Interesting new Ricoh GRIII metering - expose to the right finally?

Started 4 months ago | Discussions
OP Raist3d Forum Pro • Posts: 38,673
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

TN Args wrote:

Raist3d wrote:

TN Args wrote:

Raist3d wrote:

SirPeepsalot wrote:

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

The Nikon Z7 I played with yesterday has a metering mode called "highlight weighted", too.

Cool about time more cameras offer the option

I get the impression that real ETTR automatic metering would be too hard to implement.

  • Firstly, images with specular highlights still need them to clip. If you don't let them clip, images will be near-black and full of noise in the main subject areas. The camera could guess which areas are specular and need to clip "no matter what", but it would only be guessing.

You can certainly have the camera meter for the highlights and keep the highlights at all cost (that's why it's a metering option and not a mandate of mode of operation).  Depending on the contrast given the dynamic and shadow range I am seeing on the Ricoh GRiii shadow range, it may be fine for many situations.

  • Secondly, just green channel or all 3 channels? Meters normally only meter green, because green clipping means loss of detail in the image. Red or blue clipping not so much: that leads to a shift in hue, generally speaking. It would surprise me to hear that Ricoh are ETTR-ing on all 3 channels.

That can only be found through testing. Ideally watch over all the three channels.

  • Thirdly, JPEG or raw? I assume you mean raw, but what will that mean when shooting JPEG-only? Or raw + JPEG?

Yes, I meant raw, but raw is the beginning of the JPEG.

I actually like the way Olympus cameras such as my E-M5 II have a metering mode called "High Spot". You put the spot meter area over the brightest tone that you want to retain detail, and it sets that to +2.7 EV, which IIRC is about 95% luminance.

High spot is the first step towards ETTR meeting I have seen, but you have to purposely meter the highlights. Of course, I haven't used the Ricoh GRiii yet, so I am not sure how long the camera takes to expose correctly in this mode.

Anything brighter, e.g. specular highlights, will then clip. The EC dial still works, so if you think your camera can handle more than +2.7EV in raw, then apply EC to the "High Spot" meter.

You could also apply EC to the Ricoh GRiii metering in this mode.

Iliah Borg has a nice article on how to do raw ETTR with precision. It's pretty complex. Would be interesting to see a camera automate all that, with precision, successfully.

cheers

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- I photograph black cats in coal mines at night...
“The further a society drifts from truth the more it will hate those who speak it.” - George Orwell

SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?
1

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

Shawn67 wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

... without clipping and loosing important detail.

Shawn

Yes, that's how I understand ETTR. To the right, but not beyond.

But my concern a lot of posts ago was: Will it protect from (over) saturated colors of flowers and the like?

That too is just (per-channel) highlight protection.

I'm green, but not that green. The question was does it work!?

IDK but no reason for it not to, in a mirrorless camera metering is done by image analysis directly from the sensor unlike in a DSLR where a separate sensor is used which may or may not be color-aware.

Olifaunt Contributing Member • Posts: 762
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

teemodk wrote:

Shawn67 wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

... without clipping and loosing important detail.

Shawn

Yes, that's how I understand ETTR. To the right, but not beyond.

But my concern a lot of posts ago was: Will it protect from (over) saturated colors of flowers and the like?

I am not sure if this is really about exposure. Saturation clipping is more about color profile. For example, the GR-ii standard JPEG color profile almost never over-saturates flowers in my experience; they are rendered very well almost all the time. But saturated colors get clipped all the time in "Positive film" mode, for the same image, same exposure.

This is different from the sky clipping to green (a common issue with GR-ii) because of overexposure of the blue and red channels.

SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

Olifaunt wrote:

teemodk wrote:

Shawn67 wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

... without clipping and loosing important detail.

Shawn

Yes, that's how I understand ETTR. To the right, but not beyond.

But my concern a lot of posts ago was: Will it protect from (over) saturated colors of flowers and the like?

I am not sure if this is really about exposure. Saturation clipping is more about color profile. For example, the GR-ii standard JPEG color profile almost never over-saturates flowers in my experience; they are rendered very well almost all the time. But saturated colors get clipped all the time in "Positive film" mode, for the same image, same exposure.

There can be profile-independent saturation clipping, if all sensor sites under the red CFA masks in an area are saturated you will lose detail regardless of the color profile used.

