EXR Confusion not fully resolved

Started 11 months ago | Discussions
Axel Vercauteren
Forum MemberPosts: 92
Like?
EXR Confusion not fully resolved
11 months ago

Elyharbour wrote:

alexisgreat wrote:

The moral of this story is that if you really want hardware EXR highlight protection you must manually select ISO 100 or 200 at M size DR 200 or 400. DO NOT USE AUTO ISO or you will get the same kind of highlight protection that L size uses, regardless of the ISO selected by the camera. Since AUTO ISO is not selectable in M mode, this conflict never happens.

It seems nobody has reacted to this statement (or have I missed it?). This is very important, as I suspect many of us have Auto ISO selected on the assumption that when low ISOs are chosen through Auto, EXR hardware DR is implemented in M size with DR400.

For shooting JPEG only Kim Letkeman recommends Auto ISO. This would be rather pointless if hardware EXR is not involved, which I can't imagine to be true but you never know with EXR, do you? Perhaps Kim or another expert can shed some light on this.

Is Alex right? Could Kim possibly be wrong? I suspect many of us would like to know.

Kim Letkeman
Forum ProPosts: 33,140Gear list
Like?
Re: EXR Confusion not fully resolved
In reply to Axel Vercauteren, 11 months ago

Axel Vercauteren wrote:

Elyharbour wrote:

alexisgreat wrote:

The moral of this story is that if you really want hardware EXR highlight protection you must manually select ISO 100 or 200 at M size DR 200 or 400. DO NOT USE AUTO ISO or you will get the same kind of highlight protection that L size uses, regardless of the ISO selected by the camera. Since AUTO ISO is not selectable in M mode, this conflict never happens.

It seems nobody has reacted to this statement (or have I missed it?). This is very important, as I suspect many of us have Auto ISO selected on the assumption that when low ISOs are chosen through Auto, EXR hardware DR is implemented in M size with DR400.

For shooting JPEG only Kim Letkeman recommends Auto ISO. This would be rather pointless if hardware EXR is not involved, which I can't imagine to be true but you never know with EXR, do you? Perhaps Kim or another expert can shed some light on this.

Is Alex right? Could Kim possibly be wrong? I suspect many of us would like to know.

Kim could not possibly be wrong, trust me on that

People spend an inordinate amount of time trying to find the perfect combination of settings for every possible situation, which will absolutely ensure that they almost never get a good image because they are too busy trying to remember the perfect settings. Seriously.

Others are certain that shooting at L size all the time is the only way because with trivial examples it actually looks better. But that is also hopeless as a general strategy, as it never uses EXR technology.

So my recommendation, forged by extensive testing of the F70EXR in 2009 and not changed since then -- despite testing many samples of every EXR sensor and generation of body -- because the basic issue has never changed. My perspective on the hardware versus software EXR issue is:

  • Use P mode with auto ISO and DR400 to get hardware DR when possible and whatever other DR solution Fuji uses otherwise (and I am not entirely convinced that it is purely the software DR mechanism)
  • If you shoot L size, you have to tolerate DR200 in bright light so that you can avoid the double whammy of raised ISO and noise reduction on smaller pixels ... 
  • Once ISO has gone above 400, the Fuji does whatever it does. In my experience, the DR400 setting remains optimal because as light dims, the presence of artificial lighting in the frame still looks considerably better with the smoother rolloff of DR400

All of what says that DR400 in P mode with AUTO ISO remains optimal as a general purpose setting. And for most people, it is all they ever need to make the best use of the best mode on an EXR camera.

 Kim Letkeman's gear list:Kim Letkeman's gear list
Nikon Coolpix 990 Fujifilm FinePix F200EXR Fujifilm FinePix F770EXR Panasonic Lumix DMC-GH2 Panasonic Lumix DMC-GF3 +16 more
Reply   Reply with quote   Complain
Timur Born
Senior MemberPosts: 4,230
Like?
Re: EXR Confusion not fully resolved
In reply to Axel Vercauteren, 11 months ago

My answer was in the other thread already:

On the X10 you can use AUTO ISO with EXR DR, no problem. Even the combination of AUTO ISO with AUTO DR can result in EXR DR images (albeit AUTO ISO ever so slightly worsens the already present overexposure bug of AUTO DR).

