PDA

View Full Version : Exposure questions



drgeneric
06-24-2013, 10:43 AM
Hi There,

As far as I understand, exposing for 100% Zebra should give us the most information in the blacks and causes no clipping at all, so it seems to be the perfect exposure. Is this the case? I don't think so, but I can't figure how. And whats odd:

- When exposing for 100% zebra, nothing should be clipped, as far as I understand. However, when connecting an external SDI-Monitor with scopes, the waveform indeed clips. How is that possible?
- No matter if I'm shooting Film or Video, ProRes or RAW: When I expose for 100% Zebra, it's drastically overexposed for my eye, and I can't really judge the shot. In my understanding, all of the information could be brought back in post, but it's really difficult to judge your shot when you're looking at a camera screen that's mostly white.

Could you help me clearing these matters?

Thank's a lot,
Remo

mrbrycel
06-24-2013, 10:51 AM
When you say "expose for 100% zebra", you mean setting your zebras to show 100% IRE, and then setting your exposure to where you see no zebras at all right? I've always used this method and haven't had a problem.

Danny1280
06-24-2013, 11:07 AM
And just to piggy-back another question: Should there be absolutely no zebras showing with this method, or is it okay if a few very small specular highlights are showing zebras? I'm going to assume there should be no zebras...?

mrbrycel
06-24-2013, 11:22 AM
And just to piggy-back another question: Should there be absolutely no zebras showing with this method, or is it okay if a few very small specular highlights are showing zebras? I'm going to assume there should be no zebras...?

In most cases with zebras at 100% I've always exposed to where no zebras show to guarantee all detail, but I know there are certain situations where you'd want to let certain highlights blow out, and you won't want that detail anyway.

zwarte_kat
06-24-2013, 11:42 AM
Here is what I have found:

With Raw:

1. Nothing should ideally be clipped according to the BMC screen, if you output to a monitor in then it kind of depends on the monitor, it might apply it's own LUT or whatever depending on what you use.
2. Nothing should ideally be below about 15% in your waveforms, because these dark areas will become noisy. Depending on what you find acceptable, this could be 20% or 10% or another value.

You will often find that you have to balance between 1 and 2. An easy example is at night by blowing out lamps and other light sources. If you don't let these clip, most of the rest of your footage will probably be dark and noisy, even with the BMC's high dynamic range. Sometimes during the day it can be a bit harder to decide. You don't want to blow out your sky, but your subject is in a dark shadow, even the BMC can struggle here. ideally you adjust your lighting with a ref or strong lights.

Finding solutions to satisfy point 1 and 2, and choosing the right balance between them, is the key to healthy exposures in the BMC IMO.

That's the relatively easy part, compared with the actual art of lighting :)

rze
06-24-2013, 12:19 PM
The reason that an external monitor is different is because of your display output settings. The in camera zebras are always for the RAW data. It shows when the sensor is clipped and not when lut adjustments are clipped. Video mode is a lut on the log signal to make the image look normal in a rec709 space. That requires pushing the highlights up so if your RAW signal was close to clipping, the video mode output will surely be clipped going out the sdi if you are set to video mode. If you set to film mode for the video output the signal going through the sdi should match the zebras.

This exposing to the right just under clipping is not good advice as far as I'm concerned. You can over expose by one or two stops, but if you go over 4 stops, how will you see your image? You won't be getting more noticeable cleanliness in the darks and your tonality will be effected. Just expose normally and you will get great results and not look bad when you send your footage to the editor.

zwarte_kat
06-24-2013, 12:40 PM
I believe the BMC SDI output is always film, and cannot be set to video. It is something that a lot of people had wished to be different.
Possibly you can get a video output when RECORDING to prores video mode, but I never do this so I am not sure.

In my case I adjust the monitor's contrast, brightness, and chroma, so that I get an image that is easier to focus. I keep in mind that this is not at all the image I am recording in terms of exposure. I rely on my waveform and scopes for that.
It is a more technical approach, but I have found that it is more error proof in terms of exposure. Especially at dark scenes, even the BMC's screen can be very deceiving.

