What I learned from Gollywop -- and what I wonder

Started Mar 26, 2013 | Discussions
gollywop
Veteran MemberPosts: 6,191
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to richarddd, Mar 26, 2013

richarddd wrote:

gollywop wrote:

Pedagydusz wrote:

texinwien wrote:

[...] It will, however, also net you less dynamic range, tonal range and color sensitivity. [...]

I don't understand this part. Why, in ISOless situations, do you get less DR, for example, or color sensitivity? It does not make sense to what I thought I had understood!

The sensor's saturation capacity always drops with increased ISO, and with it the largest signal able to be captured. Thus, even if noise remains constant with ISO, DR will fall. DR is the number of stops in SatCapacity/ReadNoise.

Check out www.sensorg.com to see these relations for a number of different cameras.

That link doesn't work. See the link in my post above, http://forums.dpreview.com/forums/post/51163548

Yeah, sorry, it should be

http://www.sensorgen.info

-- hide signature --

gollywop

Reply   Reply with quote   Complain
gollywop
Veteran MemberPosts: 6,191
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to Macx, Mar 26, 2013

Macx wrote:

gollywop wrote:

bowportes wrote:

As an old fart of a photographer, I thought I knew all about exposure -- you know, aperture, shutter-speed, and ISO. The point was to nail the proper exposure as given by the camera. Behaving like I had ektachrome loaded in a digital camera, I rarely over- or under-exposed unless I had a situation of backlighting or snow. The Panny G-series cameras, by shifting the color of the histogram to yellow (caution) if I used the thumb-wheel to increase/lesson exposure, encouraged me to keep it white -- right where the camera automatically exposes.

This may be fine general practice for JPEG shooters, but it turns out it was wrong for capturing raw.

Gollywop's recent post encouraged me to think instead of maximizing the sensor's exposure to light. Rather than accepting the camera's exposure, I should overexpose images just (but not quite) to the point of clipping highlights at base ISO (ETTR). The histogram may go yellow and the image may appear too light in the viewfinder, but as long as the right side of the histogram has not reached the right edge of its axis, go ahead and overexpose. My raw-processed image will be better for it, in spite of the fact that it doesn't look as good in the EVF at +2/3 exposure as when it's not set to overexpose. Watching the histogram (or blinkies on an Olympus) is your key for how far to amplify exposure.

texinwien's post just under your OP pretty much answers your questions on the button. Now all I have to do is get you to not use the term over-expose as you have above

Over-exposre is when you blow out the sensor at base ISO. You can be excused for a confusion on this score, because in my post Exposure vs. Brightening, I used the term over-saturate for this. But it was agreed by many that the term saturation should be reserved for its meaning in describing color intensity. Thus, in my revision I am reclaiming the term over-exposure to mean the situation when so much light falls on the sensor that some photosites (sensels) are gifted with more electrons than they can hold -- blown, in the vernacular.

So rather than "I should overexpose images just (but not quite) to the point of clipping highlights at base ISO (ETTR)" you would think "I should expose at base ISO to just below the point where raw clipping occurs, ETTR." This is not over-exposure, it's just plain ETTR.

Something that's not as clear to me though. If ISO is at base (160 for most Pannys and 200 for Olympus), the lens is at maximum aperture, and shutter-speed can't be set slower, but the histogram still sits right in the middle -- not overexposed at all -- if I understand correctly, I might get some improvement in image quality by raising the ISO to 400 or 800 in order to (overexpose -> brighten) the image, as long as I'm careful not to (overexpose to the point where I'm clipping highlights -> overexpose). I'd like confirmation that this is correct.

Again here. In the above, I've altered the verbiage that has been bold-faced. The boldface regular is the original text. The boldface italics is my suggested replacement. In the second change, note that overexposure to the point of clipping is redundant.

Don't use "brighten" in this sense, though. You're raising ISO to 400 or 800 to reduce read noise and maximise available camera dynamic range, NOT because it brightens the raw file (which it can't) nor because of what it does to final image after the raw file has been developed, which would be unwanted.

The G5 appears to be a very slightly semi-ISOless camera. Read noise does fall up to 800 or without commensurate loss in DR -- but the read noise is pretty low across the board. So it would appear you could exploit some ISO increases with relative impunity if they otherwise worked to your advantage when you are unable to ETTR.

Another situation that I've commonly encountered in the past is desiring to have the lens wide open in bright daylight. This typically means ISO is at 160 (base), my F1.8 or 2.8 lens is wide open, and shutter speed is as fast as my G5 can take it. In the past, the EVF still sometimes indicated this gave too much light, so I was compelled to reduce my aperture since ISO couldn't be lowered and SS couldn't be increased, and my goal was to achieve proper (exposure -> brightness) for my raw file. But now, if I understand the Gollywop thread correctly, I should be pleased with the (overexposure -> exposure) and leave the aperture wide open, as long as (it doesn't clip highlights -> highlights aren't clipped).

Again, just sayin'.

And again, it's meaningless to use "brightness" to describe a raw file. Your final image may be bright or not, but the raw file isn't, and final image brightness has nothing to do with your camera settings.