-- hide signature --

Red flash eyes save lives and eye-sight!
http://en.wikipedia.org/wiki/Retinoblastoma

Reply   Reply with quote   Complain
alexisgreat
Veteran MemberPosts: 4,316Gear list
Like?
Re: EXR Confusion not fully resolved
In reply to Kim Letkeman, 11 months ago

Kim, I think most of my posts in that other thread got lost amid the confusion- basically what I found is that in S Priority, I could shoot M size DR 400 AUTO ISO 3200 and choose 1/4,000 sec shutter speeds in very bright mid day conditions and the camera chose ISO 100. If the EXIF information is correct, this cant be because of Hardware based EXR. Since then, I've tested other shutter speed combos and I've found a way both of us can be right

So there are obvious differences between L size DR 400 and M size DR 400 at more normally used shutter speeds (M size DR 400 maintains more detail in the shadows by about 1/3 EV.) It's only at 1/>1,000 sec shutter speeds that we run into the issue of how is the camera giving us 2 stops of highlight protection when we've run out of shutter speed?

So given that EXR hardware works properly with M size DR 400 at shutter speeds of 1/1,000 sec and slower, my new working theory is that only at shutter speeds faster than this does hardware based highlight protection fail, since there is no possible way for it to work at M size DR 400 ISO 100 1/4,000 sec shutter speeds. The two ways it might be failing is that 1) it bins pixels instead so you get EXR SN even though the camera is set for M size DR 400 and 2) it actually does take two exposures and combine them but both exposures are exactly the same- both 1/4,000 sec shutter speed. I dont think even software based highlight protection works at these extreme shutter speeds, so these images would effectively be DR 100, despite what the EXIF says. I dont think the shutter speed or the ISO are wrong, because in manual mode (where you cant use AUTO ISO) I went as far as you can go with M size DR 400, and the exposures were still 2 stops overexposed compared to my S Priority settings. I'd need a 2 stop ND filter to be able to get a similar exposure in Manual mode that I was getting in S Priority.

I believe Photoreddi was able to replicate my findings, but it would be wonderful if you could take a look at this too- please set your camera(s) to S priority M size DR 400 AUTO ISO 3200 in very bright light and the shutter speed to 1/4,000 sec and see if the camera(s) choose(s) ISO 100. And then we can try to figure out if my new working theory is right by comparing those same settings with DR 100.

 alexisgreat's gear list:alexisgreat's gear list
Olympus C-7070 Wide Zoom Fujifilm FinePix HS20 EXR Fujifilm FinePix HS50 EXR Olympus E-520 Olympus Zuiko Digital ED 14-42mm 1:3.5-5.6 +1 more
Reply   Reply with quote   Complain
alexisgreat
Veteran MemberPosts: 4,316Gear list
Like?
Re: EXR Confusion not fully resolved
In reply to Timur Born, 11 months ago

Timur, I was trying to respond to your last post in the other thread but it has maxed out- if you read my other post in this thread you'll see we're on the same page- I think the camera might actually be shooting DR 100 even though M size DR 400 with AUTO 3200 when ISO 100 was selected by the camera when I set the shutter speed to 1/4,000 sec in very bright mid day sun. It could be what some computer programs do when they cannot execute a command properly because the computations are outside their range- the camera did take two exposures, but both of them were identical 1/4,000 sec shutter speed, or it could be that it binned pixels, EXR SN-style. Either way, it would be DR 100 not software or hardware enhanced DR protection. I dont think the shutter speed or the ISO were wrong because S Priority M size "DR 400" AUTO ISO 3200 (camera selected ISO 100) with 1/4,000 sec shutter speed looked about 2 stops dimmer than M mode M size DR 400 ISO 100 with 1/1,000 sec shutter speed. When the program is within its computational range, there are no issues- that is, hardware EXR DR is being employed at M size DR 400 with AUTO ISO 3200 (camera selects ISO 100) as long as the shutter speed is 1/1,000 sec or slower.  But it does not stop you from using shutter speeds faster than this with the same settings (how is the camera supposed to know in advance that it will select ISO 100?) and so the computational paradox is resolved in such a way that DR 100 is used instead of DR 400.

 alexisgreat's gear list:alexisgreat's gear list
Olympus C-7070 Wide Zoom Fujifilm FinePix HS20 EXR Fujifilm FinePix HS50 EXR Olympus E-520 Olympus Zuiko Digital ED 14-42mm 1:3.5-5.6 +1 more
Reply   Reply with quote   Complain
Kim Letkeman
Forum ProPosts: 33,140Gear list
Like?
Re: EXR Confusion not fully resolved
In reply to alexisgreat, 11 months ago

alexisgreat wrote:

Kim, I think most of my posts in that other thread got lost amid the confusion- basically what I found is that in S Priority, I could shoot M size DR 400 AUTO ISO 3200 and choose 1/4,000 sec shutter speeds in very bright mid day conditions and the camera chose ISO 100. If the EXIF information is correct, this cant be because of Hardware based EXR. Since then, I've tested other shutter speed combos and I've found a way both of us can be right

So there are obvious differences between L size DR 400 and M size DR 400 at more normally used shutter speeds (M size DR 400 maintains more detail in the shadows by about 1/3 EV.) It's only at 1/>1,000 sec shutter speeds that we run into the issue of how is the camera giving us 2 stops of highlight protection when we've run out of shutter speed?

