why no "blown pixel count" in (raw) viewfinder? Instead of histogram?

Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

RussellInCincinnati said:

Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

Click to expand...

I am a RAW( ;-)) beginner in photography, having purchased my first "serious" camera, M4/3, a couple of years ago and (finally) began shooting RAW about a year ago. Slow learner, over-influenced by well-meaning folks who (wrongly) advised me to "get experience" shooting JPEGS.

This is the very first time I venture forth into this hallowed PST Forum. So beware, what I say is from the pov of a beginning learner of this game.

I went through a very confusing time reading all sorts of noise about exposure triangles etc before I finally learned how to get acceptable exposures shooting RAW, ETTR, spot metering to survey scene luminance, live view histogram, back button AELock, release, chimp RGBY histogram to confirm, calibrated eyeballs in RawDigger. UniWB is too complicated for me to implement. I have thought about the Oly E-M5, E-M1 with their live-view blinkies but don't need another camera (with another UI to learn)!!

I finally got out of the deep hole of misunderstanding caused by the Exposure Triangle paradigm by reading Gollywop's Exposure Vs Brightness and am starting to understand most of it. (In spite of some very "difficult" definitions in the DPR Glossary ( like DPR "Exposure" , and DPR "ISO" where in the first case the luminance (light/area) is replaced by "amount of light" (which causes lots of confusion at DPR where many authors don't realize that exposure is a function of light/area and noise is a function of total amount of light and therefore exposure X sensor area) and in the second case the stated relationship between ISO setting and noise denies the existence of ISO-less sensors. )

Back to your excellent suggestion:

Beginning RAW shooters like myself need to detect two possible conditions for loss of highlights at the time of making the exposure.

1. over-saturated sensels due to overexposure ( due to: too open f/ and/or too short shutter interval, for the scene luminance.
2. clipped data due to over-brightening (excessive gain, due to setting ISO too high)
3. (a third possibility, excessive compression at the high end, is something that I neither understand nor worry about, yet!)

In the case of 1. above: I will dial in a change of f/ and/or shutter either directly or by EC.

In the case of 2. above: I will lower ISO if I have a camera that is ISO variant ( a camera that applies analogue gain.) to keep the exposure as high as possible. (If I am already at base ISO, I stop what I am doing and try to figure out what is going on!!!)

In the case of 3. above: I won't do anything at image capture but will pull down and/or spread out the highlights in post conversion processing.

Your suggestion will deal with Case 1. .... in a very elegant way!

But.....

Histograms would give me a Case 2 warning ---- in both live-view and post-exposure (since both are informed by jpegs which are influenced by ISO settings.) They give me other information also, when I survey the scene luminance. (I do not yet have the well trained eyes of an experienced photographer.) I can always turn them off if they clutter too much.

( I don't know if the Olympus blinkies are initiated by overexposure or overbrightening .... or both? ie are they informed by the sensor or by the jpeg sidecar in RAW.)

Related subjects are

• The UI for ISO-less cameras and
• Improvement in (sidecar) JPEG quality when shooting RAW so that we don't have to use RAW + JPEGfine to check fine focus when chimping.

so maybe the issue of levels of (exposure, brightening) in the UI should be considering in light of the above related UI issues.

My 2 cents worth from a beginner.

Tom

RussellInCincinnati said:

Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

Click to expand...

as noted - get Olympus m43 cameras - they (camera mode, WB, OOC JPG settings, etc) can be perfectly tuned to display raw clipping (not some stupid %, but actually in the exact places in frame where the clipping happens) in ambient light in EVF... otherwise lobby your camera manufacturer to repeat what Olympus did, it is easy

exdeejjjaaaa said:

RussellInCincinnati said:

Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

Click to expand...

as noted - get Olympus m43 cameras - they (camera mode, WB, OOC JPG settings, etc) can be perfectly tuned to display raw clipping (not some stupid %, but actually in the exact places in frame where the clipping happens) in ambient light in EVF... otherwise lobby your camera manufacturer to repeat what Olympus did, it is easy

Click to expand...

Do you know

if the "blinkies" are informed by the sensor to give an indication of over-saturation of the sensels

OR

if the "blinkies" are informed by the jpeg (both in live view and post exposure) to give an indication of data clipping in the ADUs due to too much gain.

OR

both?

