Any P4K ISO400 vs ISO3200 test samples?

The place for questions about shooting with Blackmagic Cameras.
  • Author
  • Message
Offline
User avatar

Dmytro Shijan

  • Posts: 1760
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Any P4K ISO400 vs ISO3200 test samples?

PostSat Oct 20, 2018 6:32 pm

Did anyone have some ISO400 (first ISO range) vs ISO3200(Second ISO range) DNG test samples?
Wonder if underexposed shoot at ISO3200 with recovered shadows in post may have more dynamic range than same shoot at ISO400?
BMMCC/BMMSC Rigs Collection https://bmmccrigs.tumblr.com
My custom made accessories for BMMCC/BMMSC https://lavky.com/radioproektor/
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Any P4K ISO400 vs ISO3200 test samples?

PostSat Oct 20, 2018 8:20 pm

The relationship between stops of dynamic range above and below middle grey are illustrated in a chart in the BMPCC4K camera manual. The maximum overall dynamic range uses the 400 band. ISO 100-200 devotes more stops to the shadows, but significantly less stops to the highlights.

ISO 400 4 stops above, 9.1 stops below
ISO 3200 3.7 above, 8.7 below middle grey

If your scene has important details through the shadows, shooting at ISO 200 or 1250 gives you at least 10 stops below middle grey. Both good choices. ISO 1250 seems to be a popular choice balancing dynamic range and noise for those situations where you need to use a higher ISO.

If your scene has important details through the highlights, shooting either ISO 640-800 or ISO 4000-6400 may be good choices. We haven’t seen much in those ranges to judge a best practice.



Sent from my iPhone using Tapatalk
Rick Lang
Offline
User avatar

Dmytro Shijan

  • Posts: 1760
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Oct 21, 2018 7:53 am

Yes, those are theoretical values. but when shoot RAW all those charts and meanings are useless. I just use native camera ISO and boost or lower expose in post.
I mean that i was looking for high iso and low iso sample exposed same way to compare. I was thinking to test this viewtopic.php?f=21&t=65149 "boost expose and lower gain" method to extend dynamic range from underexposed shoots with high and low iso and compare how far they can go into shadow.

Actually i found some samples but those are not exact what i looking for:
There are different iso samples with red lamp clipping but exposed differently. Also there are bmpcc vs p4k indoors dynamic range samples but also exposed differently. BMPCC is 2 stops underexposed but when you recover shadows noise amount looks similar as normally exposed P4K. https://www.dropbox.com/sh/d6p9k043vnvv ... 4Aw-a?dl=0

Here are also high iso (ISO3200) samples slightly underexposed to preserve extreme highlights. Nose from recovered shadows looks far better than similar low iso (ISO400) samples
https://www.dropbox.com/s/87wr8c8tmjzpb ... s.rar?dl=0
BMMCC/BMMSC Rigs Collection https://bmmccrigs.tumblr.com
My custom made accessories for BMMCC/BMMSC https://lavky.com/radioproektor/
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Oct 21, 2018 3:47 pm

Looking at your original post, is this what you want to see?

Expose a scene properly for with the camera set to ISO 400. Let’s pretend the exposure was T4 etc.

Then with the camera set to ISO 3200, expose the scene at the identical T4 etc.?

That’s underexposed an additional three stops. Then in post you want to raise the shadows three stops to compare noise?

Apologies if I’ve misinterpreted your request.


Sent from my iPhone using Tapatalk
Rick Lang
Offline
User avatar

Dmytro Shijan

  • Posts: 1760
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Oct 21, 2018 4:18 pm

Not exactly. I mean two scenes that visually looks exact the same, but one shoot at ISO400, and another at ISO3200. Both shoot at same lens f-stop but difference in light compensated by adjusting shutter angle.
BMMCC/BMMSC Rigs Collection https://bmmccrigs.tumblr.com
My custom made accessories for BMMCC/BMMSC https://lavky.com/radioproektor/
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Oct 21, 2018 5:26 pm

Thanks, Dmitry. I’m sure someone with the Pocket4K camera will provide that shortly.