So given that EXR hardware works properly with M size DR 400 at shutter speeds of 1/1,000 sec and slower, my new working theory is that only at shutter speeds faster than this does hardware based highlight protection fail, since there is no possible way for it to work at M size DR 400 ISO 100 1/4,000 sec shutter speeds. The two ways it might be failing is that 1) it bins pixels instead so you get EXR SN even though the camera is set for M size DR 400 and 2) it actually does take two exposures and combine them but both exposures are exactly the same- both 1/4,000 sec shutter speed. I dont think even software based highlight protection works at these extreme shutter speeds, so these images would effectively be DR 100, despite what the EXIF says. I dont think the shutter speed or the ISO are wrong, because in manual mode (where you cant use AUTO ISO) I went as far as you can go with M size DR 400, and the exposures were still 2 stops overexposed compared to my S Priority settings. I'd need a 2 stop ND filter to be able to get a similar exposure in Manual mode that I was getting in S Priority.

I believe Photoreddi was able to replicate my findings, but it would be wonderful if you could take a look at this too- please set your camera(s) to S priority M size DR 400 AUTO ISO 3200 in very bright light and the shutter speed to 1/4,000 sec and see if the camera(s) choose(s) ISO 100. And then we can try to figure out if my new working theory is right by comparing those same settings with DR 100.

I suppose I was not clear ... on cameras without a real aperture, I use P mode because of the risk of ruining images by hitting various limitations of exposure.

Which means that the camera chooses everything. In very right light, it always chooses ISO 100 and some combination of shutter and aperture to get a good exposure. As light dims, it raises ISO as necessary, and after 400 ISO I still get decent DR400 results, which means that I have no reason to shoot any other way.

I didn't really understand the upside of shooting in the complex fashion you described. Do you have an elevator pitch for that?

 Kim Letkeman's gear list:Kim Letkeman's gear list
Nikon Coolpix 990 Fujifilm FinePix F200EXR Fujifilm FinePix F770EXR Panasonic Lumix DMC-GH2 Panasonic Lumix DMC-GF3 +16 more
Reply   Reply with quote   Complain
alexisgreat
Veteran MemberPosts: 4,316Gear list
Like?
Re: EXR Confusion not fully resolved
In reply to Kim Letkeman, 11 months ago

I use S Priority because I find my hands are steadier than the camera thinks they should be, therefore when I'm using AUTO ISO 3200, I can get away with slower shutter speeds (so the camera selects lower ISO) at longer focal lengths. With P mode and A Priority, the camera was consistently selecting ISO around 400 even in somewhat bright light. I never even thought about this limitation until I realized that there were some shutter speeds I didn't have access to in M mode M size DR 400 that I do have access to in S Priority- and the only difference was that AUTO ISO was being used.

 alexisgreat's gear list:alexisgreat's gear list
Olympus C-7070 Wide Zoom Fujifilm FinePix HS20 EXR Fujifilm FinePix HS50 EXR Olympus E-520 Olympus Zuiko Digital ED 14-42mm 1:3.5-5.6 +1 more
Reply   Reply with quote   Complain
jcmarfilph
Senior MemberPosts: 7,592Gear list
Like?
Re: EXR Confusion not fully resolved
In reply to alexisgreat, 11 months ago

alexisgreat wrote:

I use S Priority because I find my hands are steadier than the camera thinks they should be, therefore when I'm using AUTO ISO 3200, I can get away with slower shutter speeds (so the camera selects lower ISO) at longer focal lengths. With P mode and A Priority, the camera was consistently selecting ISO around 400 even in somewhat bright light. I never even thought about this limitation until I realized that there were some shutter speeds I didn't have access to in M mode M size DR 400 that I do have access to in S Priority- and the only difference was that AUTO ISO was being used.

You may have missed my findings about it...

IMHO, there will be zero or negligible difference in DR and highlight at 1/4000s between MSize DR400 (if it will work properly) and LSize DR100 at ISO-100

-=[ Joms ]=-

 jcmarfilph's gear list:jcmarfilph's gear list
Fujifilm X20 Fujifilm FinePix HS50 EXR +1 more
Reply   Reply with quote   Complain
Timur Born
Senior MemberPosts: 4,230
Like?
Re: EXR Confusion not fully resolved
In reply to alexisgreat, 11 months ago

alexisgreat wrote:

I use S Priority because I find my hands are steadier than the camera thinks they should be, therefore when I'm using AUTO ISO 3200, I can get away with slower shutter speeds (so the camera selects lower ISO) at longer focal lengths. With P mode and A Priority, the camera was consistently selecting ISO around 400 even in somewhat bright light. I never even thought about this limitation until I realized that there were some shutter speeds I didn't have access to in M mode M size DR 400 that I do have access to in S Priority- and the only difference was that AUTO ISO was being used.

Try using AUTO ISO in combination with "+MOTION" IS, if you don't already. It will increase ISO when it detects motion or camera-shake, but will allow lower ISO compared to not using "+MOTION" IS when the frame is steady (aka the camera chooses higher ISO than necessary when "+MOTION" IS is turned off).

-- hide signature --

Red flash eyes save lives and eye-sight!
http://en.wikipedia.org/wiki/Retinoblastoma

Reply   Reply with quote   Complain
photoreddi
Senior MemberPosts: 4,694
Like?
Re: EXR Confusion not fully resolved
In reply to jcmarfilph, 11 months ago

jcmarfilph wrote:

alexisgreat wrote:

I use S Priority because I find my hands are steadier than the camera thinks they should be, therefore when I'm using AUTO ISO 3200, I can get away with slower shutter speeds (so the camera selects lower ISO) at longer focal lengths. With P mode and A Priority, the camera was consistently selecting ISO around 400 even in somewhat bright light. I never even thought about this limitation until I realized that there were some shutter speeds I didn't have access to in M mode M size DR 400 that I do have access to in S Priority- and the only difference was that AUTO ISO was being used.

You may have missed my findings about it...

IMHO, there will be zero or negligible difference in DR and highlight at 1/4000s between MSize DR400 (if it will work properly) and LSize DR100 at ISO-100

No matter what shutter speed, aperture or ISO is used, whether the camera is set to L size or M size, there will be zero differences between one photo's blown highlights and another photo's blown highlights, so you've simply stated a photographic truism for the highlights that should be obvious.

What isn't obvious is how you can think that you know what the dynamic range of those two photos are just by looking at the two photos. The dynamic range capability should be the same if both photos were shot using the same camera using the same ISO, the same resolution and the same resolution. But the resolutions aren't the same. On a pixel basis the larger M size pixels capture more photons so their DR will be greater than the L size pixels (substitute photosites for pixels here if you prefer.) But the captured dynamic ranges appear quite different based on histograms that are too small to be meaningful. Both photos appear to contain significant parts of the image that show totally blown highlights yet these aren't seen in the histograms. How would you explain your conclusion that "there will be zero or negligible difference in DR and highlight" other than by working backwards from preexisting, probably biased assumptions?

Reply   Reply with quote   Complain
Timur Born
Senior MemberPosts: 4,230
Like?
Re: EXR Confusion not fully resolved
In reply to photoreddi, 11 months ago

Sorry to correct you, but L and M size collect the same amount of photons as sensor size is the same. What happens in M size (DR 100) is that noise is calculated out by averaging two sensels into one. It's the same as looking at the "Print" vs. "Screen" resolution graphs at DxOmark. The benefit of the EXR sensor versus downsizing in post is that the averaging on the EXR sensor always happens between sensels of same filter-color and close proximity.

-- hide signature --

Red flash eyes save lives and eye-sight!
http://en.wikipedia.org/wiki/Retinoblastoma

Reply   Reply with quote   Complain
alexisgreat
Veteran MemberPosts: 4,316Gear list
Like?
Re: EXR Confusion not fully resolved
In reply to jcmarfilph, 11 months ago

Thanks Joms, was the left one shot at M size DR 400 AUTO ISO (camera chose ISO 100) and 1/4,000 sec?  It really is hard to tell them apart.

 alexisgreat's gear list:alexisgreat's gear list
Olympus C-7070 Wide Zoom Fujifilm FinePix HS20 EXR Fujifilm FinePix HS50 EXR Olympus E-520 Olympus Zuiko Digital ED 14-42mm 1:3.5-5.6 +1 more
Reply   Reply with quote   Complain
photoreddi
Senior MemberPosts: 4,694
Like?
Re: EXR Confusion not fully resolved
In reply to Timur Born, 11 months ago

Timur Born wrote:

Sorry to correct you, but L and M size collect the same amount of photons as sensor size is the same. What happens in M size (DR 100) is that noise is calculated out by averaging two sensels into one. It's the same as looking at the "Print" vs. "Screen" resolution graphs at DxOmark. The benefit of the EXR sensor versus downsizing in post is that the averaging on the EXR sensor always happens between sensels of same filter-color and close proximity.