Thank you,

tom

(btw, I agree that the OLY implementation of the blinkies for RAW shooter is a great improvement, as it gives the photographer a good idea of where (in the frame) the blowing/clipping is occurring. I wish that my cameras had that feature. It is feature that almost led me to buy a E-M5!)

RussellInCincinnati said:

Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

Click to expand...

Histograms of the three channels are not strictly necessary but they are useful, for instance, to see at a glance:

1) how far short/over of ETTR you are
2) how far short/over blocking shadows you are
3) what the DR of the scene is
4) whether you are exposing to the right/left the right stuff - perhaps you want to let some highlights blow but not others or conversely let some shadows be blocked but not others - and the ones you may want to keep may be obvious in the histogram

Also, as mentioned in the other threads, a histogram is not complete without blinkies that show you what pixels you are letting blow/block. RawDigger docet.

All of this does not mean that it wouldn't be a good idea to have some sort of dedicated blown/blocked raw data indicator. It would be an excellent idea, as requested without satisfaction for years. So let's keep asking.

Jack

Jack Hogan said:

RussellInCincinnati said:

Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

Click to expand...

Jack Hogan wrote: Russ, this is the third thread you start on the same topic. Why spread the love? Keep it all neatly packaged so that future generations will not be confounded by the shallowness of your threads

Click to expand...

The plan is for future generations to have fun piecing together the aggregate wisdom. Not unlike the continuing work with computers, to puzzle out the fitment of scattered fragments of the Dead Sea Scrolls.

Jack Hogan said:

Histograms of the three channels are not strictly necessary but they are useful, for instance, to see at a glance:

1) how far short/over of ETTR you are
2) how far short/over blocking shadows you are
3) what the DR of the scene is
4) whether you are exposing to the right/left the right stuff - perhaps you want to let some highlights blow but not others or conversely let some shadows be blocked but not others - and the ones you may want to keep may be obvious in the histogram

Also, as mentioned in the other threads, a histogram is not complete without blinkies that show you what pixels you are letting blow/block. RawDigger docet.

Click to expand...

My not previously clearly stated thought is that thinking you "need" the camera to show you exactly where the bright pixels, for all the reasons you detail, is about as important as me "needing" to photograph a full color chart in order to get a "perfect" white balance. (Which have wasted time with in the past). Of course to get "perfect color balance" you "should" photograph a huge color chart with scores of patches. The color patches photographed in the dominant scene light, in order to supply your raw processing software with the perfect mapping of scene colors to output colors.

But realistically, when doing anything but precise copy-photography of artwork, or product photography for an art director, nobody cares about having a color balance better than what you get by simply setting the white balance curves off a single snapshot of a white card.


In this ordinary (complex mixed lighting) tourist-type photo, as long as the white t-shirt reads as white, the additional color accuracy that might have been added by photographing a full color chart wouldn't have amounted to a hill of beans.

Analogously, am claiming that all you probably "need" to know as an expose-to-the-right person is how many pixels are blown out high. Why do you really need special blinkies to tell you what pixels are blown out, when what you are most often looking for is almost no blown out pixels? And why do you need blinkies to show you what's blown, when an observant photographer (which anyone who is thinking about this kind of thing must be) can guess from the unadorned display what's going to be burnt out?

The points being both to not (a) overload a display with "unnecessary" markings, (b) not to present "unnecessary" info that mostly slows down interpreting the most important overexposure info, and (c) also to minimize the amount of work camera firmware has to do to get us the most-important expose-to-the-right information.

I don't even think we need to know which of red, green or blue channel pixels are blown. Just show us the single number of total blown-in-raw-file pixels, and we'd be far ahead of where we are now.

Jack Hogan said:

All of this does not mean that it wouldn't be a good idea to have some sort of dedicated blown/blocked raw data indicator. It would be an excellent idea, as requested without satisfaction for years. So let's keep asking.

Click to expand...

To which am adding let's ask for the bare minimum on-screen info we could use to avoid blown pixels. Save histograms and blinky warnings for a version 2 of the idea.

exdeejjjaaaa said:

RussellInCincinnati wrote: Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

exdeejjjaaaa said:

exdeejjjaaaa wrote: as noted - get Olympus m43 cameras - they (camera mode, WB, OOC JPG settings, etc) can be perfectly tuned to display raw clipping (not some stupid %, but actually in the exact places in frame where the clipping happens) in ambient light in EVF... otherwise lobby your camera manufacturer to repeat what Olympus did, it is easy

Click to expand...