I'm in full agreement with everything else

You are indeed correct.  I have misspoken.  One can brighten the raw file (in the sense of increasing its data values) but brightness refers properly only the a final image.  So above with the (exposure -> brightness) it ought either be "brightness of my final image" or "brightening of my raw file" depending upon which he really had in mind.

-- hide signature --

gollywop

Reply   Reply with quote   Complain
texinwien
Senior MemberPosts: 3,117Gear list
Like?
Re: Many thanks to Golly, Bowie, and Tex -- very illuminating!
In reply to Hen3ry, Mar 26, 2013

Hen3ry wrote:

I have to (shamefacedly) admit that Golly's post left me floundering.

No shame! It's just complicated enough to require some concentration, close focus and clear thinking. Not one of those subjects most can listen two with 20% of their attention and be certain they've understood it, afterwards. I've seen it trip up guys who appear to have engineer or research-level knowledge of the important topics.

Now that Bowie has interpreted it and Tex polished it a little, I grok it. Thanks to you both.

Glad it helped. I hesitate to say much about it in case I mess something up. I'm still learning and internalizing it all, too

In fact, it is right on the money of what I have been thinking about and doing more and more of with my E-PL3 -- paying a lot of attention to the histogram and to getting the highlights up to the right end since the E-PL3 doesn’t take kindly at all to under exposure (and is pretty desperate at 800 ISO in my view -- not that that worries me at all).

I can then mess with the histogram (or even -- extreme, edgy stuff -- the curves!!!) in PP to darken the pic a bit or whatever.

It will be interesting to hear how this works for you. Would you plan on shooting RAW then and applying the curves to the RAW? Or still going with JPEG only?

Something my focus on the highlights has made me more conscious of is which highlights I might be clipping. Do they matter?

This is one thing I love about the live view blinkies on the E-M5 (and I guess the other two newest Olympus models). They'll tell you which highlights are in danger of clipping, and you can decide whether you're worried about them or not.

From time to time I had been disappointed by results in "correctly" exposed pix where preserving the highlights had pushed more important stuff into nether world of Oly 12 MPX JPEG shadows.

This could help you out here, especially combined with live-view blinkies that show you which highlights are about to get creamed.

I have now learned to scan the pic for highlights and shadows -- and on occasion, to let those highlights blow so important matter is lifted out of the shadows into the mid tones.

Thanks again all three -- I see more clearly now (but I will continue in JPEG where possible!)

Glad my restatement was clear, and I also hope it passes muster with those who're way more informed than I.

Have fun, and let us know how it works out for you!

tex

 texinwien's gear list:texinwien's gear list
Canon EOS 300D Panasonic Lumix DMC-GM5 Olympus E-M5 II Panasonic Lumix G 20mm F1.7 ASPH Olympus M.Zuiko Digital 45mm 1:1.8 +17 more
Reply   Reply with quote   Complain
Vlad S
Senior MemberPosts: 3,434Gear list
Like?
Both
In reply to bowportes, Mar 26, 2013

bowportes wrote:

Vlad S wrote:

Go ahead with the rules suggested above, but try to test it for your typical scenes. In the past we had threads with test images, and everyone saw what they wanted to see. I think this indicates that the difference is not that big, but everybody has their own criteria what's important and what's not. At any rate, I would encourage you to do your own tests, with your own style of processing, rather than take anyone's word for it.

Vlad

Vlad, I don't know whether the "testing" you are referring to is specific to the question of raising ISO beyond base, or if it refers to the entire idea of going for "full exposure". ??

I saw a thread about Nikon D7000, which is one of the most touted examples of the "ISOless" cameras, and the comment was that the base ISO + PP worked almost as well as high ISO, even though the image looked grossly underexposed, but the chroma noise was a little higher than in the image taken at higher ISO. You need to remember, that all the considerations about the effect of the ISO rest on the assumption "all other things being equal." In reality, the ISO amplification in the camera and the pushing algorithms in the raw converter probably differ in details. If you are trying to eke out every drop of performance, you need to know which of the two algorithms provides better result.

As far as the full exposure is concerned, there is no doubt that maximizing the light lowers the noise, but the practical limitation is that it is quite rare that you will have a lot of headroom without clipping highlights somewhere. It works very well on a foggy day, but in contrasty conditions there may not much of a chance to increase exposure non-destructively.

Most photographers make a conscious decision about clipping some highlights, at least the specular ones, but how far they are willing to go is a very personal choice. In the vast majority of cases covered in this forum, it was 1/3 to 2/3 of a stop above the camera meter. The reduction of noise with such a modest exposure increase can be barely noticeable. Here is an earlier discussion of what techniques work and how well, and the 2nd post in the thread links to some data and discussion of the problem.