OP Raist3d Forum Pro • Posts: 38,673
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- I photograph black cats in coal mines at night...
“The further a society drifts from truth the more it will hate those who speak it.” - George Orwell

teemodk Senior Member • Posts: 1,200
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

Olifaunt wrote:

teemodk wrote:

Shawn67 wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

... without clipping and loosing important detail.

Shawn

Yes, that's how I understand ETTR. To the right, but not beyond.

But my concern a lot of posts ago was: Will it protect from (over) saturated colors of flowers and the like?

I am not sure if this is really about exposure. Saturation clipping is more about color profile. For example, the GR-ii standard JPEG color profile almost never over-saturates flowers in my experience; they are rendered very well almost all the time. But saturated colors get clipped all the time in "Positive film" mode, for the same image, same exposure.

This is different from the sky clipping to green (a common issue with GR-ii) because of overexposure of the blue and red channels.

Sure, a lot of detail are clipped by Positive Film, not only highlights but also much of the shadows. It simplifies the image, which may be the reason it is so popular.

I was thinking more about the initial exposure found in the RAW file. As discussed above, many cameras don't expose per channel.

 teemodk's gear list:teemodk's gear list
Oppo R7 Plus
SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights ! The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

OP Raist3d Forum Pro • Posts: 38,673
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all the way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- I photograph black cats in coal mines at night...
“The further a society drifts from truth the more it will hate those who speak it.” - George Orwell

OP Raist3d Forum Pro • Posts: 38,673
On EV compensation...

EV compensation is effectively merely a tool/way to control/achieve ETTR. You could just shoot pure manual with no EV compensation and do the same.

IT's all about exposure, and in EV terms, for a typical ETTR shot at least with scenes of high contrast you will be compensating negatively, not positively.

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- I photograph black cats in coal mines at night...
“The further a society drifts from truth the more it will hate those who speak it.” - George Orwell

SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all he way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

That is just a regular exposure strategy to make use of the entire sensor dynamic range while prioritizing highlight detail over shadow detail. It is just as applicable to film (slide more than negative) as it is to digital.

ETTR is an exposure strategy specific to digital because it takes advantage of how digital information is captured and quantized as more tones can be distinguished in the shadows by exposing them as midtones and in midtones by exposing them as highlights.

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

Well I can't force you to but if you take a moment to think about the purpose it serves and its uniqueness to digital capture you'll see that it is the only the thing that makes it a 'thing' to begin with.

Really it shouldn't be difficult to distinguish between things that are diametrically opposite - if you were to work in full manual mode highlight protection strategy would produce a picture either as intended to be viewed or one that needs shadows to be pushed in post whereas ETTR would produce a picture that needs to be pulled (or curve-adjusted downards) in post.

teemodk Senior Member • Posts: 1,200
Re: Interesting new Ricoh GRIII metering - expose to the right finally?
2

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all he way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

That is just a regular exposure strategy to make use of the entire sensor dynamic range while prioritizing highlight detail over shadow detail. It is just as applicable to film (slide more than negative) as it is to digital.

ETTR is an exposure strategy specific to digital because it takes advantage of how digital information is captured and quantized as more tones can be distinguished in the shadows by exposing them as midtones and in midtones by exposing them as highlights.

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

Well I can't force you to but if you take a moment to think about the purpose it serves and its uniqueness to digital capture you'll see that it is the only the thing that makes it a 'thing' to begin with.

Really it shouldn't be difficult to distinguish between things that are diametrically opposite - if you were to work in full manual mode highlight protection strategy would produce a picture either as intended to be viewed or one that needs shadows to be pushed in post whereas ETTR would produce a picture that needs to be pulled (or curve-adjusted downards) in post.

It seems to me that you did not read the link from Raist3d, which is the same I quoted from earlier. I think you are over complicating ETTR.
You simply make sure all highlights are inside the view of the histogram. This ensues that the sensor get as much light to work with as possible without blowing highlights.
If you need to push or pull the shadows depends on the dynamic range of the scene and sensor.

 teemodk's gear list:teemodk's gear list
Oppo R7 Plus
SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

teemodk wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all he way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

That is just a regular exposure strategy to make use of the entire sensor dynamic range while prioritizing highlight detail over shadow detail. It is just as applicable to film (slide more than negative) as it is to digital.

ETTR is an exposure strategy specific to digital because it takes advantage of how digital information is captured and quantized as more tones can be distinguished in the shadows by exposing them as midtones and in midtones by exposing them as highlights.