John Brawley
06-24-2013, 04:22 PM
Hi There,

As far as I understand, exposing for 100% Zebra should give us the most information in the blacks and causes no clipping at all, so it seems to be the perfect exposure. Is this the case? I don't think so, but I can't figure how. And whats odd:

- When exposing for 100% zebra, nothing should be clipped, as far as I understand. However, when connecting an external SDI-Monitor with scopes, the waveform indeed clips. How is that possible?
- No matter if I'm shooting Film or Video, ProRes or RAW: When I expose for 100% Zebra, it's drastically overexposed for my eye, and I can't really judge the shot. In my understanding, all of the information could be brought back in post, but it's really difficult to judge your shot when you're looking at a camera screen that's mostly white.

Could you help me clearing these matters?

Thank's a lot,
Remo

100% zebra shows you what *IS* clipped.

If you have Zebra's then you have clipping.

You use the 100% Zebra to indicate where the clipping is and then YOU as the DP decide where the clipping point should be. Ideally, you shouldn't have clipping on important visual information.

jb

Frank Glencairn
06-24-2013, 04:36 PM
What John said.

Also mind, that even if all zebra is just away (a hair below clipping), single channels still can clip, even if you don't see zebras..
Most of the time, this is fixed with the "recover highlight" check-box in Resolve though.

Maybe I start seeing things, but I'm under the impression, that with the last update, the was the "iris" button works was changed.
Before it was like, when you hit the iris button, your zebras just want away. Now it seems to stop down a bit more, to prevent single channels from clipping.

You can try that. ND or stop down, till the zebra disappears, than hit iris, and you gonna see that it stops down a bit more now.

Frank

PaulDelVecchio
06-24-2013, 06:08 PM
It's my experience that when your levels are below around 50%, you get noise in the image. On my SmallHD monitor when using False Color, if something in the frame is not ORANGE which is the skin tone level around 50%, there is noise in the image that may or may not be acceptable. How acceptable it is depends on what you're shooting. If you're shooting a beauty video, it might not be acceptable. If you're shooting a film, it may be, depending on your tastes.

I'm still using the firmware from Feb 2013, as I have not updated my camera. It may have changed since then, but I doubt it. Ultimately, you have to test to see what is acceptable for you and your project.

I've had in depth discussions with people who don't like the noise, and think that 50% level is way too high. Some have asked me what the point of having 13 stops of DR is if everything below 40-50% on the scopes is noisy. Some don't mind the noise. Depends on your taste and you have to ask yourself if that's acceptable or not.

Kholi
06-24-2013, 06:26 PM
Definitely to taste, but if fifty percent is where one's tolerances sit this is far from the ideal camera.

So the same for the most revered digital image right now, Alexa, which is only SLIGHTLY cleaner raw-to-raw image.

FS100, C300, EPIC, F55 are cameras that people may want to look at if the noise is too much. I've noticed that each of those are fairly clean in most circumstances, the Epic @ 5K can be noisy yet when mastered at 2K with appropriate NR it pretty much vanishes. And, of course, FS100, C300 for those that don't want to remove noise in post. Cameras that do the work for you. Too many choices.

On Zebras: wonder if any of the BMD Engineers can confirm that 100 percent does not represent when any one channel clips. And, technically, if "highlight recovery" restores the clipped area, then it was never clipped to begin with? I just stick with the 100 percent for RAW and ProRes, and try to keep anything important from riding lower than 80 IRE.

PaulDelVecchio
06-24-2013, 06:47 PM
Definitely to taste, but if fifty percent is where one's tolerances sit this is far from the ideal camera.

So the same for the most revered digital image right now, Alexa, which is only SLIGHTLY cleaner raw-to-raw image.

FS100, C300, EPIC, F55 are cameras that people may want to look at if the noise is too much. I've noticed that each of those are fairly clean in most circumstances, the Epic @ 5K can be noisy yet when mastered at 2K with appropriate NR it pretty much vanishes. And, of course, FS100, C300 for those that don't want to remove noise in post. Cameras that do the work for you. Too many choices.

On Zebras: wonder if any of the BMD Engineers can confirm that 100 percent does not represent when any one channel clips. And, technically, if "highlight recovery" restores the clipped area, then it was never clipped to begin with? I just stick with the 100 percent for RAW and ProRes, and try to keep anything important from riding lower than 80 IRE.

Yeah we've talked plenty about this before. I've had a director tell me he thought the image was noisy even after we pumped a lot of light into the scene and he was getting frustrated. He even questioned my knowledge of the camera saying that I must have a setting wrong. I told him to take a look at the settings because there barely are any. It just needs a lot of light. All of this was discussed before the project, but I guess it still took him by surprise. He kept comparing it to his Nikon DSLR though. Eek.