Sent from my iPhone using Tapatalk
Rick Lang
Offline
User avatar

Cris.Cunningham

  • Posts: 12
  • Joined: Sun Oct 18, 2015 9:16 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostFri Nov 09, 2018 11:18 pm

rick.lang wrote: shooting either ISO 640-800 or ISO 4000-6400 may be good choices. We haven’t seen much in those ranges to judge a best practice.
Sent from my iPhone using Tapatalk


I've received 3 of my cameras. I'm getting FPN (I think) from 640 - 800. I'll try and upload later today or tomorrow. 2 of the 3 cameras duplicate this result and the 3rd remains unopened as of right now. 400 and 1250 look incredible and 3200 begins to show noise I feel I may need to denoise. Maybe I'm spoiled as a long time A7sii owner.

I'm concerned I may have bad cameras as the results from 640 - 1000 are horrible and not useable under any circumstance.

I'll keep ya' posted.

c
Offline
User avatar

Dmytro Shijan

  • Posts: 1760
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Any P4K ISO400 vs ISO3200 test samples?

PostFri Nov 09, 2018 11:45 pm

I just found this video by Samuel H. From author comments both high and low ISO have similar noise level and similar dynamic range if recover shadows and add essential noise reduction.

BMMCC/BMMSC Rigs Collection https://bmmccrigs.tumblr.com
My custom made accessories for BMMCC/BMMSC https://lavky.com/radioproektor/
Offline
User avatar

joe12south

  • Posts: 782
  • Joined: Tue Jan 16, 2018 4:14 pm
  • Location: Nashville, TN
  • Real Name: Joseph Moore

Re: Any P4K ISO400 vs ISO3200 test samples?

PostFri Nov 09, 2018 11:58 pm

The P4K handles over-exposure better than under. I don't think there is any good reason to under-expose it. You're just gonna reveal noise.

I've basically come to think in terms of it having 3 standard ISO's: 400 / 1250 / 3200 and one "extended" ISO: 6400.

I'm not saying that's how it technically works (I know it's not) but in terms of practical thinking to get the best balance of sensitivity and dynamic range, it's a good mindset. 400 and 1250 are no compromise. 3200 is a bit noisy, 6400 can be used if you have no other choice.
Dedicated curmudgeon. Part-time artiste.
Offline

mwalker1000

  • Posts: 80
  • Joined: Sun Aug 19, 2018 1:57 am
  • Real Name: Mike Walker

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Nov 10, 2018 10:10 am



The 400 & 1250 noise seems pretty identical. 3200 is pretty impressive. Anything past 6400 is blasphemy.
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Any P4K ISO400 vs ISO3200 test samples?

PostSat Nov 10, 2018 3:08 pm

Cris.Cunningham wrote:
rick.lang wrote: shooting either ISO 640-800 or ISO 4000-6400 may be good choices. We haven’t seen much in those ranges to judge a best practice.
Sent from my iPhone using Tapatalk
...
I'm concerned I may have bad cameras as the results from 640 - 1000 are horrible and not useable under any circumstance...


Cris, do you also feel ISO 4000-6400 are horrible similar to 640-1000? I don’t expect I’ll find a little noise at 3200 a problem, but I’d like another stop to 6400 and have it look pretty good. To keep what I mean by “pretty good” in context, I stopped using ISO 1600 for the most part on the URSA Mini 4.6K as I found it was too noisy.

I don’t know if you (or anyone else) have the Mini 4.6K to be able to judge the BMPCC4K at ISO 6400 against the Mini 4.6K at 1600.

Thanks for any guidance you can provide. I’m trying to plan how I’ll use the BMPCC4K for client shoots in December. Not easy when I don’t have the camera!

The other recent posts above seem more encouraging than your remarks. But I’m not sure Resolve noise reduction is an option for me as my iMac 2015 likely is not up to that task.


Sent from my iPhone using Tapatalk
Rick Lang
Offline

John Paines

  • Posts: 5777
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Nov 10, 2018 3:45 pm

rick.lang wrote:do you also feel ISO 4000-6400 are horrible similar to 640-1000? I don’t expect I’ll find a little noise at 3200 a problem, but I’d like another stop to 6400 and have it look pretty good


Oh, man. By no stretch is 640-1000 "horrible". Anything above 400 (to 1000) has *slightly* more noise (than 400), assuming the subject is suitable for exposure at the lower gain circuit (if you're shooting in the dark, start with 1250 or above!).