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

Well I can't force you to but if you take a moment to think about the purpose it serves and its uniqueness to digital capture you'll see that it is the only the thing that makes it a 'thing' to begin with.

Really it shouldn't be difficult to distinguish between things that are diametrically opposite - if you were to work in full manual mode highlight protection strategy would produce a picture either as intended to be viewed or one that needs shadows to be pushed in post whereas ETTR would produce a picture that needs to be pulled (or curve-adjusted downards) in post.

It seems to me that you did not read the link from Raist3d, which is the same I quoted from earlier. I think you are over complicating ETTR.
You simply make sure all highlights are inside the view of the histogram. This ensues that the sensor get as much light to work with as possible without blowing highlights.
If you need to push or pull the shadows depends on the dynamic range of the scene and sensor.

Actually the wikipedia article over-complicates the issue by trying to cram together arbitrary chunks of digital exposure theory with semi-relevant trivia, it is not technically wrong when it says "exposing to the important highlights (abbr. ETTIH), which is in fact just a slight generalization of ETTR. " but it makes about as much sense as saying that food is a slight generalization of pizza.

Ultimately it all comes down to the question - what do you expect a camera to do with a low-dr scene when the highlights are nowhere near clipping with any of the traditional metering modes ?

Olifaunt Contributing Member • Posts: 762
Re: Interesting new Ricoh GRIII metering - expose to the right finally?
2

SirPeepsalot wrote:

...it makes about as much sense as saying that food is a slight generalization of pizza.

Well, yeah, you have described North American nutrition in a nutshell

https://www.thejournal.ie/us-congress-rules-that-pizza-is-a-vegetable-282033-Nov2011/

SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

Olifaunt wrote:

SirPeepsalot wrote:

...it makes about as much sense as saying that food is a slight generalization of pizza.

Well, yeah, you have described North American nutrition in a nutshell

https://www.thejournal.ie/us-congress-rules-that-pizza-is-a-vegetable-282033-Nov2011/

Well here we have the opposite problem - people have decided that Ricoh GRIII is a pizza based on claimed presence of tomato sauce.

teemodk Senior Member • Posts: 1,200
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all he way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

That is just a regular exposure strategy to make use of the entire sensor dynamic range while prioritizing highlight detail over shadow detail. It is just as applicable to film (slide more than negative) as it is to digital.

ETTR is an exposure strategy specific to digital because it takes advantage of how digital information is captured and quantized as more tones can be distinguished in the shadows by exposing them as midtones and in midtones by exposing them as highlights.

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

Well I can't force you to but if you take a moment to think about the purpose it serves and its uniqueness to digital capture you'll see that it is the only the thing that makes it a 'thing' to begin with.

Really it shouldn't be difficult to distinguish between things that are diametrically opposite - if you were to work in full manual mode highlight protection strategy would produce a picture either as intended to be viewed or one that needs shadows to be pushed in post whereas ETTR would produce a picture that needs to be pulled (or curve-adjusted downards) in post.

It seems to me that you did not read the link from Raist3d, which is the same I quoted from earlier. I think you are over complicating ETTR.
You simply make sure all highlights are inside the view of the histogram. This ensues that the sensor get as much light to work with as possible without blowing highlights.
If you need to push or pull the shadows depends on the dynamic range of the scene and sensor.

Actually the wikipedia article over-complicates the issue by trying to cram together arbitrary chunks of digital exposure theory with semi-relevant trivia, it is not technically wrong when it says "exposing to the important highlights (abbr. ETTIH), which is in fact just a slight generalization of ETTR. " but it makes about as much sense as saying that food is a slight generalization of pizza.

Ultimately it all comes down to the question - what do you expect a camera to do with a low-dr scene when the highlights are nowhere near clipping with any of the traditional metering modes ?

If I set the camera for highlight exposure, I expect it to expose to get all highlights below max value. I would expect the jpeg to make sense as created by the camera, but the RAW shoud stay unmodified.

 teemodk's gear list:teemodk's gear list
Oppo R7 Plus
SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all he way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

That is just a regular exposure strategy to make use of the entire sensor dynamic range while prioritizing highlight detail over shadow detail. It is just as applicable to film (slide more than negative) as it is to digital.

ETTR is an exposure strategy specific to digital because it takes advantage of how digital information is captured and quantized as more tones can be distinguished in the shadows by exposing them as midtones and in midtones by exposing them as highlights.