Vlad

 Vlad S's gear list:Vlad S's gear list
Olympus OM-D E-M5 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 100-300mm F4-5.6 OIS Panasonic Lumix G 20mm F1.7 ASPH Olympus M.Zuiko Digital ED 9-18mm 1:4.0-5.6 +2 more
Reply   Reply with quote   Complain
Aleo Veuliah
Forum ProPosts: 14,545Gear list
Like?
Re: Good discussion here. :) / No text.
In reply to bowportes, Mar 26, 2013

-- hide signature --

Simplicity is the ultimate sophistication.
Pure logic is the ruin of the spirit.
If plan A doesn't work, the alphabet has 25 more letters, keep calm.
Imagination is more important than knowledge.
God is the tangential point between zero and infinity.
Aleo Photo Site

 Aleo Veuliah's gear list:Aleo Veuliah's gear list
Panasonic Lumix DMC-ZS1 Panasonic Lumix DMC-FZ1000 Panasonic Lumix DMC-GF1 Panasonic Lumix DMC-G2 Panasonic Lumix DMC-GH4 +4 more
Reply   Reply with quote   Complain
gollywop
Veteran MemberPosts: 6,191
Like?
Thanks, Mom. (nt)
In reply to Aleo Veuliah, Mar 26, 2013
-- hide signature --

gollywop

Reply   Reply with quote   Complain
texinwien
Senior MemberPosts: 3,117Gear list
Like?
Re: Both
In reply to Vlad S, Mar 26, 2013

Vlad S wrote:

As far as the full exposure is concerned, there is no doubt that maximizing the light lowers the noise, but the practical limitation is that it is quite rare that you will have a lot of headroom without clipping highlights somewhere. It works very well on a foggy day, but in contrasty conditions there may not much of a chance to increase exposure non-destructively.

I think it depends on what you're comparing against. That is, increase exposure from what baseline? I find there's often significant room to increase exposure non destructively as compared to the exposure I'd get were I to try to get the brightness I desired in the the OOC JPEG (w/ E-M5).

Most photographers make a conscious decision about clipping some highlights, at least the specular ones, but how far they are willing to go is a very personal choice. In the vast majority of cases covered in this forum, it was 1/3 to 2/3 of a stop above the camera meter.

Not to be argumentative, but I would like to know your source for the 'vast majority' statement. I'm seriously not trying to start something - I would like to know more about this, since it's not ringing a bell for me.

The reduction of noise with such a modest exposure increase can be barely noticeable. Here is an earlier discussion of what techniques work and how well, and the 2nd post in the thread links to some data and discussion of the problem.

I'm not sure how helpful the link is (or the post you mentioned with its 3 other links and no other information). That is, unless someone wants to dig through the haystack in order to find the needle - it appears you've linked perhaps 600 posts on the subject (4 discussions, each at 150 posts, although two of the sublinked discussions yield 'page not found' errors for me, at the moment).

Perhaps your statements are more geared to the m43 world, as a whole, but aren't as applicable to some specific models? In which case, I think it makes sense to state that specifically, if there is a significant variance between different m43 models such that ETTR might make more sense with some cameras than it would with others.

tex

 texinwien's gear list:texinwien's gear list
Canon EOS 300D Panasonic Lumix DMC-GM5 Olympus E-M5 II Panasonic Lumix G 20mm F1.7 ASPH Olympus M.Zuiko Digital 45mm 1:1.8 +17 more
Reply   Reply with quote   Complain
Pedagydusz
Veteran MemberPosts: 4,678Gear list
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to gollywop, Mar 26, 2013

OK, fine. I understand it now. Thanks!

-- hide signature --
 Pedagydusz's gear list:Pedagydusz's gear list
Olympus PEN E-PL5 Panasonic Lumix DMC-GX7 Adobe Photoshop CS6
Reply   Reply with quote   Complain
JoeVC
Contributing MemberPosts: 666
Like?
Re: Many thanks to Golly, Bowie, and Tex -- very illuminating!
In reply to Hen3ry, Mar 27, 2013

This has been a very informative post. Thanks for everyone's input.

One other thing that might help in higher ISO situations is to do a white balance calibration against a white card, it could help all three color channels be better exposed.

~Joe

Reply   Reply with quote   Complain
Great Bustard
Forum ProPosts: 25,965
Like?
Please don't do that.
In reply to Hen3ry, Mar 27, 2013

Hen3ry wrote:

I grok it.

As a personal favor, please find another way to say that.  For example, "I understand" will do nicely. 

Reply   Reply with quote   Complain
Hen3ry
Forum ProPosts: 11,699Gear list
Like?
Re: Many thanks to Golly, Bowie, and Tex -- very illuminating!
In reply to texinwien, Mar 27, 2013

texinwien wrote:

Hen3ry wrote:

In fact, it is right on the money of what I have been thinking about and doing more and more of with my E-PL3 -- paying a lot of attention to the histogram and to getting the highlights up to the right end since the E-PL3 doesn’t take kindly at all to under exposure (and is pretty desperate at 800 ISO in my view -- not that that worries me at all).

I can then mess with the histogram (or even -- extreme, edgy stuff -- the curves!!!) in PP to darken the pic a bit or whatever.

It will be interesting to hear how this works for you. Would you plan on shooting RAW then and applying the curves to the RAW? Or still going with JPEG only?

I'll stick with JPEGs for most of whast I do -- they do the job. I am looking at some other stuff and also there are a few situations where I might start switching to RAW+JPEG and see how I go there.

Something my focus on the highlights has made me more conscious of is which highlights I might be clipping. Do they matter?

This is one thing I love about the live view blinkies on the E-M5 (and I guess the other two newest Olympus models). They'll tell you which highlights are in danger of clipping, and you can decide whether you're worried about them or not.