OTOH, there's no reason to shoot at anything but 400 in the lower range, since 640-1000 doesn't do anything in-camera that you can't do in post equally well. You clip at exactly the same f-stop, from 100 to 1000. The other values are just getting mapped lower or higher. But the sensor performance doesn't change, until you get into the higher circuit.

3200 has a bit more noise and contrast than 400 but is thoroughly usable. Higher values will depend on your tolerance for noise.
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Nov 10, 2018 9:06 pm

Thanks, John. Always good observations. When I get the camera, I’ll give ISO 6400 a try. For much of what I do, I’ll be living in the upper band most of the time for client work. Personal stuff can go with ISO 400 most of the time.

I’ve decided to add the SLR Magic 1.33x-65 Anamorphot Adapter and that requires a minimum T5.6 setting for longer focal lengths and T2.8 up to 50mm taking lenses. So whenever I use that anamorphic look, it’s going to need the higher band as I’ll likely be shooting T5.6 or T8 a lot. That does make the Adapter easier to focus. More on this in another thread. Just mentioned because I can use those few extra stops of sensitivity in the higher band.


Sent from my iPhone using Tapatalk
Last edited by rick.lang on Sun Dec 09, 2018 7:12 pm, edited 1 time in total.
Rick Lang
Offline

Mark Grgurev

  • Posts: 802
  • Joined: Fri Nov 03, 2017 7:22 am

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Nov 10, 2018 9:55 pm

Low ISO Range
Image

High ISO Range
Image
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Any P4K ISO400 vs ISO3200 test samples?

PostSun Nov 11, 2018 2:38 am

Very nice. Very nice. Although the far right does show more chroma noise on the blinds.


Sent from my iPhone using Tapatalk
Last edited by rick.lang on Sun Nov 11, 2018 6:06 pm, edited 1 time in total.
Rick Lang
Offline

Philip Lipetz

  • Posts: 23
  • Joined: Thu Aug 23, 2012 1:14 am

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Nov 11, 2018 3:37 pm

Mark Grgurev wrote:Low ISO Range
Image

High ISO Range
Image


400 and 3200?
Offline
User avatar

joe12south

  • Posts: 782
  • Joined: Tue Jan 16, 2018 4:14 pm
  • Location: Nashville, TN
  • Real Name: Joseph Moore

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Nov 11, 2018 3:53 pm

Much of the high ISO testing I see is underexposed. Underexposure ALWAYS leads to extra noise. So here's a quick test at ISO6400 that is more or less properly exposed. (It's surprisingly tricky to create balanced lighting at such low levels.)

Graded + noise reduction:
Image

Original frame:
http://www.humcrush.com/grabs/6400/p4k-6400.dpx
Dedicated curmudgeon. Part-time artiste.
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Any P4K ISO400 vs ISO3200 test samples?

PostSun Nov 11, 2018 6:12 pm

Joe, well done. 6400 does look very usable with only a hint of an issue on the plaster wall in the shadows on the right. This looks better in terms of noise in the final image than Mark’s comparison of ISO 400 to 3200, but of course you don’t want to go to it unless you really need it.


Sent from my iPad using Tapatalk
Rick Lang
Offline

Mark Grgurev

  • Posts: 802
  • Joined: Fri Nov 03, 2017 7:22 am

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Nov 11, 2018 6:35 pm

Philip Lipetz wrote:400 and 3200?


The reason I said low ISO and high ISO is because the specific ISO within those ranges is irrelevant.

I forget what the ISO was set to in-camera. It might have been 400 and 1250 but in post I might have set them to like 800 and 4000 or something. Regardless, I exposed with the zebras. I exposed so that the highlights are clipped just as much in both.

As joe12south mentioned, a lot of people will set their ISO high and then expose for middle grey. That gives the illusion of more "highlight dynamic range" but all you're really doing is under exposing the sensor and bringing highlights closer to the midtones. If you're trying to preserve highlights then just expose so your highlights are just under clipping (or just started clipping if you have faith in Highlight Recovery) so that you sacrifice your shadows as little as possible and then choose whatever ISO you want within that range.

If you have a low dynamic range scene and your highlights are nowhere near clipping and you can give the sensor more light, then do the same thing as before. Bring the highlights to just under clipping and allow your shadows to get more light and then drop the ISO. That keeps them away from the noise floor and keeps them clean.

The one thing that's true about all sensors is that they like light. Giving them as much as they can handle will always give you the cleanest images.

Btw, for anybody interested. Here's those same images plus shot of the original BMPCC graded to kind of look the same. I didn't have a second tripod, a way to mount one camera on the other, or two lenses that allowed me to match focal lengths so this was the best I could do

Pocket 4K Low ISO
Image

Pocket 4K High ISO
Image

BMPCC
Image
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Any P4K ISO400 vs ISO3200 test samples?

PostMon Nov 12, 2018 7:03 am

Pocket4K low range and Pocket look quite similar, so hats off to BMD.


Sent from my iPad using Tapatalk
Rick Lang
Offline

Philip Lipetz

  • Posts: 23
  • Joined: Thu Aug 23, 2012 1:14 am

Re: Any P4K ISO400 vs ISO3200 test samples?

PostWed Nov 14, 2018 10:13 am

Mark Grgurev wrote:
Philip Lipetz wrote:400 and 3200?


The reason I said low ISO and high ISO is because the specific ISO within those ranges is


Only if you are shooting raw, and I did not see that specified in the original post, so I assumed it could be ProRes
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Dec 08, 2018 6:17 pm

I don't fully understand the thrust of this thread. So far I've been mostly shooting indoors with the pocket's iso set to 3200. The image is a little grainy even when well exposed. I'm wondering at what point it might be better to shoot at 400 and boost levels in post. I haven't done tests, but assume there's a grey area, where the image could be somewhat under exposed at 400 but still come out better than well exposed at 3200. Anyone care to shed some 'light' on this?
Offline

dsorbera

  • Posts: 75
  • Joined: Thu Nov 29, 2018 10:29 pm
  • Real Name: Daniel Sorbera

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Dec 08, 2018 9:25 pm

Here is a quick reference comparing iso 200, 400, 800, 1600, and 3200 with the Ursa Mini Pro 4.6k running the latest firmware.

Filmed on both cameras in UHD Prores HQ in film mode. I applied the Extended video LUT in post, but did no other post processing.
The way I set exposure was to set my zebras at 100%, and then adjust my aperture until the zebras just started to disappear on the white chip of my color card.
The pocket 4k was using a metabones 0.64x and both cameras were using the sigma 18-35. Apertures shown are what each camera told me it was at, no converting on my end.



Note the crazy color cast on the URSA (bottom) compared to the pocket which keeps it's color very well.

I also filmed every ISO of the pocket and have some very interesting conclusions, but I'll have to upload the video later as I don't have time to edit it right now. Long story short, iso 100-400 is great, iso 640-1000 is ugly (really ugly), iso 1250 is *amazingly* clean, and anything above 3200 should be avoided at all costs.
Offline

dsorbera

  • Posts: 75
  • Joined: Thu Nov 29, 2018 10:29 pm
  • Real Name: Daniel Sorbera

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Dec 08, 2018 9:43 pm

Here is the full ISO test of the Pocket 4k.



Make sure to watch in UHD for best results!


*EDIT* The youtube/vimeo compression makes it almost useless (which really says something about how we all fuss over image quality when most of our content is going to be delivered online highly compressed....).
For best results download the full files from my drive here.

URSA comparison,
https://drive.google.com/open?id=17AWzw ... mPFwwKFo8z

Pocket 4k full ISO,
https://drive.google.com/open?id=1X1G8z ... aKSrxic2W6
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Any P4K ISO400 vs ISO3200 test samples?

PostSat Dec 08, 2018 10:48 pm

dsorbera wrote:... Long story short, iso 100-400 is great, iso 640-1000 is ugly (really ugly), iso 1250 is *amazingly* clean, and anything above 3200 should be avoided at all costs.


Ouch! Say it ain’t so! I have previously concluded shooting ISO 800 and boosting in post is preferable to capturing ISO 1600 due to the noise on the URSA Mini 4.6K; so I can understand what you’re saying. But I had thought others had said ISO 6400 on the BMPCC4K was acceptable. I don’t mind some ‘film grain’ as that can enhance its appeal. But “avoided at all costs” is a heartbreaker. I guess I’ll just have to learn this lesson the hard way!


Sent from my iPhone using Tapatalk
Rick Lang
Offline

dsorbera

  • Posts: 75
  • Joined: Thu Nov 29, 2018 10:29 pm
  • Real Name: Daniel Sorbera

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Dec 08, 2018 11:13 pm

Download the videos from my drive and judge for yourself. In my opinion things get too noisy at 640, and get freakishly clean again at 1250.
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 1:15 am

Understand. But I’m hoping ISO 6400 is useful. Will take a look later when I’m on my computer.


Sent from my iPhone using Tapatalk
Rick Lang
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 5:54 am

Why would anything other than 400 and 3200 - the two native ISOs - be useful at all when shooting RAW? ... except for monitoring
Offline
User avatar

AndreeMarkefors

  • Posts: 281
  • Joined: Fri Sep 30, 2016 11:41 am
  • Location: Zürich, Switzerland

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 8:46 am

Benjamin de Menil wrote:Why would anything other than 400 and 3200 - the two native ISOs - be useful at all when shooting RAW? ... except for monitoring


When shooting raw, you have the two modes: the low range and the high range. They are two different shooting modes on the camera.

Regardless of what ISO you set up to 1000, you're in the 100-1000 range. You can change the ISO at your own discretion in post within that range. If you shoot at ISO 400, Resolve won't let you select ISO 1600 in post, since that is in the high ISO range (starting from 1250).

If shooting raw, when selecting ISO, think of it more as rating the camera or scene, because the ISO you set might influence your choice of aperture (and maybe shutter). Changing ISO in post will never effect the amount of light you capture while shooting, but setting ISO before shooting might do so, by influencing you to open up or close down (actually changing how much light the sensor gets)—this is why it's important to consider ISO even if "it's just meta data".

If you default to rating your camera at ISO 400. but then you feel the image is more than a stop too dark at your selected aperture/shutter settings, consider going into the high range by re-rating it to ISO 1250.

Try to feed your sensor as much light as possible. That said, I'm not a fan of stopping down to save a specular highlight from blowing out (0.5% of the frame) just to severely under expose 99.5% of the scene.
Current camera: Canon R5C
Mac Pro 2019: 12c | Dual W6800X Duo | DeckLink Mini Monitor 4K + LG 55C8
MacBook Pro M3 Max 16" 16/40, 64GB
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 12:50 pm

Thanks Andree.

Concerning the range vs specific native iso. Suppose I have iso at 3200, and a window is blown out. If I get rid of the clipping by lowering iso to 1250, have I really eliminated clipping? If the native iso of the high range is 3200, haven't I captured with clipping? If the raw data is the same either way, I don't see how lowering the iso within range wouldn't obscure clipping...
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 1:19 pm

My understanding (and observation using the URSA Mini 4.6K) is that clipping on a digital sensor package can only be reduced by reducing the amount of light on the sensor which means greater ND or smaller aperture or smaller shutter angle. If you are clipping ISO is irrelevant on a digital camera within a given native range.

On a film camera (analogue) using a slower speed film like ISO 50 would reduce clipping compared to using a higher rated film speed like ISO 400.

Sent from my iPhone using Tapatalk
Rick Lang
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 1:47 pm

I just ran some test with the pocket 4K and sure enough, changing the iso within range has no effect on clipping. Interestingly, this means that reducing from iso 1250 to 1000 reduces highlight clipping massively, while going from 3600 to 1250 not at all.

Within range, digital iso is a change of center grey and not a change in sensitivity - totally different from film or analog. I once read some advice regarding digital iso: to use a higher iso if you're worried about preserving highlight detail and a lower iso if you're worried about shadows - which is counter-intuitive.

In the case of dual iso (or analog gain), the reverse is true when selecting your range. It can be confusing because two functions are being performed by a single control.
Offline
User avatar

Uli Plank

  • Posts: 21109
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 4:24 pm

It has been like this from the days of the first Red: if you are recording RAW, changing ISO from the native value is only changing the way the image is presented on the monitor. You are kind of "lying" to the camera. If you call it less sensitive than it really is, you will counteract by higher exposure, so you get earlier into clipping and keep the shadows cleaner. The inverse also applies.