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

Well I can't force you to but if you take a moment to think about the purpose it serves and its uniqueness to digital capture you'll see that it is the only the thing that makes it a 'thing' to begin with.

Really it shouldn't be difficult to distinguish between things that are diametrically opposite - if you were to work in full manual mode highlight protection strategy would produce a picture either as intended to be viewed or one that needs shadows to be pushed in post whereas ETTR would produce a picture that needs to be pulled (or curve-adjusted downards) in post.

It seems to me that you did not read the link from Raist3d, which is the same I quoted from earlier. I think you are over complicating ETTR.
You simply make sure all highlights are inside the view of the histogram. This ensues that the sensor get as much light to work with as possible without blowing highlights.
If you need to push or pull the shadows depends on the dynamic range of the scene and sensor.

Actually the wikipedia article over-complicates the issue by trying to cram together arbitrary chunks of digital exposure theory with semi-relevant trivia, it is not technically wrong when it says "exposing to the important highlights (abbr. ETTIH), which is in fact just a slight generalization of ETTR. " but it makes about as much sense as saying that food is a slight generalization of pizza.

Ultimately it all comes down to the question - what do you expect a camera to do with a low-dr scene when the highlights are nowhere near clipping with any of the traditional metering modes ?

If I set the camera for highlight exposure, I expect it to expose to get all highlights below max value. I would expect the jpeg to make sense as created by the camera, but the RAW shoud stay unmodified.

Then you have completely missed the point of ETTR, with ETTR a "flat" scene would come out way over-exposed (but not clipped) in RAW and would need to be pulled back to the original presentation intent but with richer tonality and lower noise, the camera may or may not do that automatically for the JPEG.

teemodk Senior Member • Posts: 1,200
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all he way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

That is just a regular exposure strategy to make use of the entire sensor dynamic range while prioritizing highlight detail over shadow detail. It is just as applicable to film (slide more than negative) as it is to digital.

ETTR is an exposure strategy specific to digital because it takes advantage of how digital information is captured and quantized as more tones can be distinguished in the shadows by exposing them as midtones and in midtones by exposing them as highlights.

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

Well I can't force you to but if you take a moment to think about the purpose it serves and its uniqueness to digital capture you'll see that it is the only the thing that makes it a 'thing' to begin with.

Really it shouldn't be difficult to distinguish between things that are diametrically opposite - if you were to work in full manual mode highlight protection strategy would produce a picture either as intended to be viewed or one that needs shadows to be pushed in post whereas ETTR would produce a picture that needs to be pulled (or curve-adjusted downards) in post.

It seems to me that you did not read the link from Raist3d, which is the same I quoted from earlier. I think you are over complicating ETTR.
You simply make sure all highlights are inside the view of the histogram. This ensues that the sensor get as much light to work with as possible without blowing highlights.
If you need to push or pull the shadows depends on the dynamic range of the scene and sensor.

Actually the wikipedia article over-complicates the issue by trying to cram together arbitrary chunks of digital exposure theory with semi-relevant trivia, it is not technically wrong when it says "exposing to the important highlights (abbr. ETTIH), which is in fact just a slight generalization of ETTR. " but it makes about as much sense as saying that food is a slight generalization of pizza.

Ultimately it all comes down to the question - what do you expect a camera to do with a low-dr scene when the highlights are nowhere near clipping with any of the traditional metering modes ?

If I set the camera for highlight exposure, I expect it to expose to get all highlights below max value. I would expect the jpeg to make sense as created by the camera, but the RAW shoud stay unmodified.

Then you have completely missed the point of ETTR,

No, the point is to get max/optimal data out of your sensor without blowing highlights. Whether it is worth the hassle is another debate. I don't get why you insist that we all are missing the point of ETTR.

with ETTR a "flat" scene would come out way over-exposed (but not clipped) in RAW and would need to be pulled back to the original presentation intent but with richer tonality and lower noise,

Yes of course:
    "I would expect the jpeg to make sense as created by the camera"
and:
   "If you need to push or pull the shadows depends on the dynamic range of the scene and sensor."

the camera may or may not do that automatically for the JPEG.

It probably will not.
If I take a picture with my GRII of a flat scene, I get a flat jpeg. I get what I see. With filters I can spice it up and using the RAW, I can do more.

 teemodk's gear list:teemodk's gear list
Oppo R7 Plus
SirPeepsalot Contributing Member • Posts: 550
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all he way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

That is just a regular exposure strategy to make use of the entire sensor dynamic range while prioritizing highlight detail over shadow detail. It is just as applicable to film (slide more than negative) as it is to digital.

