Sunny 16+

Started by Aradi, June 17, 2019, 10:16:14 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Aradi

Hello,

First, I would like to thank all ML contributors for their time and effort. I have enjoyed ML on my 60d, and currently on my 6d, and really appreciate the extra skills it teaches my cameras. Thank you.

Now, my simple proposition might go against ML highly technical fiber, but I'm sure it's feasible so please bear with me. I suggest ML allows the user to choose the lighting condition, and then set the appropriate SS/A/ISO accordingly, observing the relevant limits of the various settings. The user can then engage Exposure Lock and manipulate this 'correct' exposure setting to achieve a more desirable depth of field, or a particular level of noise, or even override the settings altogether for high/low key effect... etc.

The reason I'm asking is that I have used the sunny 16 rule and its close siblings for quite a while, and the results are never further than one half of a stop from where I expect them. And, while everybody can easily memorise and apply the sunny 16 rule -1/-2 stops, the full tables for exposure values with the full description of each lighting condition, those tables are not that easy to commit to memory. For that step I refer to an app on my mobile, and then copy the suggested numbers in my camera. It would be great if ML could take over the app role, as well as ensure the settings are copied accurately :-)

I'm sure you're all aware that many versions of those tables are widely available in books and over the web, and hope that if you decide to implement this feature, please keep the lighting conditions descriptions in a text file that the user can modify/translate... Perhaps not everyone likes the longer descriptions that I prefer, and they should have a way to modify them to their liking.

Please let me know what you think.

Regards

Walter Schulz

You may want to add something about why ETTR or zebra overlays won't do the job.

Aradi

Hello Schulz,

Thank you for your prompt response.

ETTR and zebra overlays work perfectly well, on reflected light. They try to fit the measured signal as best as possible given the sensor's characteristics. Sunny 16 et al. on the other hand work on incident light, with little regard to the sensor or otherwise. To be sure, both metering methods have their pros and cons. In this particular case, a sunny 16 exposure would require more noise reduction, while ETTR calls for exposure adjustment in post processing.

Perhaps, if one has the time, a more 'accurate' approach would be a sunny 16 exposure for reference, AND an ETTR exposure for raw cleanliness and minimal clipping. That is surely something I would like to test in the near future.

What do you think?


Aradi

Hello Audionut,

This is also auto exposure, where the camera meter measures reflected light and acts accordingly. I would like to decide on the incident light, and manually let the camera know that it's a heavy overcast scene, so that it sets the appropriate shutter/aperture/ISO to achieve an EV of 12.


Audionut

If you're going to manually do something, then just manually adjust the exposure settings to suit your peculiar needs.

This sounds like a classic case of moving the goal posts.

Aradi

Dear Audionut,

Camera meters measure REFLECTED light, which Canon exploits to make available aperture priority, shutter priority, program mode...etc. ML too exploits the same meter to provide AETTR, zebras, Auto exposure...etc. All cool and useful stuff that I really appreciate and enjoy, thank you. However reflected light meters fail photographers every now and again, like when the subject is black, or white, or when you use polarising filters...etc.

INCIDENT light on the other hand is more robust, with high repeatability. You stick your incident light meter next to your subject's face, point it to the camera position, and take a reading. This might not be convenient at times, like when the subject is way too far, so photographers have relied for decades on the sunny 16+ guidelines. You just look around and notice that the subject is in sky lit shade, then you lookup those conditions on a piece of paper, et voila, f/4 @ 1/125 for your ISO100.

Now my suggestion was about that piece of paper. That piece of paper, the one with lighting conditions and camera settings, evolved into quite a few mobile apps and online exposure calculators, so evidently not peculiar from a photography perspective. I hope you agree that that same useful information can be incorporated into ML, with the logical extra step of ML setting the exposure parameters to match the manually selected lighting condition.

Lastly, I would very much appreciate it if you were to observe the forum rules, and either stick to the subject of the discussion or refrain from responding. I'm honestly not sure about peculiar, nor do I see where goal posts came from, let alone the prior irrelevant response. Indeed, I'm surprised that the differences between reflected and incident light metering need to be clarified at all on a photography related forum.

Please let me know if it could be done, and thanks a lot in anticipation.

a1ex

Quote from: Aradi on June 20, 2019, 11:58:36 PM
Indeed, I'm surprised that the differences between reflected and incident light metering need to be clarified at all on a photography related forum.

I'm not familiar with this concept (maybe because I have absolutely no experience with film), and I've never used this technique. I'm also not convinced about the usefulness of this technique over ETTR, especially given the limited dynamic range of Canon sensors. So, Walter's reply was not irrelevant at all - you need to come up with some good explanations, relevant links or whatever.

