic multi-aspect sensor.
 
No one asked what lens was used?
Low light isn´t just a matter of how much practicals are in the scene, but how much light is allowed to enter the camera.
Putting "disappointed" as a conclusion in the title suggests at least some thorough understanding of approaching low-lit scenes, otherwise "confused" "need help" or similar tags seem more adequate.
So what lens and at what F-stop (or T) was the test shot on?
 
Last edited:
LyleGorch said:
Can you provide a verifiable source for this nugget of information? . . .you know, the Sony part?
Click to expand...

They were talking about it on the Personal View website... you know, the guys who hacked and performed their "magic" on the GHx line of cameras. It's been studied and the consensus seems to indicate that Panasonic farmed out the sensor to Sony rather than using their own in-house multi-aspect sensor as was the case in the GH2 (which has little visible aliasing or moire due to a better pixel binning process). When asked about the GH3 using a Sony image sensor, Panasonic basically said "no comment" rather than emphatically stating it was a Panasonic chip.
 
JohnTollwannabe said:
They were talking about it on the Personal View website... you know, the guys who hacked and performed their "magic" on the GHx line of cameras. It's been studied and the consensus seems to indicate that Panasonic farmed out the sensor to Sony rather than using their own in-house multi-aspect sensor as was the case in the GH2 (which has little visible aliasing or moire due to a better pixel binning process). When asked about the GH3 using a Sony image sensor, Panasonic basically said "no comment" rather than emphatically stating it was a Panasonic chip.
Click to expand...

As someone who spends way too much time over at PV every stinking day I think I am safe in suggesting that there is not, or will there ever be anything resembling "consensus" about anything. The Sony sensor thing remains a rumor.
 
You must log in or register to reply here.
Back
Top