ETTR is an exposure strategy specific to digital because it takes advantage of how digital information is captured and quantized as more tones can be distinguished in the shadows by exposing them as midtones and in midtones by exposing them as highlights.

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

Well I can't force you to but if you take a moment to think about the purpose it serves and its uniqueness to digital capture you'll see that it is the only the thing that makes it a 'thing' to begin with.

Really it shouldn't be difficult to distinguish between things that are diametrically opposite - if you were to work in full manual mode highlight protection strategy would produce a picture either as intended to be viewed or one that needs shadows to be pushed in post whereas ETTR would produce a picture that needs to be pulled (or curve-adjusted downards) in post.

It seems to me that you did not read the link from Raist3d, which is the same I quoted from earlier. I think you are over complicating ETTR.
You simply make sure all highlights are inside the view of the histogram. This ensues that the sensor get as much light to work with as possible without blowing highlights.
If you need to push or pull the shadows depends on the dynamic range of the scene and sensor.

Actually the wikipedia article over-complicates the issue by trying to cram together arbitrary chunks of digital exposure theory with semi-relevant trivia, it is not technically wrong when it says "exposing to the important highlights (abbr. ETTIH), which is in fact just a slight generalization of ETTR. " but it makes about as much sense as saying that food is a slight generalization of pizza.

Ultimately it all comes down to the question - what do you expect a camera to do with a low-dr scene when the highlights are nowhere near clipping with any of the traditional metering modes ?

If I set the camera for highlight exposure, I expect it to expose to get all highlights below max value. I would expect the jpeg to make sense as created by the camera, but the RAW shoud stay unmodified.

Then you have completely missed the point of ETTR,

No, the point is to get max/optimal data out of your sensor without blowing highlights. Whether it is worth the hassle is another debate. I don't get why you insist that we all are missing the point of ETTR.

with ETTR a "flat" scene would come out way over-exposed (but not clipped) in RAW and would need to be pulled back to the original presentation intent but with richer tonality and lower noise,

Yes of course:
"I would expect the jpeg to make sense as created by the camera"
and:
"If you need to push or pull the shadows depends on the dynamic range of the scene and sensor."

the camera may or may not do that automatically for the JPEG.

It probably will not.
If I take a picture with my GRII of a flat scene, I get a flat jpeg. I get what I see. With filters I can spice it up and using the RAW, I can do more.

It's not about whether the result will look flat or not it's about whether it will be gray or near-white, the RAW would need to be near-white for ETTR to have taken place which I'm almost sure will not be the case with the GRIII.

teemodk Senior Member • Posts: 1,200
Re: Interesting new Ricoh GRIII metering - expose to the right finally?

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

Raist3d wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot wrote:

teemodk wrote:

SirPeepsalot

Coincidentally - a clean black wall shot with full ETTR should also come out white, so an in-camera ETTR implementation should probably include a safeguard on the maximum amount of stops to overexpose when possible.

Only if you want to use the jpeg directly from the camera. If you want the least amount of noise in the raw image, you want to give it as much light as possible.

At some point the marginal benefit becomes negligible and if you ETTR a uniform dark surface your image will be dominated by haze and a heightened black point all while requiring multiple additional stops of shutter speed.

Might be true, I have never felt the need to photograph black walls
Anyway, it's always the user that is to blame for the outcome, never the camera. Learning to use the equipment is more important than features. But if the new exposure mode actually works it would indeed be very helpful.

There is almost no chance that the GRIII does meaningful ETTR

Well... have you tried it? It’s a new metering mode

Don't have one to try, is there a manual online ? How is it described in there ?

yes and yes It certainly would be good to try a real one to see how well it works My comment comes precisely because I looked at the manual/ but I agree needs to be tested in the real world

Here's the manual:

http://www.ricoh-imaging.co.jp/english/r_dc/support/manual/pdf/grd3/GR_DIGITAL_III_E.pdf

Page 65 lists the traditional Multi, Center-weighted and Spot modes, what were you refering to ?

nm this is the GRD-III (facepalm ricoh)

So this is the GRIII manual:

http://www.ricoh-imaging.co.jp/english/support/man-pdf/gr-3.pdf

And it lists a "highlight-weighted" mode that "Meters exposure in multiple areas of the frame with emphasis on the bright areas."