It really depends on how you rate the camera. For me, I set the ISO to 200 and shoot like that because that allows me to see what I'm shooting a bit better than keeping it at 800, which will look very "white." Someone correct me if I'm wrong, but if I'm shooting RAW, it still is a fact that shooting at ISO 200 does not decrease your dynamic range, it just displays it differently, correct? I know there was some back and forth about different curves being applied and whether or not you lost dynamic range, so I'm just double-checking my facts here. Either way, what I've been getting with the camera set to 200 is amazing and without a LUT, that's the best option to actually see what you're shooting.

Kholi
06-24-2013, 06:53 PM
I don't think 200's been confirmed but to be honest I feel like there IS a difference when shooting 200, and so even when I'm doing a sort of "Viewing lut" approach by changing ISO, I stay at 400 or 800. But, might be me playing tricks on m'self. 1600 ISO is definitely changed.

By using the TV Logic with a fake look on it, I've been able to judge what's too far under and what's not, and have been getting cleaner images since.

His Nikon DSLR is definitely going to be cleaner, given that it's reducing noise (and IQ in general) when going to a compressed image. I guess I'm being redundant, sorta feels like anyone afraid of noise to that degree would never shoot Alexa or F65, F35, or Film for that matter. xD

PaulDelVecchio
06-24-2013, 07:33 PM
His Nikon DSLR is definitely going to be cleaner, given that it's reducing noise (and IQ in general) when going to a compressed image. I guess I'm being redundant, sorta feels like anyone afraid of noise to that degree would never shoot Alexa or F65, F35, or Film for that matter. xD

Definitely. I keep referencing The Village and 28 Weeks Later. I like the way they're shot. Whether you like the films or not, check out how much noise is in the image. It's WAY more than what I normally get out of the BMCC, especially in 28 Weeks Later. I believe 28 Weeks Later was shot on Super 16 for the majority, with VFX plates shot on 35.

I really wish there was a way to view the image at 200 without actually changing the setting. I know it's been talked about a lot, but it would be awesome. If 200 truly doesn't affect the image, then I'm fine with just setting it to 200.

rze
06-24-2013, 09:32 PM
Shooting prioritizing your clipping point rather than middle grey is weird to me. Sacrificing 2 stops of light to just see your image seems like a little much. Noise reduction would be a far simpler if thats a subjective problem, especially since the BMC is usually too sharp anyway. I know the bmc is strong but I would rather clip than have the contrast ratios on the talent be ambiguous.

Kholi
06-24-2013, 09:59 PM
Shooting prioritizing your clipping point rather than middle grey is weird to me.

Had a discussion the other day with a DP, we got into this. It's a fear of clipping highlights that's been blown out of proportion (not meaning this in a bad way or as an attack on anyone, I do it as well) because.. well, blown highlights tend to look VERY ugly in regards to digital images. RED, 5D, C300, I find that if you do that you can never get things back to where you truly want them. The final image looks dark, drab, worn out as if someone's been beating on it to get it where it's at.

I guess I'm saying it just doesn't "lift" well.

For whatever reason, I do find Blackmagic footage much more forgiving in regards to color and contrast if I've gone and underexposed the image to guard highlights, but there is a noise penalty. Noise doesn't bother me until it's really heavy (subjective word there), so it's never really a big deal. But I'm okay with underexposing some stuff with this camera.

Mostly, I'll try to land what I believe is an optimal exposure for a primary subject, though, which is mainly a human being in my case. If that means losing a window then that window should've been dealt with externally, not in camera. Those were earlier lessons learned.

After attending LA Film Festival and catching CONCUSSION (Shot on RED) and The Conjuring. I don't know what the latter was shot on, but it being a Horror picture one would think it would have been "dark". It was bright, but still a "dark" movie, like it was exposed if that makes sense. CONCUSSION could've easily been mistaken for a horror if cut the trailer the right way, and it wasn't a horror... I'll tell you that.

I'm really happy with about ninety percent of what BV1 offers, being able to have a nicely exposed product at the end of it all's on the list of positives for me, even if the method that I choose isn't THE most ideal.

