Ugly Hedgehog - Photography Forum
Home Active Topics Newest Pictures Search Login Register
Main Photography Discussion
ISO-less exposure
Page <prev 2 of 15 next> last>>
May 9, 2018 15:00:13   #
BebuLamar
 
MrBob wrote:
This is prob. one of the more interesting posts I have read in a long time... brings up all sorts of things to ponder on. I would like to see more comments by other experienced folks with a lot more expertise than myself. A VERY interesting topic Cat... Thanks for posting.


That is the effect of the 14 bit RAW files. If the RAW files are only 8 bit you will get the posterization effect when you increase brightness in post.

Reply
May 9, 2018 15:02:40   #
R.G. Loc: Scotland
 
CatMarley wrote:
....The ability to move the exposure by 3 or 4 EV without loss of quality just by shooting raw could come in handy.


It's an interesting point that goes against what is commonly perceived as being the right way to do things. However, I suspect that it only works with fairly modern cameras. I can remember when the Nikon D750 came out, a reviewer posted a totally black image from a D750 that he successfully lightened in PP to a normal level of brightness, and the noise level was impressively low.

I suspect the reason why what you say is against perceived wisdom is because until recently, lifting shadows meant bringing out noise - without exception - and that applied to raw files as well as jpegs. I certainly have no shortage of memories of finding out exactly that. It'd be interesting to find out if your findings can be expected from all makes and models. I suspect that it's only certain brands and the most recent models that it applies to.

Reply
May 9, 2018 15:06:10   #
CatMarley Loc: North Carolina
 
E.L.. Shapiro wrote:
Extremely interesting!

For folks like me, who spent a lifetime dealing with film and A.S.A., ISO, din, Weston and personalized exposure indexes, it my be difficult to relate all of that directly to digital/electronic sensitivity ratings. Of course, physically speaking, there is quite a difference between silver halide/chemical technology and the physics of the movement of electrons in a circuit. Film, with its emulsion issues of speed, grain, acutance and the enormous variations of all of theses factors as they are effected by as many variations in processing is a science and sometimes a form of alchemy or witch's brews in itself. Digital photography is more stable, repeatable and predictable- at least to me.

In a practical usage sense, however, it seems to me that are cretin correlations or parallels of the two technologies that hold true regardless of their theoretical differences such as format/sensor size and an increase of nose/grain as the exposure index is increased and how theses phenomena relate to the quality of the final image, especially in critical scenarios or requirements such as where high degrees of enlargement are indicated.

In my own professional/commercial my practice is simply to shoot everything in RAW, oftentimes in manual mode and selection ISO speeds accordingly. Mostly I opt for the lowest speed that is practical but have not experienced any serious quality losses at moderate to higher speeds. I seldom have the need to go as high as 1600 let alone higher. Am I missing something?

Oh!- this is in no way a facetious question. Down here in "commercial land" we get so involved in workaday stuff and production routines that we can miss out on some of the newfangled good stuff and latest discoveries or theories. Please elaborate!

My question is, regardless of where the "amplification" takes place in the digital chain of events, is the any difference in photographic results and does this discovery debunk the common knowledge or conventional wisdom of RAW vs. Jpeg usage, noise issues or anythg in post processing?
Extremely interesting! br br For folks like me, w... (show quote)


Yes I think it does have practical application. Namely that you do not have to worry about underexposure if you shoot raw. This applies to people who shoot mainly jpegs, and often are faced with badly underexposed files because they forgot to change their ISO. If they had shot raw + fine, they could dig a perfect exposure out of the gloom.

Reply
 
 
May 9, 2018 15:08:44   #
E.L.. Shapiro Loc: Ottawa, Ontario Canada
 
Oh- yes! With NEGATIVE films it was common practice to expose for the shadows and the print down for the highlights. With transparency and digital work the practice was/is to expose for the middle tomes and make cretin that the highlights are in range and fill the shadows adequately in that highlights are easily blown out and significant loss of shadow detail in underexposure is usually irretrievable.

Does this relate to this theory in any way?

Reply
May 9, 2018 15:13:19   #
CaptainC Loc: Colorado, south of Denver
 
What camera?