Click to expand...

Name-calling a minimalist, super quick to interpret and easy to implement display of blown-pixel-count "stupid", compared to display clipping blinkies, does not prove that we need the chartjunk of blinking pixels. I certainly don't need to know where the blown pixels are, nearly as badly as I need to know whether or not there are a significant number of blown pixels at all. And I've already got focus peaking markings to interpret on the main display, maybe I don't want to interpret blinking/crawling/whatever clipped pixel marks as well.

For example if my display says I've got 20 blown out pixels out of 20 million, I know that my exposure is not too high, who cares where the 20 blown pixels are.

The advantage of blinkies is that you can see whether they occur in specular highlights. But I'm not very fond of them on the PC in LR, they can be very visually domiant. LR at least lets you turn them off/on easily

But I hadn't thought before about how you could get both peaking and blinkies without getting visual overlad.

I still like bobn's mock-up best. There was another one done on LuLa, I'll see if I can find that

GeorgianBay1939 said:

exdeejjjaaaa said:

RussellInCincinnati said:

Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

Click to expand...

as noted - get Olympus m43 cameras - they (camera mode, WB, OOC JPG settings, etc) can be perfectly tuned to display raw clipping (not some stupid %, but actually in the exact places in frame where the clipping happens) in ambient light in EVF... otherwise lobby your camera manufacturer to repeat what Olympus did, it is easy

Click to expand...

Do you know

Click to expand...

I did tune setting with rawdigger ( http://www.rawdigger.com/ ) in hand, so yes, I do know that they indicate for me clipping in raw channels w/ very good precision

RussellInCincinnati said:

RussellInCincinnati said:

RussellInCincinnati wrote: Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

RussellInCincinnati said:

exdeejjjaaaa wrote: as noted - get Olympus m43 cameras - they (camera mode, WB, OOC JPG settings, etc) can be perfectly tuned to display raw clipping (not some stupid %, but actually in the exact places in frame where the clipping happens) in ambient light in EVF... otherwise lobby your camera manufacturer to repeat what Olympus did, it is easy

Click to expand...

Click to expand...

Name-calling a minimalist, super quick to interpret and easy to implement display of blown-pixel-count "stupid", compared to display clipping blinkies, does not prove that we need the chartjunk of blinking pixels. I certainly don't need to know where the blown pixels are, nearly as badly as I need to know whether or not there are a significant number of blown pixels at all. And I've already got focus peaking markings to interpret on the main display, maybe I don't want to interpret blinking/crawling/whatever clipped pixel marks as well.

For example if my display says I've got 20 blown out pixels out of 20 million, I know that my exposure is not too high, who cares where the 20 blown pixels are.

Click to expand...

yes, yes = but we are not talking about 20 blown out pixels here... in live view blinkies give you both where the clipping is happening and how much is happening (because, you know, sometimes you make a decision to have a good amount of clipping intentionally and not only in specular highlights)

exdeejjjaaaa said:

GeorgianBay1939 said:

exdeejjjaaaa said:

RussellInCincinnati said:

Maybe many of us don't really need a full exposure histogram in our viewfinders. For example would prefer an option to just show me 3 numbers in the corner of the viewfinder, simply describing the percentage of blown-out red, green and blue pixels at the current exposure setting. See http://www.dpreview.com/forums/post/52447035

Click to expand...

as noted - get Olympus m43 cameras - they (camera mode, WB, OOC JPG settings, etc) can be perfectly tuned to display raw clipping (not some stupid %, but actually in the exact places in frame where the clipping happens) in ambient light in EVF... otherwise lobby your camera manufacturer to repeat what Olympus did, it is easy

Click to expand...

Do you know

Click to expand...

I did tune setting with rawdigger ( http://www.rawdigger.com/ ) in hand, so yes, I do know that they indicate for me clipping in raw channels w/ very good precision

Click to expand...

Ok, that is good to know .... that the blinkies are triggered by the output from the sensor BEFORE the application of analogue gain (ISO brightening).

But I would like to have BOTH ... I would ALSO like to have an indicator of when I am overbrightening although that is not normally a problem especially with an ISO-less sensor (which my camera does not have.)

I suspect that the blinkies in post-exposure EVF JPEGS in my Panasonics are a combo of blowing (too much overexposure, saturation) AND clipping (overbrightening). But I don't KNOW. : (

BYW, you put me onto RawDigger a few months ago. It is a great piece of s/w that I now use a lot.