No, you relish trying to correct people but in this case it's YOU that are wrong. You're talking about the entire sensor, ignoring or not recognizing that I wrote this :

On a pixel basis the larger M size pixels capture more photons so their DR will be greater than the L size pixels (substitute photosites for pixels here if you prefer.)

Now if you can demonstrate that M size pixels are the same size as L size pixels then maybe you have a point to make. That you appear to be supporting Joms's unsupportable statements is interesting if not amusing. Great minds think alike, or so they say.

Reply   Reply with quote   Complain
Timur Born
Senior MemberPosts: 4,230
Like?
Re: EXR Confusion not fully resolved
In reply to photoreddi, 11 months ago

photoreddi wrote:

No, you relish trying to correct people but in this case it's YOU that are wrong.

I relish my daughter getting healthy, not correcting people on an anonymous forum.

You're talking about the entire sensor, ignoring or not recognizing that I wrote this :

On a pixel basis the larger M size pixels capture more photons so their DR will be greater than the L size pixels (substitute photosites for pixels here if you prefer.)

But you are right, I puzzled up your "pixel basis" statement with a "sensor basis" one. Sorry for that.

Now if you can demonstrate that M size pixels are the same size as L size pixels then maybe you have a point to make.

Here I have to disagree again, though. M size pixels are the same as L size pixels, it's still a 12 mp sensor capturing 12 mp data. It is not a 6 mp sensor where each pixel has a deeper photon well with higher linear dynamic range. Instead the 12 mp pixels are averaged down to 6 mp and then demosaiced, they are not summed up to a higher signal strength. Averaging lowers noise and thus increases signal-to-noise ratio.

You can do the very same using an L size image and downsizing it after demosaicing. The benefit of the EXR sensor based averaging is that only pixels of same filter-color and very close proximity (nearly showing the same information) are averaged. You can get an idea of post-demosaicing results by looking at the measurements of the FinePix-F800EXR at DxOmark, comparing Screen (16 mp pixel) vs. Print (8 mp averaged) results.

http://www.dxomark.com/Cameras/Fujifilm/FinePix-F800EXR

That you appear to be supporting Joms's unsupportable statements is interesting if not amusing. Great minds think alike, or so they say.

I assume this is meant to be an intelligent but snide remark. People on this forum always amaze me with their blatant demonstration of human nature.

-- hide signature --

Red flash eyes save lives and eye-sight!
http://en.wikipedia.org/wiki/Retinoblastoma

Reply   Reply with quote   Complain
photoreddi
Senior MemberPosts: 4,694
Like?
Re: EXR Confusion not fully resolved
In reply to Timur Born, 11 months ago

Timur Born wrote:

photoreddi wrote:

No, you relish trying to correct people but in this case it's YOU that are wrong.

I relish my daughter getting healthy, not correcting people on an anonymous forum.

You're talking about the entire sensor, ignoring or not recognizing that I wrote this :

On a pixel basis the larger M size pixels capture more photons so their DR will be greater than the L size pixels (substitute photosites for pixels here if you prefer.)

But you are right, I puzzled up your "pixel basis" statement with a "sensor basis" one. Sorry for that.

Thank you. It's good to see that unlike some, you're willing to be reasonable.

.

Now if you can demonstrate that M size pixels are the same size as L size pixels then maybe you have a point to make.

Here I have to disagree again, though. M size pixels are the same as L size pixels, it's still a 12 mp sensor capturing 12 mp data. It is not a 6 mp sensor where each pixel has a deeper photon well with higher linear dynamic range. Instead the 12 mp pixels are averaged down to 6 mp and then demosaiced, they are not summed up to a higher signal strength. Averaging lowers noise and thus increases signal-to-noise ratio.

