GH5 really weird noise when doing VFR on GH5s !!!

Bake the image in when shooting over 60p. Also nail your exposure and lighting with the lowest ISO you can use. It will help keep it cleaner but wont be anywhere as clean as 60P.
 
okay...so to you my body wouldn't be faulty ?

I don't know if I was really expecting too much, it's just that....for exemple before I bought it I looked at this https://vimeo.com/211425610
and I really don't see anything like what I'm getting with mine. Even with NR in resolve and grading, it looks way cleaner....

Hi coutard,

I also think there is something wrong with my gh5s and before seeing your post, was thinking I was the only one with a gh5s that sucks in lowlight.
Its not just about the noise, in my footage it looks like horizontal noise stripes with a magenta cast, and if I try the 2500 native iso with 1080p vfr @ 200, its completely unusable.
Again, I have the same feelling after seeing a lot of good footage in lowlight, even with higher isos. In a cave with bats - https://www.youtube.com/watch?v=V1NnizowA2Q
I would like to know if you found a solution or if you returned your camera?
 
It is strange. The 180fps at 6400 iso looked really really clean in that video.

We don't know what picture profile he was shooting in though.

Mr.Edd, do you get those issues in all picture profiles?
 
I live about 5 minutes from Denver Zoo, have both GH5 and GH5s if you want to do a comparison. I am actually confused if you are citing the example in the bat cave as an example of good or an example of bad. If there was horizontal banding it didn't jump out at me. Yes, I thought the 180fps at 6400 iso was clean as well, agree with sgttom. I also have Sony F55 cinema cam. Hate to admit this, but low light of the GH5s is leagues better, so I'm not sure what other cams are really cleaner than GH5s in low light, I mean I can name a few, Sony full frame mirrorless A7SII A7RII A7RIII I would assume, but at equal depth of field maybe not even those?
 
I live about 5 minutes from Denver Zoo, have both GH5 and GH5s if you want to do a comparison. I am actually confused if you are citing the example in the bat cave as an example of good or an example of bad. If there was horizontal banding it didn't jump out at me. Yes, I thought the 180fps at 6400 iso was clean as well, agree with sgttom. I also have Sony F55 cinema cam. Hate to admit this, but low light of the GH5s is leagues better, so I'm not sure what other cams are really cleaner than GH5s in low light, I mean I can name a few, Sony full frame mirrorless A7SII A7RII A7RIII I would assume, but at equal depth of field maybe not even those?

I was citing the link as a good example since I can´t get anything like that.
I´ve also check the gh5s Manual page 176 and it seems that when VFR and V-LOG L is selected, the ISO selection (LOW/HIGHT) changes.
https://cdn.static-bl.com/images/manual/GH5S-Manual.pdf
 
sgttom,

Yes, in all picture profiles but the V-log is more evidente since it doesn´t have contrast and your clearly see what is happening in the shadows
 
I would want to check that Extended ISO is "OFF." The manual notes that "noise and horizontal stripes appear more often." It doesn't elaborate.

I have always kept extended ISO off because someone posted that it adversely affects DR, although I found no mention of this in the manual.

I usually feel that Natural is a cleaner profile for low light than V-Log/L but I did make an exception during a July 4th fireworks show I used V-Log/L to protect the highlights and preserve color. That turned out very well because when the fireworks are exploding it is actually very bright, you're exposing for that, and in between when it is dark it is also very clean because you've recorded the noise below the noise floor by exposing for the brighter part of the show.
 
Last edited:
You mean, the "extended ISO" function would affect DR even when not really using the extended ISOs ? Like, shooting at 800 ?
 
GH5s Horizontal Noise/Banding/Striping Update

GH5s Horizontal Noise/Banding/Striping Update

I made 3 exposures of a white wall in a very dim room. All were shot at F/1.2 1/60th:

1.) HLG 320 ISO (Dual Gain = Low)
2.) Nat 160 ISO (Dual Gain = Low)
3.) Nat 800 ISO (Dual Gain = High)

I put them all on the same timeline using Resolve Color Managed Space.

Image #2 was darker because the gain was less. In post I pushed #2 to match #1. Image #3 was brighter because the gain was more. I decreased #3 to match #2. Now, all 3 images were matched to the same brightness. On the waveform monitor, the image level for the matched images was about IRE 3%, or 10 bit code value 32.

After the post adjustments, images #1 and #2 looked to be the same amount of noise which was very evident as horizontal striping. Image #3 was the cleanest, no horizontal striping was evident. Had there been highlights in the images, #3 would have clipped 1-1/3 stops sooner than #2.

Next I repeated the experiment with the lens cap on. I could scarcely if at all, see any noise in any of the images so it was not possible nor necessary to match the brightness; there wasn't any. It took the brightness of the (very dim) white wall in the earlier experiment to shine through the transparency of the noise so that you could see it. I could not distiguish a difference in noise levels with the waveform monitor either, just too low and at the very bottom of the scale.

I could however, clearly measure the noise ceiling with the histogram, which was about equal for all 3 color channels. Image #1 had the lowest noise ceiling at 2% IRE. #2 was next at 2.5%. The highest measured noise ceiling was #3 at 3%, the setting which earlier manifested no striping; the least visible noise, measured the highest; a bit surprising.

My Conclusions:

1.) Photographic detail exists below the noise ceiling. If there is detail below the ceiling and above the recording floor, the noise will be visible. The noise ceiling appears to be about 2-3%.

2.) As the photographic details rise to the level of the noise ceiling and beyond, the noise will fade out.

3.) It's desirable that the signal to noise ratio be higher. It may be very high already. Visible noise seems to be a consequence of great low light sensitivity in combination with high dynamic range which makes it possible to record details and see noise into the darker areas of an image that are below the noise ceiling simultaneously with mid-tones and highlights in the noise-free areas above.

4.) The GH5s also has great low light sensivity, making the possibility of more shots at levels below the noise ceiling all the more likely.

5.) I'm not offering a solution. The visible noise, particularly characterized as horizontal banding stripes is very ugly, ruinous. These are just my observations and opinions about it.
 
I just did a full shoot today with Extended ISO off and I have to say the difference between shadow noise on the GH5S in V-Log IS NIGHT AND DAY!! Even when shooting in standard ISOs like 400/800/2500 I'm going to do some proper 'on / off' tests tomorrow, but I'm so much happier with my rushes from the GH5S today!
 
thanks for your tests and insights.
I'll do some testing without extended ISO as well as I have a shoot planned in 10 days !
 
So there is definitely a problem with VFR on the GH5s - Anything above 60FPS creates extremely strong horizontal stripes / interference in shadow areas.

To test - if you set your camera to 1920 x 1080 in any picture profile (but V-Log will show it more) and set to VFR, then set VFR to 60FPS. Hit the record button and watch the LCD. The image on the LCD doesn't change. Now, change settings on VFR to 72 FPS, and hit the record button, the screen will go darker and more 'green' and you can see the horizontal stripes in all shadow areas suddenly appear. This is the case all the way up to 240FPS. This is a bug.

All shots will show seriously nasty horizontal stripes in the shadows on anything above 60fps in VFR on the GH5S.
 
Last edited:
Back
Top