PaulDelVecchio
06-24-2013, 10:52 PM
After attending LA Film Festival and catching CONCUSSION (Shot on RED) and The Conjuring. I don't know what the latter was shot on, but it being a Horror picture one would think it would have been "dark". It was bright, but still a "dark" movie, like it was exposed if that makes sense.

The Conjuring was shot on the Alexa. I've been looking at a lot of Alexa footage lately and comparing the BMCC to it (which is unfair because of the cost). Anyway, with the way cameras are these days, I'd love to know how certain scenes were lit because I'm wondering if they were indeed lit with a match, candle, whatever, or if there was another source used. It seems to be popular with the lower light cameras that they're lighting scenes with actual sources rather than "faking it."

3532353335343535

I'm surprised to hear The Conjuring was "exposed." It seems it would be a "darker lit" film.

Kholi
06-24-2013, 10:58 PM
The Conjuring was shot on the Alexa. I've been looking at a lot of Alexa footage lately and comparing the BMCC to it (which is unfair because of the cost). Anyway, with the way cameras are these days, I'd love to know how certain scenes were lit because I'm wondering if they were indeed lit with a match, candle, whatever, or if there was another source used. It seems to be popular with the lower light cameras that they're lighting scenes with actual sources rather than "faking it."

3532353335343535

I'm surprised to hear The Conjuring was "exposed." It seems it would be a "darker lit" film.

Thanks for the info. I had a very strong feeling that it was Alexa, but I'll tell ya... some of those exterior shots were pretty film-like, and had you said film I would've believed it. Definitely surprised me to see how bright the movie was.

SPOILER

There are scenes where it's dark like that but they're mostly leading into the next scare. The rest of the picture, even when we're sitting with the family at tables etc.'s pretty darned nicely exposed... in a way that film critics would approve of, that's what I mean.

SPOILER END

Dark, underexposed digital picture has worn me down, though. It's just a preference, but I'm over it.

CONCUSSION wasn't a horror at all, but check out some of the stills if you can find any. I'm comparing these two because I sawr them both the same day, and really took notice of just how "dim" CONCUSSION's final image turned out.

Oh, and on match lighting: I would doubt that where Alexa is concerned. It's not a lowlight camera, at all. Very noisy. Unless they removed a lot of noise and did shoot match, but it didn't look like match light only to be honest. I could be wrong!

CaptainHook
06-25-2013, 02:06 AM
On my SmallHD monitor when using False Color, if something in the frame is not ORANGE which is the skin tone level around 50%, there is noise in the image that may or may not be acceptable.

I've kinda been finding the same, except i would say for my tastes it's only noisy if i need to lift exposure in any of those areas under 50%. If i'm pulling down the exposure overall, it's generally fine.

I'm using iso 200/400 a lot as 'preview lut' as well, except i'm still some what at a loss with Prores since my monitor can't seem to compensate enough and i tend to shoot prores more than raw. I just shot my 2nd TVC as 2nd unit DP on the BMCC last week for a Dove ad (the first Dove ad to be shot in NZ), and part way through the first day i had to switch from raw to prores for the rest of the shoot as they simply couldn't handle all the data from the BMCC (it really needed it owns dedicated DIT station that could handle 3 simultaneous backups of raw). The Alexa was shooting Prores, so hard to justify to people that the little $3000 cam is costing more in data wrangling than the a cam. With budgets shrinking, i can only imagine this being more common place unless i can work myself up to bigger shoots with much larger budgets (yeah, i would think a nationwide Dove TVC would qualify too but not in this country) or as storage and data management becomes much more cost effective - in which case is the BMCC the right choice for those shoots anyway?

I think i'm gonna need to just pony up and buy my own on-set LUT box (ideally portable that i can stick on my rig) - this camera is cheap for a reason huh. ;)

Frank Glencairn
06-25-2013, 02:22 AM
On Zebras: wonder if any of the BMD Engineers can confirm that 100 percent does not represent when any one channel clips. And, technically, if "highlight recovery" restores the clipped area, then it was never clipped to begin with? I just stick with the 100 percent for RAW and ProRes, and try to keep anything important from riding lower than 80 IRE.

I got the word from Kristian, since I asked him about zebras and single channel clipping.
So yeah, it comes straight from the horses mouth.

Highlight recovery - as far as I know - restores or kind of extracts the missing information from the non clipping channels.
A technique that was done manually by colorists since a long time, it's a convenient check-box now.