Reply
May 9, 2018 15:16:54   #
LWW Loc: Banana Republic of America
 
CatMarley wrote:
I was unaware until recently that the camera exposes everything based on aperture and shutter speed alone. The ISO you dial in is a simple after the fact amplification of the signal, and it happens during the processing, not the exposure itself. So taking a shot with ISO 1600 or 200 makes no difference if you are shooting raw. You can apply the 3EV boost using your ISO dial or pushing the exposure 3EV during conversion of the raw file. Here are 2 jpegs, one was taken at ISO 200 and one at 1600. The 200 jpg straight from the camera was black, and the 1600 jpg was properly exposed. Aperture and shutter were the same both shots. the raw 200 was pushed 3 EV during conversion from raw to jpg. Bet you can't tell the difference!
I was unaware until recently that the camera expos... (show quote)

Bingo.

Reply
May 9, 2018 15:17:52   #
CatMarley Loc: North Carolina
 
R.G. wrote:
It's an interesting point that goes against what is commonly perceived as being the right way to do things. However, I suspect that it only works with fairly modern cameras. I can remember when the Nikon D750 came out, a reviewer posted a totally black image from a D750 that he successfully lightened in PP to a normal level of brightness, and the noise level was impressively low.

I suspect the reason why what you say is against perceived wisdom is because until recently, lifting shadows meant bringing out noise - without exception - and that applied to raw files as well as jpegs. I certainly have no shortage of memories of finding out exactly that. It'd be interesting to find out if your findings can be expected from all makes and models. I suspect that it's only certain brands and the most recent models that it applies to.
It's an interesting point that goes against what i... (show quote)


I am not a camera engineer, and I am sure that some folks on this forum can supply the answers about the circuitry. I don't think the processing is all that different though from make to make. I am using a Fuji XT-2. And it seems that this camera, at least, exposes every shot dependent, just as in film days, only on the aperture and shutter speed. The native zero amplification in this camera is ISO 200. I am uncertain if amplification is written into the raw file data or if it is applied only when the raw data is translated into the readable file (tiff or jpeg). It certainly does not appear to make any difference, as I have demonstrated.

Reply
 
 
May 9, 2018 15:28:10   #
AlohaBob Loc: Los Angeles, CA
 
CatMarley wrote:
I finally read the book!


Thanks for responding. What book are you referring to?

Reply
May 9, 2018 15:40:33   #
CatMarley Loc: North Carolina
 
AlohaBob wrote:
Thanks for responding. What book are you referring to?


"The Fujifilm X-T2 120 X-Pert Tips" by Rico Pfirstinger

Reply
May 9, 2018 15:51:52   #
Rich1939 Loc: Pike County Penna.
 
I am not qualified to give an in depth dissertation on this but it would come under ISO Invariance. Basically what you have found will apply up to a certain ISO level, that will vary from camera to camera. After that level noise will begin to show up. Try the same test again but use a higher ISO like 3200 or 6400. At some point noise will be an issue.

Reply
May 9, 2018 15:55:28   #
CaptainC Loc: Colorado, south of Denver
 
This observation is true of many (not all) of the newer cameras. Try that with a camera built in 2006 an you would destroy the image with noise. As a Nikon user I can make my observations from that standpoint ands I can tell you the D810 and D850 are "ISO Invariant." You can do just as you indicated, underexpose buy 2-3, maybe 4 stops and still get a VERY usable image. There is some degradation, but it is really minor and in most cases not apparent. Try that on my D2X and all you get is ugly noise.

This does not work as well on the other end. OVER expose by four stops and then what is gone is pretty much...well...gone. One can still recover highlights quite well for a one-stop over exposure or even a bit more, but there is not the latitude one has with an underexposure.

This is not a reason to not get a good exposure correct with the initial exposure.

Reply
 
 
May 9, 2018 15:55:53   #
selmslie Loc: Fernandina Beach, FL, USA
 
CatMarley wrote:
... So taking a shot with ISO 1600 or 200 makes no difference if you are shooting raw. You can apply the 3EV boost using your ISO dial or pushing the exposure 3EV during conversion of the raw file. ...

You are right but this is only true if you are recording a scene with a narrow dynamic range like the one you posted.

For example, if a scene covers only 5 stops, the raw file might contain a range of values (in a 14-bit raw file) between:

1. 512 through 16383 at ISO 1600
2. 256 through 8191 at ISO 800
3. 128 through 4095 at ISO 400
4. 64 through 2048 at ISO 200
5. 32 through 1023 at ISO 100

All of these can be made to look the same after you adjust only the Exposure slider during post processing. All of the noise levels and colors will be the same because the physical exposure is the same. As CaptainC points out, you need to be using a modern ISO invariant sensor.

The only difference will be in the appearance of the JPEG that comes straight from the camera. For the range of raw values I listed above, it is likely that only #2 will look right. #1 will be overexposed and #3, 4 and 5 will look underexposed by 1, 2 and 3 stops.

For a scene with a wider DR you can't really get away with this much variation. However, since colors are only really strong around middle gray +/- 1 stop and clear details +/- 2 stops, the important parts of any scene will still cover less than 5 stops.

Reply
May 9, 2018 16:17:15   #
srt101fan
 
CatMarley wrote:
I was unaware until recently that the camera exposes everything based on aperture and shutter speed alone. The ISO you dial in is a simple after the fact amplification of the signal, and it happens during the processing, not the exposure itself. So taking a shot with ISO 1600 or 200 makes no difference if you are shooting raw. You can apply the 3EV boost using your ISO dial or pushing the exposure 3EV during conversion of the raw file. Here are 2 jpegs, one was taken at ISO 200 and one at 1600. The 200 jpg straight from the camera was black, and the 1600 jpg was properly exposed. Aperture and shutter were the same both shots. the raw 200 was pushed 3 EV during conversion from raw to jpg. Bet you can't tell the difference!
I was unaware until recently that the camera expos... (show quote)


You say:
"the camera exposes everything based on aperture and shutter speed alone. The ISO you dial in is a simple after the fact amplification of the signal"

"The camera doesn't care it shoots everything at its native ISO and only applies the amplification after the fact."


So:
I set my camera to aperture priority, ISO 200, and f/5 and aim it at a target: the camera sets shutter at 1/50.

Keeping everything else the same I now set the ISO to 1000. Based on your comments I would expect the shutter to stay at 1/50.

It doesn't, it goes to 1/250. And it shows that value in the viewfinder before I take the shot.

This doesn't seem to track with your assertions. What am I missing?

Edit: Camera is Nikon 5300 set to record Raw.

Reply
May 9, 2018 16:36:54   #
Rongnongno Loc: FL
 
Oh, another finally found the quality offered by an invariant sensor. This information is valid only for the latest top of the line cameras.

Been posting about this for a long time.

There is a caveat thought... Do not trust this so much that you start not caring about ISO. Also use the ISO setting that offers the best DR (hint: This is not the base ISO).

Reply
May 9, 2018 16:39:03   #
selmslie Loc: Fernandina Beach, FL, USA
 
srt101fan wrote:
You say:
"the camera exposes everything based on aperture and shutter speed alone. The ISO you dial in is a simple after the fact amplification of the signal"

"The camera doesn't care it shoots everything at its native ISO and only applies the amplification after the fact."


So:
I set my camera to aperture priority, ISO 200, and f/5 and aim it at a target: the camera sets shutter at 1/50.

Keeping everything else the same I now set the ISO to 1000. Based on your comments I would expect the shutter to stay at 1/50.

It doesn't, it goes to 1/250. And it shows that value in the viewfinder before I take the shot.

This doesn't seem to track with your assertions. What am I missing?

Edit: Camera is Nikon 5300 set to record Raw.
You say: br "the camera exposes everything ba... (show quote)

The sensor has only one level of sensitivity. It does not change when you change the ISO.

Changing the ISO setting only causes the camera to suggest or calculate a different exposure in order to find the right the gain or amplification of the signal received by the sensor to produce a "normal" range of values in the raw and JPEG file.

Reply
Page <prev 2 of 15 next> last>>
If you want to reply, then register here. Registration is free and your account is created instantly, so you can post right away.
Main Photography Discussion
UglyHedgehog.com - Forum
Copyright 2011-2024 Ugly Hedgehog, Inc.