No, M size pixels are not the same size as L size pixels. Pixels are (as you indicate) mathematical creations of a complex demosaicing algorithm. The photosites used to create the M size pixels are the same size as the photosites used to create L size pixels, but there are twice as many L size pixels as M size pixels so they can't be the same size. You seem to think that to create an M size pixel, the demosaicing process has to create individual L size pixels and then average them down to M size pixels. That's one way that Fuji could have done it, but it's more likely that Fuji starts by combining pixel pairs and then uses a more straightforward Bayer demosaic on the averaged pairs. This would produce higher resolution M size images than what you've suggested. Unless you have a link to a Fuji document that explicitly states how the pixels are created, I'd have to think that the Occam's Razor principle suggests that the simpler explanation (or in this case, process), while not foolproof, provides more support for my theory than yours.

.

You can do the very same using an L size image and downsizing it after demosaicing. The benefit of the EXR sensor based averaging is that only pixels of same filter-color and very close proximity (nearly showing the same information) are averaged. You can get an idea of post-demosaicing results by looking at the measurements of the FinePix-F800EXR at DxOmark, comparing Screen (16 mp pixel) vs. Print (8 mp averaged) results.

http://www.dxomark.com/Cameras/Fujifilm/FinePix-F800EXR

Could you post a link to the page that has the "mp pixel" scores along with the controls that allow Screen vs Print to be selected, or describe how to get to it? All I see are DxO's bar charts showing Overall Score, Color Depth Score, Dynamic Range Score and Low Light ISO Score. This is on the Scores page. The Measurements page makes charts available that show ISO Sensitivity, SNR 18%, Dynamic Range, Tonal Range, Color Sensitivity, Full SNR, Color Response and Full CS (Color Sensitivity), but nothing for sensor resolution.

.

That you appear to be supporting Joms's unsupportable statements is interesting if not amusing. Great minds think alike, or so they say.

I assume this is meant to be an intelligent but snide remark. People on this forum always amaze me with their blatant demonstration of human nature.

I've seen that sort of thing from you as well, or at least I think that I have. If not, I probably was conflating you and Trevor. Speaking of whom, I wonder if you agree with his opinion that EXR sensors have no more dynamic range than non-EXR sensors, even when EXR DR mode is considered?

Reply   Reply with quote   Complain
Timur Born
Senior MemberPosts: 4,230
Like?
Re: EXR Confusion not fully resolved
In reply to photoreddi, 11 months ago

photoreddi wrote:

No, M size pixels are not the same size as L size pixels. Pixels are (as you indicate) mathematical creations of a complex demosaicing algorithm. The photosites used to create the M size pixels are the same size as the photosites used to create L size pixels, but there are twice as many L size pixels as M size pixels so they can't be the same size.

I guess this is more a question of how we define things than how we understand them. If you take two consecutive shots on any camera and then combine them, does that increase pixel size or just sample size? There is a difference in my definition.

You seem to think that to create an M size pixel, the demosaicing process has to create individual L size pixels and then average them down to M size pixels.

No, I even described it the way you describe it in the paragraph below. But language barriers surely add to the confusion. What happens is that raw data of all L sized pixels is saved in the RAW file and then adjacent pixels of same filter-color are preferably averaged *before* demosaicing. With certain settings the averaging happens in-camera, with others the RAW converter software has to handle this (some do better than others).

That's one way that Fuji could have done it, but it's more likely that Fuji starts by combining pixel pairs and then uses a more straightforward Bayer demosaic on the averaged pairs. This would produce higher resolution M size images than what you've suggested.

This is how I described it before and even underlined that the supposedly advantage of this EXR approach is that downsizing/averaging can be done before demosaicing on a (mostly) Bayer type pattern. People seem to agree that this also creates better resolving (as in detail, not resolution as in pixels) detail compared to standard Bayer pattern cameras of same MP count (SN mode, not HR mode). My personal jury is still out on that.

Unless you have a link to a Fuji document that explicitly states how the pixels are created, I'd have to think that the Occam's Razor principle suggests that the simpler explanation (or in this case, process), while not foolproof, provides more support for my theory than yours.

The simple explanation is that collecting the same information two times and then averaging the data leads to a higher signal-to-noise ratio, because noise happens randomly. Thus it can be expected that the chance for noise to hit the same (pair of) pixel(s) twice is smaller than just taking a single sample.

But my understanding or interpretation is that averaging two samples is *not* the same as having larger sized pixels (of same tech level) with larger photon wells. Averaging a 1-100 pixel with another 1-100 pixel is not the same as reading a single 1-200 pixel. Which of the two provides better results is something I don't know enough about at this time.

Could you post a link to the page that has the "mp pixel" scores along with the controls that allow Screen vs Print to be selected, or describe how to get to it? All I see are DxO's bar charts showing Overall Score, Color Depth Score, Dynamic Range Score and Low Light ISO Score. This is on the Scores page. The Measurements page makes charts available that show ISO Sensitivity, SNR 18%, Dynamic Range, Tonal Range, Color Sensitivity, Full SNR, Color Response and Full CS (Color Sensitivity), but nothing for sensor resolution.

In the measurement graphs there are two buttons in the upper left corner saying "Screen" and "Print". "Screen" is the native resolution of the measured sensor, "Print" is always sampled down to 8 mp.

I've seen that sort of thing from you as well, or at least I think that I have.

Actually I gave an example in just the very sentence you replied to in this paragraph. Two can play that way, but mine may or may not have been more subtle.

If not, I probably was conflating you and Trevor. Speaking of whom, I wonder if you agree with his opinion that EXR sensors have no more dynamic range than non-EXR sensors, even when EXR DR mode is considered?

Is that what Trevor says that EXR DR does not provide more dynamic range over non EXR sensor of *same* size? Anyway, when two sensors of same size are based on roughly the same technology then they more or less offer the same dynamic range. It's all about collecting photons, turning photons efficiently into electrons and keeping read noise (of the electronic circuits) down.

What EXR DR does in essence is nothing else than taking two consecutive exposures at lower resolution and then combine them into one. You can do the very same with every camera out there and likely get better results at same sensor size and (HR) pixel count than what EXR DR gives. The benefit of EXR is (only) that it works handheld with perfect pixel-alignment and doesn't require post-processing to merge the two exposures. Drawbacks compared to truly consecutive shots are increased noise in highlights and shadows (only half the sensor size used to capture data) and lack of control over the blending process (every RAW converter does it differently and there are blending artifacts and not so well done blending even in-camera). Both methods struggle with moving targets if the longer exposed half uses too slow a shutter speed, but EXR has a slight advantage in that part of both exposures happen at the same time (on the X10 one sensor half starts later and then both stop at the same time).

-- hide signature --

Red flash eyes save lives and eye-sight!
http://en.wikipedia.org/wiki/Retinoblastoma

Reply   Reply with quote   Complain
Timur Born
Senior MemberPosts: 4,230
Like?
Re: EXR Confusion not fully resolved
In reply to photoreddi, 11 months ago

photoreddi wrote:

No, M size pixels are not the same size as L size pixels. Pixels are (as you indicate) mathematical creations of a complex demosaicing algorithm. The photosites used to create the M size pixels are the same size as the photosites used to create L size pixels, but there are twice as many L size pixels as M size pixels so they can't be the same size.

I think what's adding to the confusion is the ambivalent use of the word pixel. Sometimes we mean sensor photosites/sensels by writing pixels, sometimes we mean the demosaiced outcome.

-- hide signature --

Red flash eyes save lives and eye-sight!
http://en.wikipedia.org/wiki/Retinoblastoma

Reply   Reply with quote   Complain
jcmarfilph
Senior MemberPosts: 7,592Gear list
Like?
Re: EXR Confusion not fully resolved
In reply to alexisgreat, 11 months ago

alexisgreat wrote:

Thanks Joms, was the left one shot at M size DR 400 AUTO ISO (camera chose ISO 100) and 1/4,000 sec? It really is hard to tell them apart.

Sorry I forgot to label it but you can view it in original and you will see the resolution of the LSize is larger. The MSize is the left one.

-=[ Joms ]=-

 jcmarfilph's gear list:jcmarfilph's gear list
Fujifilm X20 Fujifilm FinePix HS50 EXR +1 more
Reply   Reply with quote   Complain
photoreddi
Senior MemberPosts: 4,694
Like?
Re: EXR Confusion not fully resolved
In reply to Timur Born, 11 months ago

Timur Born wrote:

photoreddi wrote:

No, M size pixels are not the same size as L size pixels. Pixels are (as you indicate) mathematical creations of a complex demosaicing algorithm. The photosites used to create the M size pixels are the same size as the photosites used to create L size pixels, but there are twice as many L size pixels as M size pixels so they can't be the same size.

I guess this is more a question of how we define things than how we understand them. If you take two consecutive shots on any camera and then combine them, does that increase pixel size or just sample size? There is a difference in my definition.

I take this as a non-answer.

.

You seem to think that to create an M size pixel, the demosaicing process has to create individual L size pixels and then average them down to M size pixels.

No, I even described it the way you describe it in the paragraph below. But language barriers surely add to the confusion. What happens is that raw data of all L sized pixels is saved in the RAW file and then adjacent pixels of same filter-color are preferably averaged *before* demosaicing.

Leaving aside that I haven't been discussing RAW files (files are the things written to memory cards) but JPEGs, this is what I wrote but which is not what you described. Maybe you think that you did, but if so, you need to be careful to avoid writing "pixel" when it seems that you were describing "photosite".

.

.

Unless you have a link to a Fuji document that explicitly states how the pixels are created, I'd have to think that the Occam's Razor principle suggests that the simpler explanation (or in this case, process), while not foolproof, provides more support for my theory than yours.

The simple explanation is that collecting the same information two times and then averaging the data leads to a higher signal-to-noise ratio, because noise happens randomly. Thus it can be expected that the chance for noise to hit the same (pair of) pixel(s) twice is smaller than just taking a single sample.

But you were talking about pixels previously, not photosites, and you continue doing that here. Pixels have already been demosaiced, and that's akin to blurring them with an AA filter. Combining/averaging two blurred pixels results in reduced resolution compared with combining the un-mosaiced dual photosites. Following this is the Bayer-ish demosaicing process that creates the final pixels. This is advantageous for M size resolution which is already much less than the camera's stated resolution, and as reviewers are wont to say, M size resolution is very good indeed.

.

.

Could you post a link to the page that has the "mp pixel" scores along with the controls that allow Screen vs Print to be selected, or describe how to get to it? All I see are DxO's bar charts showing Overall Score, Color Depth Score, Dynamic Range Score and Low Light ISO Score. This is on the Scores page. The Measurements page makes charts available that show ISO Sensitivity, SNR 18%, Dynamic Range, Tonal Range, Color Sensitivity, Full SNR, Color Response and Full CS (Color Sensitivity), but nothing for sensor resolution.

In the measurement graphs there are two buttons in the upper left corner saying "Screen" and "Print". "Screen" is the native resolution of the measured sensor, "Print" is always sampled down to 8 mp.

I've seen that control before but not here. Do you see it here?

.


If not, I probably was conflating you and Trevor. Speaking of whom, I wonder if you agree with his opinion that EXR sensors have no more dynamic range than non-EXR sensors, even when EXR DR mode is considered?

Is that what Trevor says that EXR DR does not provide more dynamic range over non EXR sensor of *same* size?

Yes. He calls EXR DR mode a "party trick".

.

Anyway, when two sensors of same size are based on roughly the same technology then they more or less offer the same dynamic range. It's all about collecting photons, turning photons efficiently into electrons and keeping read noise (of the electronic circuits) down.

No, it's usually about collecting photos, but not when the two halves get different exposures, as happens with EXR DR mode. Resolution is reduced but the dynamic range increases one or two stops worth. I'm surprised that you'd disagree with this as it was stated (with illustrations of the CFA layout) in DPR's review of Fuji's first EXR camera and repeated in subsequent reviews of other EXR cameras. If you disagree with this, you'd have to also say that HDR photography that combines individual photos show using different exposures doesn't produce photos with increased DR.

.

What EXR DR does in essence is nothing else than taking two consecutive exposures at lower resolution and then combine them into one. You can do the very same with every camera out there and likely get better results at same sensor size and (HR) pixel count than what EXR DR gives. The benefit of EXR is (only) that it works handheld with perfect pixel-alignment and doesn't require post-processing to merge the two exposures.

Wow. You're still not acknowledging the use of different exposures by the two sensor halves in EXR DR photos. Do you really not understand this after so much time has passed? I guess that you and Trevor both subscribe to the "party trick" theory after all.

Reply   Reply with quote   Complain
photoreddi
Senior MemberPosts: 4,694
Like?
Re: EXR Confusion not fully resolved
In reply to Timur Born, 11 months ago

Timur Born wrote:

photoreddi wrote:

No, M size pixels are not the same size as L size pixels. Pixels are (as you indicate) mathematical creations of a complex demosaicing algorithm. The photosites used to create the M size pixels are the same size as the photosites used to create L size pixels, but there are twice as many L size pixels as M size pixels so they can't be the same size.

I think what's adding to the confusion is the ambivalent use of the word pixel. Sometimes we mean sensor photosites/sensels by writing pixels, sometimes we mean the demosaiced outcome.

Yes, as I've been saying for ages, and as I've been telling you here in this thread. But that's not the only confusion or disagreement here, as I've just pointed out in my other reply that concludes with a mention of Trevor's "party trick" theory.

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads