Gemini not capturing blacks. Help.

[Jason beaumont]

Damn...she hot!
Who is she? :001_tt1:

IG is @camille.sofia super cool too!

 

[Sabovic Adis]

IG is @camille.sofia super cool too!

Thanks, Jason!

 

[Les Hillis]

Looks good Jason (even if Low-Light mode would have been better for that shot), glad it worked out.

 

[Phil Holland]

I've only worked with Gemini a handful of times, but my general way of working was using the Low Light Mode most of the time and Standard for Daylight or extremely well lit situations, like tabletop or talking head.

 

[Hrvoje Simic]

Can someone tell me what I did wrong, or has this happened to them. My camera just started doing this.

Of the four shots in the link:

First is underexposed + 8:1 bites off a chunk of what is left = unusable
Second is okay
Third is okay
Fourth is underexposed = unusable

There is nothing wrong with the camera.
You just didn't capture sufficient light to quantize into data. Your signal chain for monitoring may have been misleading so make sure you check that and stick with what is reliable and tested, and on that base test all the preview transformations in specific shooting conditions.






Your images with proper transformation applied to get a sense of realistic exposure levels:

 

[Jens Jakob Thorsen]

You could hire a DOP?

 

[Mike P.]

May I ask why you didn't you use LL mode when you're more or less going natural/ambient light? The beauty of LL mode isn't just that it "gives you clean ISO2500", but that it actually shifts the dynamic range to open up more stops in the low-end (it actually empties out the shadow goalpost.)

And why only 4k WS? 5k (or even 4.5k) WS will net you substantially more image, which only serves to look better/cleaner when downscaled to 4k or 1080p. Also the shot of the girl (the most useable in my opinion) was still f4!!? You could've shot wide open, ETTR, and brought the ISO down in post for a fatter histogram/image.

Try looking at your left (shadow) goalpost. If you can, keep it empty, but if its not, toggle to gio/video/raw view to see what's actually purple and that'll tell you what is actually black. In terms of underexposure, again, use the video view for false colour (Caucasian skin tones should ideally be green/pink, unless it's actually in shadow). Beyond that, you're dealing with underexposure, so tread lightly (or do more tests until you know what you can recover in post).

EDIT: I realize others may have already said this, I just read the OP. So ignore if its redundant.

 

[rand thompson]

Here are the only two recoverable out of the 4.

 

[Jason beaumont]

I've only worked with Gemini a handful of times, but my general way of working was using the Low Light Mode most of the time and Standard for Daylight or extremely well lit situations, like tabletop or talking head.

Yeah I was doing it the opposite, I didn't use low light mode very much, but since this thread and experience im incorporating it much more.

=Hrvoje Simic;1861531]Of the four shots in the link:

First is underexposed + 8:1 bites off a chunk of what is left = unusable
Second is okay
Third is okay
Fourth is underexposed = unusable

There is nothing wrong with the camera.
You just didn't capture sufficient light to quantize into data. Your signal chain for monitoring may have been misleading so make sure you check that and stick with what is reliable and tested, and on that base test all the preview transformations in specific shooting conditions.

Yeah I see that now, thank you. I really wanted that 4th shot, I might just have to reshoot it.

You could hire a DOP?

Nah , give me a fish feed me for a day, teach me to fish, and feed me for a lifetime. Besides this was just a small passion project. It was just me and a friend and another friend holding my boom. For lighting we had 1 sky panel, and a arri 650 lol. Its not enough for a man to climb mount Everest. He has to do it with the least amount of tools possible! Nah but really, I just wanted to see how well I could bring my little script to life with limited resources.

May I ask why you didn't you use LL mode when you're more or less going natural/ambient light? The beauty of LL mode isn't just that it "gives you clean ISO2500", but that it actually shifts the dynamic range to open up more stops in the low-end (it actually empties out the shadow goalpost.)

And why only 4k WS? 5k (or even 4.5k) WS will net you substantially more image, which only serves to look better/cleaner when downscaled to 4k or 1080p. Also the shot of the girl (the most useable in my opinion) was still f4!!? You could've shot wide open, ETTR, and brought the ISO down in post for a fatter histogram/image.

Try looking at your left (shadow) goalpost. If you can, keep it empty, but if its not, toggle to gio/video/raw view to see what's actually purple and that'll tell you what is actually black. In terms of underexposure, again, use the video view for false colour (Caucasian skin tones should ideally be green/pink, unless it's actually in shadow). Beyond that, you're dealing with underexposure, so tread lightly (or do more tests until you know what you can recover in post).

EDIT: I realize others may have already said this, I just read the OP. So ignore if its redundant.

My friend who I was shooting with didn't understand low light mode. He felt like because we had lights we shouldn't use it, which I told him is not the case, but for the sake of not wasting time arguing, and since he was operating I let him have his way (never again lol). 4k WS was because my Cooke s4i mini vignette at 5k on the gemini. atleast one of them does. I don't know, I hardly ever shoot lenses wide open. I am def using my tools more often thanks.

 

[Les Hillis]

Imo those two underexposed shots are unuseable because they don't intercut with the rest of the scene/footage, but you could still shoot that dark (like in Hrvoje's grades) if the surrounding shots matched and if the individual shots didn't need to convey more information to the audience than what they still manage to do.

It's definitely a particular kind of look that wouldn't suit most kinds of scenes, but with the right subject matter they don't look so dark to me that you couldn't still figure out what's going on.

Just saying/asking.

 

[Mike P.]

Nah , give me a fish feed me for a day, teach me to fish, and feed me for a lifetime. Besides this was just a small passion project. It was just me and a friend and another friend holding my boom. For lighting we had 1 sky panel, and a arri 650 lol. Its not enough for a man to climb mount Everest. He has to do it with the least amount of tools possible! Nah but really, I just wanted to see how well I could bring my little script to life with limited resources.


My friend who I was shooting with didn't understand low light mode. He felt like because we had lights we shouldn't use it, which I told him is not the case, but for the sake of not wasting time arguing, and since he was operating I let him have his way (never again lol). 4k WS was because my Cooke s4i mini vignette at 5k on the gemini. atleast one of them does. I don't know, I hardly ever shoot lenses wide open. I am def using my tools more often thanks.

I hear ya; it's a dilemma when you don't quiet know for sure (and definitely gives you a real-world fishing lesson). And yeah, you'll be aces moving forward.

Makes sense (re: s4i/coverage)... a shame cause they usually cover when in WS.

 

[Hrvoje Simic]

Imo those two underexposed shots are unuseable because they don't intercut with the rest of the scene/footage, but you could still shoot that dark (like in Hrvoje's grades) if the surrounding shots matched and if the individual shots didn't need to convey more information to the audience than what they still manage to do.

Those are not grades.

Those are transformations showing accurate exposure for that sensor for a thick negative and healthy shadows. Which is especially important when exposing to dark scenes.

If the surrounding shots matched those uderexposed two, they would also be unusable. Gemini is capable of outstanding imagery but as with every camera, that necessitates proper exposure. Unrealistic image preview propagates overly optimistic expectations which lead to improper one.

 

[David Mullen ASC]

In theory he was right, if you have enough lights, you don’t need to use high ISOs, especially for a scene that is supposed to be dark. After all, they were doing moody night scenes in movies back when film stocks were incredibly slow.

One tip: if you are shooting in a very dimly-lit room at high ISO’s, don’t judge the exposure by the monitor because the monitor becomes the brightest object in the room, it’s like looking into a lit lampshade. You will tend to underexpose too much to make the monitor image feel dark enough.

 

[David J. Buchanan]

Yeah, I think everyone touched on the problem. it's the lighting, not the camera.

It could be useful to shoot some tests with it before dropping rental dollars on a sky panel... unless you own it.

 

[Jason beaumont]

In theory he was right, if you have enough lights, you don’t need to use high ISOs, especially for a scene that is supposed to be dark. After all, they were doing moody night scenes in movies back when film stocks were incredibly slow.

One tip: if you are shooting in a very dimly-lit room at high ISO’s, don’t judge the exposure by the monitor because the monitor becomes the brightest object in the room, it’s like looking into a lit lampshade. You will tend to underexpose too much to make the monitor image feel dark enough.

Thank you!

Yeah, I think everyone touched on the problem. it's the lighting, not the camera.

It could be useful to shoot some tests with it before dropping rental dollars on a sky panel... unless you own it.

Own it. yeah I still think the came jumps in noise really easy in standard mode, but atleast I understand it, so I can avoid (or not avoid it) , as I choose. thank you everyone