Many thanks!

t

GeorgianBay1939 said:

that the blinkies are triggered by the output from the sensor BEFORE the application of analogue gain (ISO brightening).

Click to expand...

that is your words, not mine

exdeejjjaaaa said:

GeorgianBay1939 said:

that the blinkies are triggered by the output from the sensor BEFORE the application of analogue gain (ISO brightening).

Click to expand...

that is your words, not mine

Click to expand...

I misunderstood your previous post, then. And I made the assumption that the "clipping" in Rawdigger is due to oversaturation of sensels due to over exposure ..... completely independent of clipping due to excessive gain, overloading the ADUs.

So is it known what DOES trigger the blinkies in the Oly EVFs?

t

GeorgianBay1939 said:

exdeejjjaaaa said:

GeorgianBay1939 said:

that the blinkies are triggered by the output from the sensor BEFORE the application of analogue gain (ISO brightening).

Click to expand...

that is your words, not mine

Click to expand...

I misunderstood your previous post, then. And I made the assumption that the "clipping" in Rawdigger is due to oversaturation of sensels due to over exposure ..... completely independent of clipping due to excessive gain, overloading the ADUs.

So is it known what DOES trigger the blinkies in the Oly EVFs?

t

Click to expand...

Hi,

I have read also the article you mentionned

exposure-vs-brightening

but this is still not clear... It seems that there are 2 ways to blow the pixel. With over-exposure or with high ISO.

Pratically, I don't think there are 2 different limits. In other words, highlights are clipped is this is over-brightened and you can fix it with one of the three parameters in the exposure triangle.

Do we really have to differentiate over-exposure and over-brightened and have 2 different upper thresholds to deal with, I don't think so.

Well in fact, I am a beginner and I admit this is something very unclear for me.

Christophe

Christof21 said:

GeorgianBay1939 said:

exdeejjjaaaa said:

GeorgianBay1939 said:

that the blinkies are triggered by the output from the sensor BEFORE the application of analogue gain (ISO brightening).

Click to expand...

that is your words, not mine

Click to expand...

I misunderstood your previous post, then. And I made the assumption that the "clipping" in Rawdigger is due to oversaturation of sensels due to over exposure ..... completely independent of clipping due to excessive gain, overloading the ADUs.

So is it known what DOES trigger the blinkies in the Oly EVFs?

t

Click to expand...

Hi,

I have read also the article you mentionned

exposure-vs-brightening

but this is still not clear... It seems that there are 2 ways to blow the pixel. With over-exposure or with high ISO.

Click to expand...

What follows relates to the time when making the exposure on a digital camera:

There is one way to oversaturate, or blow a SENSEL (sensor element) .... overexposure.

Resulting PIXELS (picture elements) can appear blown/clipped due to two effects:

1. the overexposure referred to above .... blown... and/or
2. by overrunning the bit depth of the digital field associated with the pixel during the analog to digital conversion by using too much analog gain (ISO) .... clipped

In both cases there is a loss of tone variation among those pixels which are blown/clipped

Member said:

Pratically, I don't think there are 2 different limits. In other words, highlights are clipped is this is over-brightened and you can fix it with one of the three parameters in the exposure triangle.

Click to expand...

To properly understand this you have to get rid of the concept of the exposure triangle and realize that exposure is only a function of:

1. f/ (actually T stop)
2. Shutter interval
3. Scene luminance

While 1 and 2 above might be affected by an ISO setting, when using P, S or A; ISO does not affect the exposure of a sensor (number of photons/effective area of the sensor)

Member said:

Do we really have to differentiate over-exposure and over-brightened and have 2 different upper thresholds to deal with, I don't think so.

Click to expand...

To get rid of pure blowing of highlights, (oversaturation, caused by overexposure) the photog MUST reduce exposure, (by stopping down and/or by reducing shutter interval for a given scene luminance)

To get rid of pure clipping of highlights, (overrunning of data, caused by too much gain high ISO) the photographer can reduce simply ISO and confirm exposure so that it is maximum before oversaturation.

A M(anual) shooter, in the above case, would just drop ISO and leave exposure (f/, ss) where it was, assuming that it was not blowing sensels.

You can also cause loss of highlights in post processing by compressing highlights and clipping of highlights .... by applying too much brightening (often misnamed "exposure" on the sliders).