So yeah sounds like "make sure the highlights are retained even at the cost of underexposure" mode which is only tangentially related to ETTR.

Afaik this is ettr The way Reichman described it a while back/ years ago at least

No, at least it's not the salient part of ETTR - in ETTR the protection of the highlights comes into play only after you apply positive exposure compensation in a scene that would have otherwise left them in tact based on your prefered metering choice, if you need to worry about highlights in the beginning then you are exposing to the LEFT.

Again, this doesn't make sense to me- at least the way it was defined by Reichman. The reason it's called expose to the right is not because of how EV compensation is used, but how the histogram looks (ie. you try to capture the highlights, thus the histogram has data in the high bits). The attempt here is to maximize data capture.

https://en.wikipedia.org/wiki/Exposing_to_the_right

The highlights will always be at the right of the histogram - that's what makes them the highlights !

No, this is missing the point. You want to make the highlights go to the very edge of the right in the histogram, preserving them. It is this key difference what makes ETTR's definition. You can always have highlights all he way *through* the right (clipping) or just "towards the right side" but without touching or close to touching the right side (under exposure).

That is just a regular exposure strategy to make use of the entire sensor dynamic range while prioritizing highlight detail over shadow detail. It is just as applicable to film (slide more than negative) as it is to digital.

ETTR is an exposure strategy specific to digital because it takes advantage of how digital information is captured and quantized as more tones can be distinguished in the shadows by exposing them as midtones and in midtones by exposing them as highlights.

The difference is how closed to midtones and the shadows are to them, so yes applying positive exposure compensation relative to the perceptually expected final result is what ETTR is all about.

This is not what is clearly defined in the link provided. And from that link also note that it does mention Michael Reichman's talk on it.

Sorry but I can't agree with your claims of the definition of what ETTR is.

Well I can't force you to but if you take a moment to think about the purpose it serves and its uniqueness to digital capture you'll see that it is the only the thing that makes it a 'thing' to begin with.

Really it shouldn't be difficult to distinguish between things that are diametrically opposite - if you were to work in full manual mode highlight protection strategy would produce a picture either as intended to be viewed or one that needs shadows to be pushed in post whereas ETTR would produce a picture that needs to be pulled (or curve-adjusted downards) in post.

It seems to me that you did not read the link from Raist3d, which is the same I quoted from earlier. I think you are over complicating ETTR.
You simply make sure all highlights are inside the view of the histogram. This ensues that the sensor get as much light to work with as possible without blowing highlights.
If you need to push or pull the shadows depends on the dynamic range of the scene and sensor.

Actually the wikipedia article over-complicates the issue by trying to cram together arbitrary chunks of digital exposure theory with semi-relevant trivia, it is not technically wrong when it says "exposing to the important highlights (abbr. ETTIH), which is in fact just a slight generalization of ETTR. " but it makes about as much sense as saying that food is a slight generalization of pizza.

Ultimately it all comes down to the question - what do you expect a camera to do with a low-dr scene when the highlights are nowhere near clipping with any of the traditional metering modes ?

If I set the camera for highlight exposure, I expect it to expose to get all highlights below max value. I would expect the jpeg to make sense as created by the camera, but the RAW shoud stay unmodified.

Then you have completely missed the point of ETTR,

No, the point is to get max/optimal data out of your sensor without blowing highlights. Whether it is worth the hassle is another debate. I don't get why you insist that we all are missing the point of ETTR.

with ETTR a "flat" scene would come out way over-exposed (but not clipped) in RAW and would need to be pulled back to the original presentation intent but with richer tonality and lower noise,

Yes of course:
"I would expect the jpeg to make sense as created by the camera"
and:
"If you need to push or pull the shadows depends on the dynamic range of the scene and sensor."

the camera may or may not do that automatically for the JPEG.

It probably will not.
If I take a picture with my GRII of a flat scene, I get a flat jpeg. I get what I see. With filters I can spice it up and using the RAW, I can do more.

It's not about whether the result will look flat or not it's about whether it will be gray or near-white, the RAW would need to be near-white for ETTR to have taken place which I'm almost sure will not be the case with the GRIII.

I didn't say that. But we can just say you are right and stop here. Or you can read what I wrote and move on. We are not so far from each other as you may think.

 teemodk's gear list:teemodk's gear list
Oppo R7 Plus
Keyboard shortcuts:
FForum MMy threads