The only difference with the P4K is the second native ISO, starting from 1250. From there on it's again the same regarding protection of highlights or cleanliness of shadows.
Maybe AI can help you. Or make you obsolete.

Studio 18.6.5, MacOS 13.6.5
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G, iMac 2017, eGPU
Offline
User avatar

AndreeMarkefors

  • Posts: 281
  • Joined: Fri Sep 30, 2016 11:41 am
  • Location: Zürich, Switzerland

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 4:38 pm

Yes, it's never really possible to recover clipped data. So the question simply becomes: what is truly clipped and what is just a "monitoring issue" where things appear to be blown out, but are actually still intact in the raw data.

RED always had great tools here with their goal posts and stop lights. Regardless of monitoring setup you can evaluate what the sensor sees—good stuff.

In the P4K we can use the False Color mode. It shows the true clipping in red color. If you set ISO to 1250 and enter False Color and then expose so that a light source shows up as yellow, but not red, and then increase ISO all the way up to 6400 you'll see the colors shift through your scene, but the very top doesn't change. The light source won't go red. Above ISO 6400, the highlights will go red again.

Even if we can change ISO in post when shooting raw, I still feel that the ISO rating is relevant. Even if 1250-6400 share the same physical setting, the higher ISOs will still push the shadows and mids more and your rating will influence your other exposure settings (unless you strictly follow the ETTR method and just set ISO to 3200 and stop down/filter down until False Color doesn't show red). I personally don't think that's the best method, but here everyone will decide for themselves.
Current camera: Canon R5C
Mac Pro 2019: 12c | Dual W6800X Duo | DeckLink Mini Monitor 4K + LG 55C8
MacBook Pro M3 Max 16" 16/40, 64GB
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 6:58 pm

AndreeMarkefors wrote:Yes, it's never really possible to recover clipped data. So the question simply becomes: what is truly clipped and what is just a "monitoring issue" where things appear to be blown out, but are actually still intact in the raw data.


Zebras and False Colour are guides and not quite exact. You can set zebras to 100% or as low as 75%. Using 100% may give you a false sense of security. In almost every scene one colour may dominate more or less. If you are only clipping in the red channel for example, zebras will not show clipping. For this reason many people use a lower threshold such as 95% so minimize the chance of one channel clipping and not being reported to you. I used to use 85%, but now use 90% as I want to be within any margin of error. I’m not talking about specular highlights or even some larger items that are not really essential to the scene perhaps such as a very bright sky or a portion of a white fence.

... In the P4K we can use the False Color mode. It shows the true clipping in red color.


Yellow is safe but warning you it’s very bright. Red indicates clipping may be happening, but it also may not. If you change the aperture slowly you get a feel for when it is slightly red and probably not clipping versus clipping beyond a doubt. This might happen on the shoulder of a white T-shirt in the sun and in post looking at the waveform or parade you see you were still safe.

Even if we can change ISO in post when shooting raw, I still feel that the ISO rating is relevant...


Yes, agree because it influences how the internal 16bit sensor linear values are assigned to the raw 12bit log.


Sent from my iPhone using Tapatalk
Rick Lang
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 7:23 pm

Yes, agree because it influences how the internal 16bit sensor linear values are assigned to the raw 12bit log.


So does it have an irreversible affect. Shooting at 3200 and then setting to 1250 in resolve isn't the same as shooting at 1250 - all other things being equal?
Offline
User avatar

AndreeMarkefors

  • Posts: 281
  • Joined: Fri Sep 30, 2016 11:41 am
  • Location: Zürich, Switzerland

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 7:24 pm

rick.lang wrote:Zebras and False Colour are guides and not quite exact.


Sure. I'd much prefer tools designed just for the purpose of visualising what the sensor is experiencing. I guess I'm pretty pragmatic and try to find a solution with what I have at my disposal.

I don't use zebras at all personally, but I have found False Color to be very educational in showing where I can expect the signal to clip with highlight recovery turned. on, when I try to reel the data back into range. Additionally, I find it useful to see where some of the other colors land in terms of light distribution; if I stop down to eliminate red but at the same time see that the rest of the scene ends up dark blue... well, that's a call to action as good as any.

That said, as I'm writing this I checked the zebra implementation on the P4K and found that it's much cleaner than my previous experiences in other cameras. Might have to do with the large screen and good resolution. I often find zebras to be extremely annoying and imprecise. I still prefer False Color for the reason mentioned above, though.

I'll be the first to admit it's still early days. Much more varied experience needed before I'd make any claim knowing what's what.
Current camera: Canon R5C
Mac Pro 2019: 12c | Dual W6800X Duo | DeckLink Mini Monitor 4K + LG 55C8
MacBook Pro M3 Max 16" 16/40, 64GB
Offline
User avatar

AndreeMarkefors

  • Posts: 281
  • Joined: Fri Sep 30, 2016 11:41 am
  • Location: Zürich, Switzerland

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 7:32 pm

Benjamin de Menil wrote:Shooting at 3200 and then setting to 1250 in resolve isn't the same as shooting at 1250 - all other things being equal?


This example would give an identical result, as long as you don't change aperture or shutter to compensate for the image looking brighter/darker on the screen.

With the P4K you get two exposure triangles:

1. Low reference ISO (constant for 100-1000) + shutter + aperture
2. High reference ISO (constant for 1250-2560) + shutter + aperture

You example use two case 2 scenarios, so if you don't change shutter or aperture, you get the same amount of light on the sensor.
Current camera: Canon R5C
Mac Pro 2019: 12c | Dual W6800X Duo | DeckLink Mini Monitor 4K + LG 55C8
MacBook Pro M3 Max 16" 16/40, 64GB
Offline
User avatar

rick.lang

  • Posts: 17156
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 09, 2018 7:41 pm

[quote="dsorbera”] Long story short, iso 100-400 is great, iso 640-1000 is ugly (really ugly), iso 1250 is *amazingly* clean, and anything above 3200 should be avoided at all costs.[/quote]

I’ve pixel-peeped this until I saw the Second Coming, and based on looking at the 400% area, I’d conclude that 6400 is still a usable ISO rating. At ISO 8000, that 400% area show signs of trouble, much above that looks like a Mardi Gras festival. However in a normal scene with normal motion and subjects that have a variety of shades, it will be interesting to see. Under those conditions, I’m guessing 6400 is acceptable and possibly ISO 8000. Areas without much illumination or variance in colouration and shading may have some difficulty granted.
Rick Lang
Offline
User avatar

Uli Plank

  • Posts: 21109
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Any P4K ISO400 vs ISO3200 test samples?

PostMon Dec 10, 2018 1:33 am

I really wish BM had something like the stop lights in Red cameras. Perfect tool to see when you run into colored highlight issues.
Maybe AI can help you. Or make you obsolete.

Studio 18.6.5, MacOS 13.6.5
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G, iMac 2017, eGPU
Offline

Denny Smith

  • Posts: 13131
  • Joined: Thu Aug 01, 2013 4:19 pm
  • Location: USA, Northern Calif.

Re: Any P4K ISO400 vs ISO3200 test samples?

PostMon Dec 10, 2018 6:14 pm

You guys do realize, like all the BM Cine cameras, there is no analog gain added with ISO changes, except for the low to high range native setting when the dual gain sensor is switched. So at 1250 you are overexposing the sensor in high range by one stop, so exposing ETTR, and getting good results with little noise, no surprise here.

The two native ISOs are just that, “correct” exposure for most (but not all) situations. However, at ISO 1000 ( a low range setting) the sensor is being underexposed by several stops, so a muddy, noisy image. Any ISO settings between the two native settings are just exposure mid-tone image shifts, unlike other cameras that use analog gain for all ISO settings, above 0dB gain on the analog amp.
Cheers
Denny Smith
SHA Productions
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Dec 15, 2018 4:29 pm

All true Denny, but changing the ISO within a given range doesn't affect the reading on what is clipping. So you always will have an accurate measurement of clipping, regardless of the iso. What is changes is how to histogram and monitor distribute info across the luminescence spectrum.

I'm more concerned with loss of detail in the shadows than with non-clipped highlights - therefore I conclude that it's better to set iso to a low level within a range - ie below native for that range. Or put another way more light on the sensor is better as long as it doesn't clip.

Based on that, in the high range I'd always shoot at 1250 and not at 3200. I've in any case noticed that if I set my exposure at iso 3200, the 4K's image comes out quite grainy. There's less grain in the low range, so perhaps it's less of an issue there.
Offline

Denny Smith

  • Posts: 13131
  • Joined: Thu Aug 01, 2013 4:19 pm
  • Location: USA, Northern Calif.

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSat Dec 15, 2018 6:36 pm

Good point about clipping Benjamin, since their is no analog gain with changing ISO is a given range, it does not effect sensor clipping, thanks why using Zebras or False Color is important to prevent clipping.

I useually figure what is important in a scene and expose for that, if people are the subject, than I expose to have consistent skin tones. Then add or subtract light in the rest of the scene that I want exposed within the limits of the camera. You figure out where you want the mid tones to fall, and set the ISO to shift it where you want, kind of a modified Zone System, which I used in landscape work.
Cheers
Denny Smith
SHA Productions
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 16, 2018 4:23 pm

If shooting a person for instance, will important parts of the image become compressed if overexposing - even if highlights aren't clipping? Ie is recording of light hitting the sensor non-linear? Does the data become compressed as the sensor approaches clipping, or is the information just below clipping just as rich as that exposed to the middle of the sensor's sensitivity range?

For sure, as you approach clipping in the shadows, you lose a lot of detail. I'm wondering if the same is true in the highlights.
Offline

John Paines

  • Posts: 5777
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostSun Dec 16, 2018 4:40 pm

Benjamin de Menil wrote:If shooting a person for instance, will important parts of the image become compressed if overexposing - even if highlights aren't clipping? Ie is recording of light hitting the sensor non-linear? Does the data become compressed as the sensor approaches clipping, or is the information just below clipping just as rich as that exposed to the middle of the sensor's sensitivity range?

For sure, as you approach clipping in the shadows, you lose a lot of detail. I'm wondering if the same is true in the highlights.


Highlights recorded near clipping will not retain the same level of detail and flexibility in grading as the same levels recorded at lower levels. You can only push them so far.

But these limitations and non-linear variations are what create the non-realistic image we associate with cinema. Look at the number of shots from big budget Hollywood films, where the sky is 100% blown out. They don't care, and won't composite in an actual sky, because the subject isn't the sky and it's not a National Geographic shoot.

Same with shadows. Western art has crushed shadows for hundreds of years. This includes large format art photography and (of course) cinema. This idea of preserving as much detail as possible in shadows and highlights is completely foreign to the visual and dramatic arts.
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostMon Dec 17, 2018 12:01 pm

The question of what looks good in the final grade is a separate one. I'm wondering if it makes sense to overexpose when shooting with the pocket 4k. People say for instance 'the sensor likes light'. if highlight compression results in similar loss of detail as shadow compression, then the best is probably too stick close to native iso. In my experience however, it's better to overexpose than to under expose.
Offline
User avatar

Uli Plank

  • Posts: 21109
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Any P4K ISO400 vs ISO3200 test samples?

PostMon Dec 17, 2018 1:28 pm

That depends on the question if noise or color fidelity near the highlights is more important for you.
Maybe AI can help you. Or make you obsolete.

Studio 18.6.5, MacOS 13.6.5
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G, iMac 2017, eGPU
Offline

Benjamin de Menil

  • Posts: 330
  • Joined: Wed Feb 04, 2015 9:04 pm

Re: Any P4K ISO400 vs ISO3200 test samples?

PostTue Dec 18, 2018 4:32 pm

Uli Plank wrote:That depends on the question if noise or color fidelity near the highlights is more important for you.


So color fidelity is lost near highlights?
Offline
User avatar

Uli Plank

  • Posts: 21109
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Any P4K ISO400 vs ISO3200 test samples?

PostWed Dec 19, 2018 1:42 am

I wouldn't say 'lost', but decreasing.
Maybe AI can help you. Or make you obsolete.

Studio 18.6.5, MacOS 13.6.5
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G, iMac 2017, eGPU

Return to Cinematography

Who is online

Users browsing this forum: No registered users and 35 guests