I know what you mean. Panny has them too -- I inadvrtently switched them on, once, on the G1 and went crazy trying to switch them off. I thought the camera had gone faulty! (Well, until I read the manual.)

But I find the blinkies distracting with live subjects, so I switch them off and work out what’s going on, e.g. there's a long lead to the right and the subject if backlit with some sky in the pic? Those highlights are the sky. Decide about ignoring them.

I have now learned to scan the pic for highlights and shadows -- and on occasion, to let those highlights blow so important matter is lifted out of the shadows into the mid tones.

Thanks again all three -- I see more clearly now (but I will continue in JPEG where possible!)

Glad my restatement was clear, and I also hope it passes muster with those who're way more informed than I.

Have fun, and let us know how it works out for you!

Thanks, Tex, I will.

Cheers, geoff

 Hen3ry's gear list:Hen3ry's gear list
Panasonic Lumix DMC-G6 Olympus M.Zuiko Digital 45mm 1:1.8 Olympus M.Zuiko Digital ED 9-18mm 1:4.0-5.6 Panasonic Lumix G Vario 45-150mm F4-5.6 ASPH Mega OIS Panasonic 12-35mm F2.8 +4 more
Reply   Reply with quote   Complain
Vlad S
Senior MemberPosts: 3,434Gear list
Like?
Re: Both
In reply to texinwien, Mar 27, 2013

texinwien wrote:

Most photographers make a conscious decision about clipping some highlights, at least the specular ones, but how far they are willing to go is a very personal choice. In the vast majority of cases covered in this forum, it was 1/3 to 2/3 of a stop above the camera meter.

Not to be argumentative, but I would like to know your source for the 'vast majority' statement. I'm seriously not trying to start something - I would like to know more about this, since it's not ringing a bell for me.

It is my personal observation after years of looking at this stuff. I am not willing to go back and seek every single reference. I think it would be wasteful, considering that everybody has their own shooting style. At the same time, I do not ask to take my word for it, I ask to test it for your own style and application.

The reduction of noise with such a modest exposure increase can be barely noticeable. Here is an earlier discussion of what techniques work and how well, and the 2nd post in the thread links to some data and discussion of the problem.

I'm not sure how helpful the link is (or the post you mentioned with its 3 other links and no other information). That is, unless someone wants to dig through the haystack in order to find the needle - it appears you've linked perhaps 600 posts on the subject (4 discussions, each at 150 posts, although two of the sublinked discussions yield 'page not found' errors for me, at the moment).

The first link is not particularly important now, it was just setting the rules for the test. Here are the other links in a working condition:

Data

Discussion of the data

I am sorry that I can't give you a clear cut answer, but it is a complicated issue, and there was a disagreement of opinions, and several methods for performing ETTR. It is not possible to summarize it in one post without making it look like a dissertation. It would not be much shorter than the original discussion anyway.

I hope you can use at least some of the subject lines as a guide, or you might want to display the thread in the flat view and start from the end. There is a reason why these threads always max out, and it's always pretty much the same discussion every time. You can go back and read the thread in one sitting, or you can watch here how all the same stuff is repeated again.

Perhaps your statements are more geared to the m43 world, as a whole, but aren't as applicable to some specific models? In which case, I think it makes sense to state that specifically, if there is a significant variance between different m43 models such that ETTR might make more sense with some cameras than it would with others.

The thread covered G3 and E-M5 for sure, and I believe there might be some GH2 results. I think this would cover µ4/3 system pretty thoroughly. I don't know about other systems.

Vlad

 Vlad S's gear list:Vlad S's gear list
Olympus OM-D E-M5 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 100-300mm F4-5.6 OIS Panasonic Lumix G 20mm F1.7 ASPH Olympus M.Zuiko Digital ED 9-18mm 1:4.0-5.6 +2 more
Reply   Reply with quote   Complain
bowportes
Senior MemberPosts: 1,907Gear list
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to gollywop, Mar 27, 2013

gollywop wrote:

Macx wrote:

gollywop wrote:

bowportes wrote:

As an old fart of a photographer, I thought I knew all about exposure -- you know, aperture, shutter-speed, and ISO. The point was to nail the proper exposure as given by the camera. Behaving like I had ektachrome loaded in a digital camera, I rarely over- or under-exposed unless I had a situation of backlighting or snow. The Panny G-series cameras, by shifting the color of the histogram to yellow (caution) if I used the thumb-wheel to increase/lesson exposure, encouraged me to keep it white -- right where the camera automatically exposes.

This may be fine general practice for JPEG shooters, but it turns out it was wrong for capturing raw.

Gollywop's recent post encouraged me to think instead of maximizing the sensor's exposure to light. Rather than accepting the camera's exposure, I should overexpose images just (but not quite) to the point of clipping highlights at base ISO (ETTR). The histogram may go yellow and the image may appear too light in the viewfinder, but as long as the right side of the histogram has not reached the right edge of its axis, go ahead and overexpose. My raw-processed image will be better for it, in spite of the fact that it doesn't look as good in the EVF at +2/3 exposure as when it's not set to overexpose. Watching the histogram (or blinkies on an Olympus) is your key for how far to amplify exposure.

texinwien's post just under your OP pretty much answers your questions on the button. Now all I have to do is get you to not use the term over-expose as you have above

Over-exposre is when you blow out the sensor at base ISO. You can be excused for a confusion on this score, because in my post Exposure vs. Brightening, I used the term over-saturate for this. But it was agreed by many that the term saturation should be reserved for its meaning in describing color intensity. Thus, in my revision I am reclaiming the term over-exposure to mean the situation when so much light falls on the sensor that some photosites (sensels) are gifted with more electrons than they can hold -- blown, in the vernacular.