Member said:

Well in fact, I am a beginner and I admit this is something very unclear for me.

Click to expand...

It used to be very unclear to me also, because there is a lot of unfortunate noise out there caused by misunderstandings of the role that ISO plays when capturing an exposure. Vocabulary is a little loose also, which makes it difficult.

The time that the above becomes important is when you seriously move from shooting JPEG to shooting RAW and must know the differences in how you set your camera up and how you change your technique.

I found this useful: http://www.josephjamesphotography.com/equivalence/#exposure

Member said:

Christophe

Click to expand...

I hope that this helps.

tom

Member said:

Christof21 wrote: Do we really have to differentiate over-exposure and over-brightened and have 2 different upper thresholds to deal with, I don't think so.

Click to expand...

Agreed. The important thing for the camera to tell a raw photographer is whether or not the raw data written, which might be a number ranging from say 0 to 4094, with 4095 representing overflow, is going to be stored as 4095 or not. The raw photographer doesn't care if some certain current viewfinder preview shows a blanked-out-overshot pixel or not--again as long as the recorded raw brightness isn't over 4094.

RussellInCincinnati said:

RussellInCincinnati said:

Christof21 wrote: Do we really have to differentiate over-exposure and over-brightened and have 2 different upper thresholds to deal with, I don't think so.

Click to expand...

Agreed. The important thing for the camera to tell a raw photographer is whether or not the raw data written, which might be a number ranging from say 0 to 4094, with 4095 representing overflow, is going to be stored as 4095 or not. The raw photographer doesn't care if some certain current viewfinder preview shows a blanked-out-overshot pixel or not--again as long as the recorded raw brightness isn't over 4094.

Click to expand...

I agree --- with base ISO, RAW ETTR shooting where exposure is the clear concern.

However, if I am RAW shooting wide luminance in very low light, I think that I would want to know both IF

1. I had to make a decision to push my shadows by using ISO OR
2. to push my shadows later in post conversion processing.

If I had a pure ISO-less camera it might be clear, but if I either didn't have one or didn't know the read noise/ISO relationship, I would want to know, I think.

Tom

GeorgianBay1939 said:

Christof21 said:

GeorgianBay1939 said:

exdeejjjaaaa said:

GeorgianBay1939 said:

that the blinkies are triggered by the output from the sensor BEFORE the application of analogue gain (ISO brightening).

Click to expand...

that is your words, not mine

Click to expand...

I misunderstood your previous post, then. And I made the assumption that the "clipping" in Rawdigger is due to oversaturation of sensels due to over exposure ..... completely independent of clipping due to excessive gain, overloading the ADUs.

So is it known what DOES trigger the blinkies in the Oly EVFs?

t

Click to expand...

Hi,

I have read also the article you mentionned

exposure-vs-brightening

but this is still not clear... It seems that there are 2 ways to blow the pixel. With over-exposure or with high ISO.

Click to expand...

What follows relates to the time when making the exposure on a digital camera:

There is one way to oversaturate, or blow a SENSEL (sensor element) .... overexposure.

Resulting PIXELS (picture elements) can appear blown/clipped due to two effects:

1. the overexposure referred to above .... blown... and/or
2. by overrunning the bit depth of the digital field associated with the pixel during the analog to digital conversion by using too much analog gain (ISO) .... clipped

In both cases there is a loss of tone variation among those pixels which are blown/clipped

GeorgianBay1939 said:

Pratically, I don't think there are 2 different limits. In other words, highlights are clipped is this is over-brightened and you can fix it with one of the three parameters in the exposure triangle.

Click to expand...

To properly understand this you have to get rid of the concept of the exposure triangle and realize that exposure is only a function of:

1. f/ (actually T stop)
2. Shutter interval
3. Scene luminance

While 1 and 2 above might be affected by an ISO setting, when using P, S or A; ISO does not affect the exposure of a sensor (number of photons/effective area of the sensor)

GeorgianBay1939 said:

Do we really have to differentiate over-exposure and over-brightened and have 2 different upper thresholds to deal with, I don't think so.

Click to expand...

To get rid of pure blowing of highlights, (oversaturation, caused by overexposure) the photog MUST reduce exposure, (by stopping down and/or by reducing shutter interval for a given scene luminance)

To get rid of pure clipping of highlights, (overrunning of data, caused by too much gain high ISO) the photographer can reduce simply ISO and confirm exposure so that it is maximum before oversaturation.