Anyway - a quick search revealed this page: http://www.fredparker.com/ultexp1.htm

There, you can find a table (called "Exposure Value Chart") with various situations, such as:
Quote
11    Sunsets. Subjects in deep shade.
12    Half moon (long lens). Subject in open shade or heavy overcast.
13    Gibbous moon (long lens). Subjects in cloudy-bright light (no shadows).
14    Full moon (long lens). Subjects in weak, hazy sun.
15    Subjects in bright or hazy sun (Sunny f/16 rule).
16    Subjects in bright daylight on sand or snow.

Now, set the exposure to 1/125, f/16, ISO 100, then look in ML menu. On the right side of Expo Lock, you will see Bv 10.0 (see this article for reference). Apply eq. 9 (i.e. add 5) to get the 15 from that table.

Let's also check 1/125, f/4, ISO 100. Bv 6.0 => 6+5 = 11 in that table. It seems to match your description.

So, the exposure calculator is already implemented in ML.

Now, what's your exact request?
- ability to dial some custom Bv value? (expo lock does pretty much that)
- print the above texts somewhere on the camera screen? (apparently the author asks for at least $3500 for a license)
- print that Bv value in a different place? (such as, on the info bars in LiveView?)
- print Ev = Bv+5 instead? (not recommended, see page 10 in the linked article)
- something else?

Or maybe we can we mark this feature as "already done"?

Audionut

For those times where the reflected light meter fails "every now and then", you can use the raw spotmeter to expose the subject.

dfort

Quote from: Aradi on June 17, 2019, 10:16:14 PM
...those tables are not that easy to commit to memory. For that step I refer to an app on my mobile, and then copy the suggested numbers in my camera. It would be great if ML could take over the app role, as well as ensure the settings are copied accurately :-)

Which app are you referring to?

Quote from: Aradi on June 20, 2019, 11:58:36 PM
...reflected light meters fail photographers every now and again, like when the subject is black, or white, or when you use polarising filters...etc.

INCIDENT light on the other hand is more robust, with high repeatability. You stick your incident light meter next to your subject's face, point it to the camera position, and take a reading...

This is why incident light meters were (and still are) more popular than reflective light meters with cinematographers. Gaffers (the guys that set up the lights) usually communicate in foot candles which is a measure of incident light (reflected light is measured in foot-lambert) -- note that this professional light meter (it says professional right on it) has an fc (for foot candle) scale:



Ok--for those who live in countries with a sensible measurement system, incident light is measured in candela and reflected light in candela per square metre. Anyway, these days it seems that still photographers prefer the histogram display (helps when exposing to the "right") while cinematographers prefer to look at the waveform.

Back on topic--

Quote from: a1ex on June 21, 2019, 11:37:22 AM
I'm not familiar with this concept (maybe because I have absolutely no experience with film), and I've never used this technique.

Pretty much every box of film came with an information sheet that had a chart that looked something like this:



One of my photography instructors liked to joke around and would sometimes pull out the information sheet in front of a client and ask if he thought it was a hazy bright day or a cloudy bright day then have his assistant check for his shadow.

Rollieflex took it to another level with this chart that was permanently attached to back of some of their twin lens reflex models:



I guess someone could come up with a module that would show a diagram like this on the LCD screen but I doubt this is what @Aradi has in mind.

Quote from: Aradi on June 17, 2019, 10:16:14 PM
... The user can then engage Exposure Lock and manipulate this 'correct' exposure setting to achieve a more desirable depth of field, or a particular level of noise, or even override the settings altogether for high/low key effect... etc...

Ah ha--that's different. Exposure value (Ev) is a number that represents a combination of a camera's shutter speed and f-stop. Note that Rollieflex diagram has the exposure expressed in Ev numbers. Some of the older cameras, especially the medium format cameras with leaf shutters like the Rollieflex and Hasselblad could be locked into an Ev number so you could easily dial in your preferred shutter speed or preferred aperture without changing the overall exposure. Note that on this Hasselblad lens the depth of field indicators would adjust to the aperture and the numbers in red on the right of the shutter speed are the Ev numbers:



What @Aradi is asking for seems more like an extension of one of the lua scripts @garry23 wrote which is available on his website - here is the source. You set some combination of shutter speed, lens aperture and ISO so the module comes up with the Ev number and if you change the aperture the shutter speed automatically adjusts or if you adjust the shutter speed the aperture the aperture automatically adjusts. If you reach a limit a warning should pop up. Also -- there should be a depth of field indicator which is the part that has already been written. Is this the request or am I way off base?

So, do I practice all of this? Well, I do have an incident light meter:



I pull it out at least once a year to check if it still works -- looks like I need to replace the battery (my old Spectra didn't use batteries). How do I determine my exposure? For still photography there is this [A+] setting on all of my cameras that seems to work quite well, though it isn't compatible with some of the Magic Lantern features so when I need to switch to manual mode I use a professional technique called bracketing. I used to burn through a lot of film back in the day but with digital images I simply erase what doesn't look good and people think I'm a genius. When I shoot video I don't like auto mode because exposure can change in the middle of a shot so I simply adjust the settings until it looks good in the EVF and go with that.  ;D

a1ex

Well, believe it or not, this request is what made me take the camera (5D2) out of the bag - after a long break - and try out the Bv thingie. Previously, I wrote that exposure calculator in order to check the math behind Expo Lock - as I had no idea about these rules of thumb.

I've set the exposure as I usually do - by looking at the raw histogram in M mode. At least on 5D2, that requires a bit of trial and error, because - for some reason - LiveView raw histogram doesn't quite match the one in photo mode (I need to look into that). In bright sunlight, there's also a problem of not being able to read the fine print / low contrast stuff on the screen.

QuoteYou set some combination of shutter speed, lens aperture and ISO so the module comes up with the Ev number

This part is already done - the Bv number printed on the right of Expo Lock in ML menu. Add 5 in order to get EV100.

Quoteand if you change the aperture the shutter speed automatically adjusts or if you adjust the shutter speed the aperture the aperture automatically adjusts.

Well, Expo Lock does exactly this. Of course, I had no idea about Sunny f/16 when I wrote that feature. I just wanted equivalent settings that gave the same exposure :D

Back to my "experiment": I ended up checking this Bv value, and - to my surprise - got pretty consistent numbers: 9 (i.e. EV100 = 14) in sunlight, about 4-6 in shadows (EV 9-11) and so on. The "standard" numbers from the f/16 rule resulted in significant underexposure, but after seeing the Rolleiflex chart - in particular, the shadow length - that pretty much explains these variations!

So, I now see some value in this technique, for cross-checking the exposure numbers. Maybe also for dialing some quick exposure when you don't have much time to fiddle with settings, or when you don't trust the (semi-)automatic modes. Pretty much all I was missing was that Bv (or Ev) number printed on the main screen (both in LiveView and outside). Going back into ML menu (Expo tab) to check it required too many clicks.

TLDR: as I understand it, the request is about printing the Bv number (or the EV100 one) on the main screen. The automation (i.e. keeping it constant as you change exposure settings) is already done.




Incident light meter: hey, wouldn't the ambient light sensor - present on the 5D2 and 5D3 - be a good candidate for that? Of course, as long as you don't cover it with your hands, head or whatever :D

BTW - the 5D2 is my main camera since January, when the 5D3 went defective. The M50 is not operational yet - I still need to get some lens(es) for it.

Aradi

Dfort and a1ex, I cannot thank you enough.

Dfort got it right almost allll the way. Those diagrams are exactly what I had in mind when I posted the request. There are apps now that do the job of those diagrams, 'Photo Friend' is one of them that is available for iOS as well as android. And those apps encompass a much wider selection of lighting conditions than the Rollieflex diagram did.

And a1ex had me plan to at last run an experiment I kept putting off, details to follow. I will run the experiment tomorrow, post the results, and I hope you will like it.

I was aware ML displayed the Bv associated with the selected Tv, Av, and Sv, as well as their correct APEX values. My request is to allow the user to SPECIFY a Bv from a drop down menu or a list, and then ML would select a corresponding valid Tv, Av, and Sv combination. The user could then, thanks to Exposure Lock, change those settings to a more desirable equivalent exposure.

That we specify it in Bv, or Ev, would be of little help. It's the description of the lighting conditions that matters. And if fredparker.com author asks $3,500 for a 'license' to use his wording, then we should get those descriptions from elsewhere for free, like from wikipedia, or https://www.thephoblographer.com/2016/04/16/the-ultimate-guide-to-the-sunny-16-rule-part-2/ , or https://www.slrlounge.com/photography-essentials-the-sunny-16-rule/ , or a dozen other sources. I would also suggest those descriptions be kept in a user editable text file or csv, for every user to modify to his/her liking.

As for my experiment, it's an investigation on the effect of over/underexposure on colour accuracy. I've stumbled upon this post https://www.blog.jimdoty.com/?p=10138 a few months ago, and I have since planned to check it out for myself. I would like to learn just how far can I safely over/underexpose before colour is permanently damaged. But ever since I've read that post, I have been keen on nailing the correct exposure in camera. For correct exposure I reverted to good old Sunny 16 and co., and they have been as helpful as ever.

On exposure, IMHO, AETTR is a great feature. Rather than have the blunt auto exposure averaging the scene brightness and placing that average at say -4.0EV from highlights clipping, AETTR takes the 99.9th percentile and places it at -0.0EV from clipping. A refined highlight priority if you like. Still, it is not infallible, as specular highlights for example are sure to sway exposure way into underexposure. Spot metering is also a great feature. Still, like with all reflected light metering you need to account for the subject tone, colour, and texture. IF you spot an 18% gray card, AND your white balance is correct, only then would you get a reading comparable to that of an incident light meter. As useful as they are, I don't see how could AETTR or Spot Metering help with a bird in flight for example, or fireworks, or a moonlit beach...etc. Trust me, for those situations and more, Sunny 16 and co. save the day.

So, A1ex, I suggest we do NOT engage the meter at all. We ignore it and we SELECT the lighting conditions from a list. Our eyes become the meter.

Cordially

Kharak

Do you shoot jpeg only?

I don't see why you don't just use Raw Zebras and learn to shoot in M mode and manually ettr.

Seems extremely slow in the heat of the moment to go to a drop down menu and select a scene and then maybe choose another because it was under/overexposed, meanwhile you are taking the eyes of the action.

once you go raw you never go back

Aradi

Hello,

I shoot exclusively RAW, AND exclusively in manual mode.

Raw zebras are yet another way to interpret the reading of the same meter, with the same reflected light complications. It is that meter that I am trying to eliminate from the workflow. Describing a dozen other ways to use the same reflected light meter is totally irrelevant from the point of view of my request, because my request is just about CIRCUMVENTING that meter, IGNORING it totally.

And ettr, whether auto or manual with zebras or otherwise, is by definition NOT correct exposure. Correct exposure is when white is captured white, black is captured black, and gray is captured gray. ETTR is about capturing the brightest as white, whether that brightest was white, black, or anything in between.

And from the heat of the moment viewpoint, you change those sunny 16 settings only when the light changes. You might not touch them for hours if the weather permits. And trust me, if in the heat of the moment you meter every shot differently, you are missing exposure more often than not, and colour accuracy most probably suffers too. We'll be sure when I post my experiment details tomorrow.

Thanks

Aradi

Now when I think about it, perhaps AETTRight needs to be complemented with AETTLeft, but that would be another thread.

But isn't that basically how you would bracket shots to begin with. Except that one wouldn't have to guess where the brightest of the highlights and darkest of the shadows are.

Or maybe it would be enough for ML to print the SBR somewhere on the display. Or perhaps start with AETTR and then overexpose by 1Ev for as many shots as the SBR requires. I need to think about it before I bother you with yet another thread.

Kharak

Sounds like we have two different approaches to exposure.

I strive to capture the maximum possible DR/information in the raw file, by ettr and setting correct exposure in post. In my opinion that is the correct way of handling raw data, fill it to the brim with information then manipulate it. I used to think I got creamier highlights e.g blue sky, non blown out clouds etc by shooting the "look/exposure" in-cam e.g. setting correct exposure, sometimes leaving 2 stops of headroom before even the green channel blew out. I since learned that it was a combination of my lack of skills in color grading, using bad CG software and a dependency on lut's that caused my ettr shots to seem dull and blown out. Which in turn also made me a worse shooter.

It has to be said that i dont think one should religiously ettr, sacrificing shutter speed, depth of field or any other physical aestethic of an image one wants. The image aestethics come first, the trick is to keep the aestethics and manage to ettr.

To each his own, you got a1ex out of hibernation so it must be something worth looking at. Luckily we can all shoot and make pictures the way we want.

Hope @a1ex doesnt Stack an Overflow of work and maybe takes a quick look at some stuff before going back sleep. ;)
once you go raw you never go back

dfort

Exposing to the right isn't exactly a new technique specific to digital photography. Negative film, both color and black and white, tends to have more latitude to overexposure than underexposure. When I was working professionally most of my assignments were for magazines and they required color transparencies (chrome film) which had practically no exposure latitude at all, maybe 1/2 to 1 stop at the most. You can think of shooting chromes as the equivalent of JPEG and negative film as RAW.

If you really want to get into the nitty gritty of working with the photochemical process, get some Ansel Adams books and learn the zone system. When I first learned photography the bible of black and white was his The Camera, The Negative and The Print books. An oversimplification on how it works is to expose for the shadows and process for the highlights--though there is a lot more to it than just that.

The way I worked was to shoot Polaroids until I got the lighting and exposure working then bracket exposures on film. If bracketing wasn't an option, like when shooting fashion, I would send out clip tests to the lab and had the film pushed or pulled in processing if necessary.

Of course with digital cameras you can preview the image in camera so there is no need for any of this, right?

Back on to the Sunny 16+ topic, even Ansel Adams used it on one of his most famous images, Moonrise over Hernandez, because he couldn't find his light meter and the light was changing rapidly.



Many years ago I attended a lecture featuring Ansel Adams (yes, I'm that old) and I remember him saying that his saving grace was knowing that exposing for the full moon is the same as exposing for a sunny day -- f/16 at the shutter speed that matches your film's ASA (ISO) rating.

Aradi

Hello,

I did the experiment, and the results are in. HYG: https://www.dropbox.com/sh/e6jsdvp9x6wn66z/AACPd4sel6C1PhNhfOPXbjZ_a?dl=0

If you're not familiar with the target, here are the specs: https://www.bartneck.de/2017/10/24/patch-color-definitions-for-datacolor-spydercheckr-48/

I used a collapsible grey target for correct exposure and white balance. The 7 photos run from correct exposure -3Ev through correct exposure +3Ev. Other than adjust their exposure, nothing was done in post.

As you can see, between +2Ev and -2Ev nothing interesting happens, and only few colours shifted at -3Ev, I can see only 5 patches slightly shifted. At +3Ev too many colours were damaged, most notably the lighter skin tones in column C, and sky blue and foliage green in column A.

I shall repeat the experiment shortly for the following reasons:
- I would like an AETTR exposure too, didn't think of that this time around.
- I will use ML Advanced Bracketing to go from correct exposure +/-5Ev.
- Anything else you might like to ask for :-)

Cordially

Aradi

Hello a1ex,

Just a note, in case you accept my proposition and actually implement the sunny 16 rule into ML. This relates to the bottom two rows of the Rollieflex diagram dfort was kind enough to post earlier. Those are modifiers to all the Ev's in the table, both for the time of day/year as well as for the weather.

Here you go, and dfort please correct me if I'm wrong. If the sun is high in the sky, so that your shadow is shorter than you are then you're good with the Ev from the table. If your shadow is taller than you are, you overexpose by one stop. If your shadow is also more than twice taller than you are, you overexpose by one more stop. If shadows of objects in direct sunlight are hazy, then you overexpose by one more stop. 

Many people didn't do that, thinking that because a subject is in a skylit shade then the exposure was always the same. They were wrong, as a skylit shade is much darker at sunset than it is at noon. Still, the guys in the labs worked harder on their negatives and produced acceptable prints nonetheless :-)

Cheers

Audionut

Quote from: Aradi on June 23, 2019, 10:10:37 PM
AETTR takes the 99.9th percentile and places it at -0.0EV from clipping. A refined highlight priority if you like. Still, it is not infallible, as specular highlights for example are sure to sway exposure way into underexposure.

Quote from: RenatoPhoto on May 25, 2013, 07:24:06 PM
Highlight ignore:

  • Set the amount of pixels in the image, that are allowed to be overexposed.
  • This works like an specular highlight control.

Exposure target:

  • Where to place the highlights with respect to overexposure.

Allow Clipping:

  • Off:  Only allow overexposure as per the highlight ignore setting.
  • Green Channel:  Ignore the green channel, and calculate exposure settings from the other channels.
  • Any Channel:  Allow some overexposure, over and above the highlight ignore setting, in any channel.


Quote from: Aradi on June 23, 2019, 10:10:37 PM
Spot metering is also a great feature. Still, like with all reflected light metering you need to account for the subject tone, colour, and texture. IF you spot an 18% gray card, AND your white balance is correct, only then would you get a reading comparable to that of an incident light meter.

If shooting JPG, sure.  RAW is not affected by WB.  Spot meter the target in RAW EV, and expose the spot to a specific EV.


Quote from: Aradi on June 23, 2019, 10:10:37 PM
As useful as they are, I don't see how could AETTR or Spot Metering help with a bird in flight for example, or fireworks, or a moonlit beach...etc.

Anyone can come up with a use-case to break any exposure style.