So rather than "I should overexpose images just (but not quite) to the point of clipping highlights at base ISO (ETTR)" you would think "I should expose at base ISO to just below the point where raw clipping occurs, ETTR." This is not over-exposure, it's just plain ETTR.

Something that's not as clear to me though. If ISO is at base (160 for most Pannys and 200 for Olympus), the lens is at maximum aperture, and shutter-speed can't be set slower, but the histogram still sits right in the middle -- not overexposed at all -- if I understand correctly, I might get some improvement in image quality by raising the ISO to 400 or 800 in order to (overexpose -> brighten) the image, as long as I'm careful not to (overexpose to the point where I'm clipping highlights -> overexpose). I'd like confirmation that this is correct.

Again here. In the above, I've altered the verbiage that has been bold-faced. The boldface regular is the original text. The boldface italics is my suggested replacement. In the second change, note that overexposure to the point of clipping is redundant.

Don't use "brighten" in this sense, though. You're raising ISO to 400 or 800 to reduce read noise and maximise available camera dynamic range, NOT because it brightens the raw file (which it can't) nor because of what it does to final image after the raw file has been developed, which would be unwanted.

The G5 appears to be a very slightly semi-ISOless camera. Read noise does fall up to 800 or without commensurate loss in DR -- but the read noise is pretty low across the board. So it would appear you could exploit some ISO increases with relative impunity if they otherwise worked to your advantage when you are unable to ETTR.

Another situation that I've commonly encountered in the past is desiring to have the lens wide open in bright daylight. This typically means ISO is at 160 (base), my F1.8 or 2.8 lens is wide open, and shutter speed is as fast as my G5 can take it. In the past, the EVF still sometimes indicated this gave too much light, so I was compelled to reduce my aperture since ISO couldn't be lowered and SS couldn't be increased, and my goal was to achieve proper (exposure -> brightness) for my raw file. But now, if I understand the Gollywop thread correctly, I should be pleased with the (overexposure -> exposure) and leave the aperture wide open, as long as (it doesn't clip highlights -> highlights aren't clipped).

Again, just sayin'.

And again, it's meaningless to use "brightness" to describe a raw file. Your final image may be bright or not, but the raw file isn't, and final image brightness has nothing to do with your camera settings.

I'm in full agreement with everything else

You are indeed correct.  I have misspoken.  One can brighten the raw file (in the sense of increasing its data values) but brightness refers properly only the a final image.  So above with the (exposure -> brightness) it ought either be "brightness of my final image" or "brightening of my raw file" depending upon which he really had in mind.

Well, isn't it kind of meaningless to talk about the "brightness" of any file, including a JPEG? All of them are simply digital containers with some mix of data that have to be interpreted by software to generate an image, however bright the final visual interpretation might be. If I open a raw file in LR, it gets interpreted and presented as an image of greater or lesser brightness just like a JPEG would. The same is true for opening JPEG data. It's really only once it is printed on paper that the physical image has a certain level of brightness/luminance.

You've really got an uphill struggle for this purity of language thing. Even those of us who have started to get the point use the language incorrectly. Beyond us, you will have to battle:

Adobe: Lightroom uses "Exposure" for its brightness slider

Olympus: Their Brightness slider in Olympus Viewer 2 is called "exposure compensation"

Etc. Etc.

What is more, all of my raw developers use in-camera settings to determine how brightly to first display the image. So if I over-expose the image (over-brighten in your sense) by shifting the camera's exposure compensation wheel, then they appear that way in live-view, and also in the raw processor when the file is first loaded (unless I set the raw editor to display them differently). In each of these settings, the use of "exposure" for "luminance" in photography is so entrenched that it almost makes me wonder whether you should concede the term "exposure" to communicate brightness, which is how people think of it, and then find another term to signify sensor charge/load, which most people don't think of at all.

Anyway, you've got a great communication challenge ahead to straighten this out, but you've taught me a fair bit, so I'm rooting for you.

 bowportes's gear list:bowportes's gear list
Fujifilm X-M1 Panasonic Lumix DMC-G5 Fujifilm X-T1 Olympus M.Zuiko Digital 45mm 1:1.8 Panasonic Lumix G X Vario PZ 14-42mm F3.5-5.6 ASPH OIS +9 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 20,707Gear list
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to texinwien, Mar 27, 2013

texinwien wrote:

LincolnB wrote:

bowportes wrote:

Something that's not as clear to me though. If ISO is at base (160 for most Pannys and 200 for Olympus), the lens is at maximum aperture, and shutter-speed can't be set slower, but the histogram still sits right in the middle -- not overexposed at all -- if I understand correctly, I might get some improvement in image quality by raising the ISO to 400 or 800 in order to overexpose the image, as long as I'm careful not to overexpose to the point where I'm clipping highlights. I'd like confirmation that this is correct.

You might get less highlight clipping and less shadow noise but you'll get more sensor noise overall. There is a trade-off.

I believe that is incorrect. Assuming exposure stays the same (same luminance, aperture and shutter speed), going from a lower ISO with higher read noise to a higher ISO with lower read noise (say, from ISO 200 to ISO 400 on the E-M5) will net you less noise.

You are right in this case tex.

It will, however, also net you less dynamic range, tonal range and color sensitivity. That's where the tradeoff comes into play with this strategy. You have to decide whether the drop in read noise is worth the tradeoff in the loss of dynamic range, tonal range and color sensitivity that comes along with it (when bumping ISO from 200 to 400 on the E-M5 and keeping exposure the same, as an example).

But you are wrong here. Going from ISO 200 to 400 on the E-M5 under these circumstances (where, importantly, you cannot increase exposure any further and you are still at least one EV short of the clipping point if you stay at base ISO) will bring you only advantages (significantly lower read noise, and thus significantly lower shadow noise), no disadvantages.

What you must keep in mind here is that the DxO figures for dynamic range, tonal range, and color sensitivity are for a sensor exposed up to the clipping point. If you are below that point, the figures are correspondingly lower. What you must also keep in mind is that things like dynamic range, tonal range, and color response are not distinct from noise. They are just different ways to index noise. We need multiple measures of noise because, as with pretty much any attempt to reduce a complex reality into a single number, none of them tells us everything we might want to know.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS +28 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 20,707Gear list
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to bowportes, Mar 27, 2013

bowportes wrote:

Michael Meissner wrote:

richarddd wrote:

Another good source of information is this thread http://forums.dpreview.com/forums/post/42595072

See the links in the first post. You may have to click on the camera name in the right pane. The linked chart clearly shows the noise issue on the E-M5 between ISO 200 and 400.

Various posts over the years have said that the Panasonic sensor used by the earlier Pens tends to have less noise when you use the full stop ISO's (200, 400, 800, 1600, 3200) rather than the intermediate versions. There is an option to only use whole stops in auto ISO. It sounds like the same is probably true of the Sony sensor used in the E-M5, E-PL5, and E-PM2.

I may remember poorly, but I think I read once (somewhere in the forum) that 160, 320, 640, 1280, etc. Were the appropriate full stops on the GH2. And I had inferred from that that those were the ones to focus upon on the G5. If this is wrong, I'd like clear word on that

You are largely (though not exactly) right about the GH2 whose only "real" ISOs are 160, 320, 640, and 800. The intermediate ISOs between these are accomplished by means of digitally scaling the RAW values upwards rather than by analog amplification. The same is true about any ISO above 800. It follows that using any ISOs other than those I listed will not bring you any benefits above using the next lower ISO among those listed. They will be the same for read noise but worse for the risk of highlight clipping.

Whether the G5 behaves the same way, I don't know. And I am not sure anyone else here does either at this point. You can find out for yourself by using RawDigger and inspect the histogram for different ISOs. If digital scaling is employed, there will be gaps in the histogram (i.e., certain specific ADU values will not occur). See here for example:

http://forums.dpreview.com/forums/post/37798663

Another interesting case here is the GH3, which uses analog amplification all the way up to ISO 6400 but neither analog amplification nor digital scaling from that point on. The RAW values recorded will be the same at 6400 as at 25600. The only thing that happens on the camera when you crank up ISO from 6400 is that the metering changes, that the scaling factor used in the OOC jpeg conversion changes, and that the instruction to the RAW converter embedded in the RAWs changes to tell it to change the scaling factor in the same way as the in-camera jpeg engine.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS +28 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 20,707Gear list
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to Michael Meissner, Mar 27, 2013

Michael Meissner wrote:

richarddd wrote:

Another good source of information is this thread http://forums.dpreview.com/forums/post/42595072

See the links in the first post. You may have to click on the camera name in the right pane. The linked chart clearly shows the noise issue on the E-M5 between ISO 200 and 400.

Various posts over the years have said that the Panasonic sensor used by the earlier Pens tends to have less noise when you use the full stop ISO's (200, 400, 800, 1600, 3200) rather than the intermediate versions.

Concerning the actual behavior of at least one Panasonic sensor (GH2), see here:

http://forums.dpreview.com/forums/post/51167049

There is an option to only use whole stops in auto ISO.  It sounds like the same is probably true of the Sony sensor used in the E-M5, E-PL5, and E-PM2.

The story on the E-M5 (and most likely E-PL5, and E-PM2) is a slightly different one. See this thread:

http://forums.dpreview.com/forums/post/41988325

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS +28 more
Reply   Reply with quote   Complain
texinwien
Senior MemberPosts: 3,117Gear list
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to Anders W, Mar 27, 2013

Anders W wrote:

texinwien wrote:

It will, however, also net you less dynamic range, tonal range and color sensitivity. That's where the tradeoff comes into play with this strategy. You have to decide whether the drop in read noise is worth the tradeoff in the loss of dynamic range, tonal range and color sensitivity that comes along with it (when bumping ISO from 200 to 400 on the E-M5 and keeping exposure the same, as an example).

But you are wrong here. Going from ISO 200 to 400 on the E-M5 under these circumstances (where, importantly, you cannot increase exposure any further and you are still at least one EV short of the clipping point if you stay at base ISO) will bring you only advantages (significantly lower read noise, and thus significantly lower shadow noise), no disadvantages.

Exactly - thanks for pointing this error out and offering a clear and concise correction. I don't want to be guilty of muddying the waters on an already complex topic.

What you must keep in mind here is that the DxO figures for dynamic range, tonal range, and color sensitivity are for a sensor exposed up to the clipping point. If you are below that point, the figures are correspondingly lower.

Funny enough, I pointed out this same error in another poster's reasoning last week. Now I'm guilty of it, myself. Not sure what it is about this topic that makes it such a minefield  

What you must also keep in mind is that things like dynamic range, tonal range, and color response are not distinct from noise. They are just different ways to index noise. We need multiple measures of noise because, as with pretty much any attempt to reduce a complex reality into a single number, none of them tells us everything we might want to know.

Thanks for the helpful explanation.

tex

 texinwien's gear list:texinwien's gear list
Canon EOS 300D Panasonic Lumix DMC-GM5 Olympus E-M5 II Panasonic Lumix G 20mm F1.7 ASPH Olympus M.Zuiko Digital 45mm 1:1.8 +17 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 20,707Gear list
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to texinwien, Mar 27, 2013

texinwien wrote:

Anders W wrote:

texinwien wrote:

It will, however, also net you less dynamic range, tonal range and color sensitivity. That's where the tradeoff comes into play with this strategy. You have to decide whether the drop in read noise is worth the tradeoff in the loss of dynamic range, tonal range and color sensitivity that comes along with it (when bumping ISO from 200 to 400 on the E-M5 and keeping exposure the same, as an example).

But you are wrong here. Going from ISO 200 to 400 on the E-M5 under these circumstances (where, importantly, you cannot increase exposure any further and you are still at least one EV short of the clipping point if you stay at base ISO) will bring you only advantages (significantly lower read noise, and thus significantly lower shadow noise), no disadvantages.

Exactly - thanks for pointing this error out and offering a clear and concise correction. I don't want to be guilty of muddying the waters on an already complex topic.

What you must keep in mind here is that the DxO figures for dynamic range, tonal range, and color sensitivity are for a sensor exposed up to the clipping point. If you are below that point, the figures are correspondingly lower.

Funny enough, I pointed out this same error in another poster's reasoning last week. Now I'm guilty of it, myself. Not sure what it is about this topic that makes it such a minefield  

I don't really think that's a mystery either. The subject is simply complex in the sense that there are many variables/parameters involved and there is quite a bit to know, and keep in mind, about each of them. Hence, the likelihood of logical slip-ups increases signficantly.

Practice makes perfect though. That's why it is important to think and talk about it. Some people are wont to say that "theoretical" discussions of the present kind are of no help in the field. I think it's exactly the other way. Only by thinking and talking about it will you eventually know exactly what to do in the field, and do so intuitively and instantly. For example, I know "without thinking" when it's the proper time to switch from ISO 200 to ISO 400 on my E-M5 when I am out taking pictures. But I do so only because I have exercised my thinking about the matter quite a bit beforehand.

What you must also keep in mind is that things like dynamic range, tonal range, and color response are not distinct from noise. They are just different ways to index noise. We need multiple measures of noise because, as with pretty much any attempt to reduce a complex reality into a single number, none of them tells us everything we might want to know.

Thanks for the helpful explanation.

tex

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS +28 more
Reply   Reply with quote   Complain
Detail Man
Forum ProPosts: 15,718
Like?
Re: GH2 Information
In reply to Anders W, Mar 27, 2013

Anders W wrote:

bowportes wrote:

Michael Meissner wrote:

richarddd wrote:

Another good source of information is this thread http://forums.dpreview.com/forums/post/42595072

See the links in the first post. You may have to click on the camera name in the right pane. The linked chart clearly shows the noise issue on the E-M5 between ISO 200 and 400.

Various posts over the years have said that the Panasonic sensor used by the earlier Pens tends to have less noise when you use the full stop ISO's (200, 400, 800, 1600, 3200) rather than the intermediate versions. There is an option to only use whole stops in auto ISO. It sounds like the same is probably true of the Sony sensor used in the E-M5, E-PL5, and E-PM2.

I may remember poorly, but I think I read once (somewhere in the forum) that 160, 320, 640, 1280, etc. Were the appropriate full stops on the GH2. And I had inferred from that that those were the ones to focus upon on the G5. If this is wrong, I'd like clear word on that

You are largely (though not exactly) right about the GH2 whose only "real" ISOs are 160, 320, 640, and 800.