A M(anual) shooter, in the above case, would just drop ISO and leave exposure (f/, ss) where it was, assuming that it was not blowing sensels.

You can also cause loss of highlights in post processing by compressing highlights and clipping of highlights .... by applying too much brightening (often misnamed "exposure" on the sliders).

GeorgianBay1939 said:

Well in fact, I am a beginner and I admit this is something very unclear for me.

Click to expand...

It used to be very unclear to me also, because there is a lot of unfortunate noise out there caused by misunderstandings of the role that ISO plays when capturing an exposure. Vocabulary is a little loose also, which makes it difficult.

The time that the above becomes important is when you seriously move from shooting JPEG to shooting RAW and must know the differences in how you set your camera up and how you change your technique.

I found this useful: http://www.josephjamesphotography.com/equivalence/#exposure

GeorgianBay1939 said:

Christophe

Click to expand...

I hope that this helps.

tom

Click to expand...

Thanks for your interesting answer.

I understand the process you describe. But practically, I still think we can just look at the live histogram and behave as if we had only one threshold and simply apply the exposure triangle. Do we have to distinguish the exposure settings and the ISO to avoid clipping highlights, I never had to do so in practice.

Can your answer have a practical interest or is it only theorical ? I am not talking about post processing.

Christof21 said:

GeorgianBay1939 said:

Christof21 said:

GeorgianBay1939 said:

exdeejjjaaaa said:

GeorgianBay1939 said:

that the blinkies are triggered by the output from the sensor BEFORE the application of analogue gain (ISO brightening).

Click to expand...

that is your words, not mine

Click to expand...

I misunderstood your previous post, then. And I made the assumption that the "clipping" in Rawdigger is due to oversaturation of sensels due to over exposure ..... completely independent of clipping due to excessive gain, overloading the ADUs.

So is it known what DOES trigger the blinkies in the Oly EVFs?

t

Click to expand...

Hi,

I have read also the article you mentionned

exposure-vs-brightening

but this is still not clear... It seems that there are 2 ways to blow the pixel. With over-exposure or with high ISO.

Click to expand...

What follows relates to the time when making the exposure on a digital camera:

There is one way to oversaturate, or blow a SENSEL (sensor element) .... overexposure.

Resulting PIXELS (picture elements) can appear blown/clipped due to two effects:

1. the overexposure referred to above .... blown... and/or
2. by overrunning the bit depth of the digital field associated with the pixel during the analog to digital conversion by using too much analog gain (ISO) .... clipped

In both cases there is a loss of tone variation among those pixels which are blown/clipped

Member said:

Pratically, I don't think there are 2 different limits. In other words, highlights are clipped is this is over-brightened and you can fix it with one of the three parameters in the exposure triangle.

Click to expand...

To properly understand this you have to get rid of the concept of the exposure triangle and realize that exposure is only a function of:

1. f/ (actually T stop)
2. Shutter interval
3. Scene luminance

While 1 and 2 above might be affected by an ISO setting, when using P, S or A; ISO does not affect the exposure of a sensor (number of photons/effective area of the sensor)

Member said:

Do we really have to differentiate over-exposure and over-brightened and have 2 different upper thresholds to deal with, I don't think so.

Click to expand...

To get rid of pure blowing of highlights, (oversaturation, caused by overexposure) the photog MUST reduce exposure, (by stopping down and/or by reducing shutter interval for a given scene luminance)

To get rid of pure clipping of highlights, (overrunning of data, caused by too much gain high ISO) the photographer can reduce simply ISO and confirm exposure so that it is maximum before oversaturation.

A M(anual) shooter, in the above case, would just drop ISO and leave exposure (f/, ss) where it was, assuming that it was not blowing sensels.

You can also cause loss of highlights in post processing by compressing highlights and clipping of highlights .... by applying too much brightening (often misnamed "exposure" on the sliders).

Member said:

Well in fact, I am a beginner and I admit this is something very unclear for me.

Click to expand...

It used to be very unclear to me also, because there is a lot of unfortunate noise out there caused by misunderstandings of the role that ISO plays when capturing an exposure. Vocabulary is a little loose also, which makes it difficult.

The time that the above becomes important is when you seriously move from shooting JPEG to shooting RAW and must know the differences in how you set your camera up and how you change your technique.

I found this useful: http://www.josephjamesphotography.com/equivalence/#exposure

Member said:

Christophe

Click to expand...

I hope that this helps.

tom

Click to expand...

Thanks for your interesting answer.

I understand the process you describe. But practically, I still think we can just look at the live histogram and behave as if we had only one threshold and simply apply the exposure triangle. Do we have to distinguish the exposure settings and the ISO to avoid clipping highlights, I never had to do so in practice.

Click to expand...

Member said:

Can your answer have a practical interest or is it only theorical ? I am not talking about post processing.

Click to expand...

This is further to my http://www.dpreview.com/forums/post/52500925

Day before yesterday I experimented with my new GX7, as I had no knowledge of the nature of its ISO/Read Noise relationship (as given by Sensorgen for most modern cameras).

Before getting more serious about it, I saw an interesting skyscape from the highway near where I live.... This is the scene out of the camera, after RAW/DNG conversion in Lightroom 4.3:


RAW, ETTR, just blowing highlights above 5 trees, Base ISO 200.

And this is the result after pushing/pulling it in LR:


After processing. M(anual), ISO 200, Lots of noise. If I really wanted that shot, I would've bracketed it for HDR fusion. What you see above is brighter than what I say, but I wanted to get all of the black holes opened up.

The exposure was determined by the need to get it as high as possible without blowing the highlights in that patch of sky.

Since that exposure would leave the foreground and most of the sky very dark I was left with the decision of whether

1. to brighten the shadows using gain (ISO) and risk blowing more of the highlights OR
2. to shoot at lowest possible ISO and to brighten the image in post conversion processing.

I decided to shoot at base ISO200 to see how much I could push the GX7 shadows.

IF I had reliable exposure data related to the sensel blowing I might've been more precise about what exposure to use. IF I had reliable brightness data related to data clipping I might've been more precise about what ISO (analog gain) to use.

AS it turned out my exposure setting, based on the Liveview histogram and the chimping of the RGBY post exposure histogram turned out to be ok, as indicated in these Rawdigger checks:




13kP overexposure, 5 MP underexposed




12 bit raw out of camera. I don't think that I'd want to increase this by even 1/3 EV at exposure.



A few minutes later, in my driveway I took some shots of a similar scene and found that I got less (or at least better) noise at 400 ISO and brightening in LR then shooting at 200 ISO and then brightening one stop more in LR. I didn't change the exposure (f/ or ss) at all during these later trials.

I believe that both the LiveView histogram and post exposure histograms are made from their respective jpegs, both of which are affected by the ISO setting. I could have the exposure right, but too much gain (ISO) on the ADC resulting in clipped highlights. I wouldn't know the culprit unless I had both exposure information off of sensor AND brightness information off of the histogram.

I am still learning this stuff, doing lots of reading and experimenting but I think the above is correct. If you think not feel free to correct me!

Tom

GeorgianBay1939 said:

GeorgianBay1939 said:

GeorgianBay1939 said:

Christof21 wrote: Do we really have to differentiate over-exposure and over-brightened and have 2 different upper thresholds to deal with, I don't think so.

Click to expand...

RussellInCincinnati wrote: Agreed. The important thing for the camera to tell a raw photographer is whether or not the raw data written, which might be a number ranging from say 0 to 4094, with 4095 representing overflow, is going to be stored as 4095 or not. The raw photographer doesn't care if some certain current viewfinder preview shows a blanked-out-overshot pixel or not--again as long as the recorded raw brightness isn't over 4094.

Click to expand...

GeorgianBay1939 wrote:...However, if I am RAW shooting wide luminance in very low light, I think that I would want to know both IF

1. I had to make a decision to push my shadows by using ISO OR
2. to push my shadows later in post conversion processing.

If I had a pure ISO-less camera it might be clear, but if I either didn't have one or didn't know the read noise/ISO relationship, I would want to know, I think.

Click to expand...

Can't quite follow what you're saying. If you have a non-ISO-less camera, and you set the ISO "too high", some kind of how-many-raw-file-pixels-will-be-overflowed indicator is all that's needed to let expose to the right folks know how much exposure they can apply. Just like if the camera is ISO-less.

Could you clarify what more info a non-ISO-less camera user needs? Am just a little too thick-headed to follow you thus far. Surely if we were in a room talking would get your idea in just a few seconds.