Captain, I use a BM HDLink LUT box on set, works pretty well. I don't load a LUT on it though, just tweak the curves in the software to get close enough,
so the agency chicks don't get a heart attack, when they see it in the village. The box velcros nicely on the back of the village monitor.

Is it just me, or did anyone else see the changes how the iris button works?

Frank

Kholi
06-25-2013, 02:58 AM
I've kinda been finding the same, except i would say for my tastes it's only noisy if i need to lift exposure in any of those areas under 50%. If i'm pulling down the exposure overall, it's generally fine.

I'm using iso 200/400 a lot as 'preview lut' as well, except i'm still some what at a loss with Prores since my monitor can't seem to compensate enough and i tend to shoot prores more than raw. I just shot my 2nd TVC as 2nd unit DP on the BMCC last week for a Dove ad (the first Dove ad to be shot in NZ), and part way through the first day i had to switch from raw to prores for the rest of the shoot as they simply couldn't handle all the data from the BMCC (it really needed it owns dedicated DIT station that could handle 3 simultaneous backups of raw). The Alexa was shooting Prores, so hard to justify to people that the little $3000 cam is costing more in data wrangling than the a cam. With budgets shrinking, i can only imagine this being more common place unless i can work myself up to bigger shoots with much larger budgets (yeah, i would think a nationwide Dove TVC would qualify too but not in this country) or as storage and data management becomes much more cost effective - in which case is the BMCC the right choice for those shoots anyway?

I think i'm gonna need to just pony up and buy my own on-set LUT box (ideally portable that i can stick on my rig) - this camera is cheap for a reason huh. ;)

Not surprising, though. At least here, there are VERY few productions shooting ARRI-RAW, most go straight to ProRes because it's more than enough. That's why everyone working in RED's camp, or dedicated to RED, yelled so long and hard about a ProRes module...they lose work to Alexa because of that first, in most cases.

Do you think the ProRes from BV1's that far off from Alexa's ProRes? ALexa's definitely more resolute right off the cuff, and you have to sharpen or shoot RAW to match it... at least from the shorter tests I've seen. What's your take on the two together?

CaptainHook
06-25-2013, 03:26 AM
Do you think the ProRes from BV1's that far off from Alexa's ProRes? ALexa's definitely more resolute right off the cuff, and you have to sharpen or shoot RAW to match it... at least from the shorter tests I've seen. What's your take on the two together?

I've only done 2 shoots with BMCC/Alexa together so far, and the first shoot was BMCC all in RAW which the main DP requested as he believed it would better match the Alexa (based purely on guess work as it wasn't tested by either of us). I only did the rushes on that shoot, and i felt the two matched pretty well. There were colour discrepancies (reds/oranges), but nothing a quick hue vs hue curve didn't solve when i did the dailies.

The Dove ad is still in edit, so i won't see the results of the grading until later this week, but i've requested to sit in during the grade which should be interesting. I will be grading the behind the scenes part of the shoot too which is destined as web content etc, which will be a mixture of the alexa/bmcc and 5D3 shot by another shooter. So that will be my first time really trying to match the two for release so i'll have more of an opinion then. I did the dailies on this shoot too and from quick impressions the alexa 'feels' more "effortless" .. pretty useless description i know but until i spend some time trying to match both that's the best i've got. :P Maybe the only other comment being the alexa appears to have more 'resolution' with the BMCC appearing 'sharper'. I will look at that more closely tho as being a beauty spot i used a glimmerglass, they tried a black promist on the alexa at the shoot but opted to go without it which should be interesting to compare.

Hopefully i can do an actual test comparing the Alexa prores and BMCC raw/prores if i get used again in that combo too.. it should really have happened at camera prep on either shoot so i will suggest we do so.

gmfb
06-25-2013, 03:37 AM
I also doubt they have much experience with the camera since NZ seems quite isolated and many crew members on both shoots had never heard of the camera, let alone seen one.

Wow, are they busy working day and night? I mean, there is the internet....

sorry for OT, but this is weird. glad you got your 2nd job nevertheless!

CaptainHook
06-25-2013, 04:35 AM
Wow, are they busy working day and night? I mean, there is the internet....

Well there's still the "old school" guys that are working 12hour+ days and no, they're not surfing the net when they get home. Most of the younger guys who are on their phones/facebook/twitter/etc inbetween setups and mealtimes were well aware what it was though. ;)