I don't think that it is clear that ISO=800 on the GH2 is unmanipulated. See bg2b's OP here:

Real ISOs are probably just 160, 320, and 640. 200, 400, and 800 show odd artifacts (gaps or near gaps) in the RAW histograms. 250, 500, and 1000 show them even more. Everything beyond ISO 1000 is completely artificial; they're obviously just digital scalings of lower ISOs.

http://forums.dpreview.com/forums/post/37785576

This is kenw's RMS Read/Dark Noise data:

1/4000 NR-2:

160 0.73554
200 1.117
250 1.5487
320 1.3875
400 1.8514
500 2.4756
640 2.6903
800 3.513
1000 5.4109
1250 5.7917
1600 7.9696
2000 13.557
2500 13.168
3200 16.576
4000 25.919
5000 28.305
6400 35.017
8000 40.398
10000 54.246
12800 68.36

http://forums.dpreview.com/forums/post/41131538

I calculated the number of EV that certain ISOs (from rated ISO=160 through ISO=800) deviate from a straight line projection (in EV) from that which would be derived from multiplying the Read/Dark Noise data by the ratio of the Saturation ISO divded by Saturation ISO=167.

Here are results using DxOMark Saturation ISO (extrapolated SatISO values for ISO=320 and 640):

Saturation ISO -------- Read/Dark Noise deviation in EV

The intermediate ISOs between these are accomplished by means of digitally scaling the RAW values upwards rather than by analog amplification. The same is true about any ISO above 800. It follows that using any ISOs other than those I listed will not bring you any benefits above using the next lower ISO among those listed. They will be the same for read noise but worse for the risk of highlight clipping.

Whether the G5 behaves the same way, I don't know. And I am not sure anyone else here does either at this point. You can find out for yourself by using RawDigger and inspect the histogram for different ISOs. If digital scaling is employed, there will be gaps in the histogram (i.e., certain specific ADU values will not occur). See here for example:

http://forums.dpreview.com/forums/post/37798663

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 20,707Gear list
Like?
Re: GH2 Information
In reply to Detail Man, Mar 27, 2013

Detail Man wrote:

Anders W wrote:

bowportes wrote:

Michael Meissner wrote:

richarddd wrote:

Another good source of information is this thread http://forums.dpreview.com/forums/post/42595072

See the links in the first post. You may have to click on the camera name in the right pane. The linked chart clearly shows the noise issue on the E-M5 between ISO 200 and 400.

Various posts over the years have said that the Panasonic sensor used by the earlier Pens tends to have less noise when you use the full stop ISO's (200, 400, 800, 1600, 3200) rather than the intermediate versions. There is an option to only use whole stops in auto ISO. It sounds like the same is probably true of the Sony sensor used in the E-M5, E-PL5, and E-PM2.

I may remember poorly, but I think I read once (somewhere in the forum) that 160, 320, 640, 1280, etc. Were the appropriate full stops on the GH2. And I had inferred from that that those were the ones to focus upon on the G5. If this is wrong, I'd like clear word on that

You are largely (though not exactly) right about the GH2 whose only "real" ISOs are 160, 320, 640, and 800.

I don't think that it is clear that ISO=800 on the GH2 is unmanipulated. See bg2b's OP here:

Could be that you are right about that. Logically, it seems that the series of "real" ISOs should end at 640 rather than 800.  But if so there appear to be some unresolved mysteries here. I remember asking Ken W about this in a prior thread and he found the ISO 800 distribution to be perfectly continuous, without the gaps we would expect as a result of digital scaling. On the other hand, he didn't find that going from 640 to 800 had any tangible benefits in terms of reduced read noise. See here:

http://forums.dpreview.com/forums/post/40403509

Real ISOs are probably just 160, 320, and 640. 200, 400, and 800 show odd artifacts (gaps or near gaps) in the RAW histograms. 250, 500, and 1000 show them even more. Everything beyond ISO 1000 is completely artificial; they're obviously just digital scalings of lower ISOs.

http://forums.dpreview.com/forums/post/37785576

This is kenw's RMS Read/Dark Noise data:

1/4000 NR-2:

160 0.73554
200 1.117
250 1.5487
320 1.3875
400 1.8514
500 2.4756
640 2.6903
800 3.513
1000 5.4109
1250 5.7917
1600 7.9696
2000 13.557
2500 13.168
3200 16.576
4000 25.919
5000 28.305
6400 35.017
8000 40.398
10000 54.246
12800 68.36

http://forums.dpreview.com/forums/post/41131538

I calculated the number of EV that certain ISOs (from rated ISO=160 through ISO=800) deviate from a straight line projection (in EV) from that which would be derived from multiplying the Read/Dark Noise data by the ratio of the Saturation ISO divded by Saturation ISO=167.

Here are results using DxOMark Saturation ISO (extrapolated SatISO values for ISO=320 and 640):

Saturation ISO -------- Read/Dark Noise deviation in EV

The intermediate ISOs between these are accomplished by means of digitally scaling the RAW values upwards rather than by analog amplification. The same is true about any ISO above 800. It follows that using any ISOs other than those I listed will not bring you any benefits above using the next lower ISO among those listed. They will be the same for read noise but worse for the risk of highlight clipping.

Whether the G5 behaves the same way, I don't know. And I am not sure anyone else here does either at this point. You can find out for yourself by using RawDigger and inspect the histogram for different ISOs. If digital scaling is employed, there will be gaps in the histogram (i.e., certain specific ADU values will not occur). See here for example:

http://forums.dpreview.com/forums/post/37798663

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS +28 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads