Anyway, for all practical purposes, I believe this test proves that in-camera ISO settings does affect the content of the RAW file in a way that cannot be compensated via available software in post.
Correct. The ISO setting IS reflected in the values ultimately written to the RAW file. ISO is not like white balance, it is analogous to shutter speed and aperture setting, and none of those three can be changed after capture. White balance can. ISO, shutter speed and aperture cannot.
Think of it this way, why is there no adjustment for ISO in any available RAW converter?
OK, I stand corrected. Glad to know it. When I design a camera, it's raw format will use a representation which will allow ISO adjustment after exposure.
When I design a camera, it's raw format will use a representation which will allow ISO adjustment after exposure.
I don't think you should.
Think of it as of a car with the manual transmission. For each gear (ISO) you'll have a certain (overlapping) range of velocities (exposures). It's not wise/optimal/possibleto try start the car in 5th gear, it's not feasible to try to drive 90 mph in 1st.
Or like a lens line. There is no such thing as quality zoom lens 10-1,000mm. You have to narrow down your soom range to achieve certain engineering goals and the end image quality. Hence we have ultrawide, wide, medium, tele, and ultra tele zoom ranges - kinda like 100, 200, 400, 800, 1600 ISOs :-)
Anyway, I'm glad we brought this issue up and got some solid practical confirmation on how these things actually work.. :
This case is a little different. I want a high dynamic range camera which can capture faces with sunsets in the background. Using a better representation of the events that actually happened on the sensor would be a step in this direction (as your experiments show.)
This case is a little different. I want a high dynamic range camera which can capture faces with sunsets in the background. Using a better representation of the events that actually happened on the sensor would be a step in this direction (as your experiments show.)
You'd have to design your sensor with like 15 (or even 20) f-stops of the range to cover the usual 5 stops of the today's dslrs and +/-5 (or more) stops for possible ISO setting. Which means 2 to the 15 (or 20th) power...
I sure like you to do so, but I'm kinda suspicious if the technology today is capable of creating a sensor with such a dynamic range
But hey, if you find a VC for that - gimme a call, I'd be glad to work on something like this :
You'd have to design your sensor with like 15 (or even 20) f-stops of the range to cover the usual 5 stops of the today's dslrs and +/-5 (or more) stops for possible ISO setting. Which means 2 to the 15 (or 20th) power...
I sure like you to do so, but I'm kinda suspicious if the technology today is capable of creating a sensor with such a dynamic range
It isn't. Not only that but making an A/D converter accurate to as high as 20 bits, to operate on battery power, and to do so extremely fast (so you can empty the sensor quickly and take another shot), is just not possible. This is why digital backs on medium format cameras are often cooled and are not capable of fast frame rates. This is primarily why Fuji's approach to extending dynamic range involves a second set of photosites that have a neutral density filter fixed above those particular photosites.
OK, I stand corrected. Glad to know it. When I design a camera, it's raw format will use a representation which will allow ISO adjustment after exposure.
As I understand it, RAW isn't really fully RAW from the sensor. The data coming off the sensor is analog and goes through some signal processing hardware before it's actually turned into a digital value that ends up in the RAW file. The ISO setting is applied as part of this signal processing. As you turn up the ISO, the analog value in the senosr is amplified before turning it into the digital value that ends up in the RAW file. The problem with that amplification process is that it also amplifies noise making noise more visible the more you amplify it.
It sounds like what we all really want is for those guys in the lab to come up with a fast, affordable sensor that captures the entire dynamic range of human-perceptible vision, from the darkest tones the human eye can see at night to the brightest tones it can see at high noon in the sun, in a single frame and with no perceptible noise down through the darkest tones, and with enough bit depth so that even the darkest tones are not starved of bits like they are today.
If we ever get that, ISO would be unnecessary. All you would need to do is apply one good curve to the data (including black point and white point, of course), and you are done, and happy.
If we ever get that, ISO would be unnecessary. All you would need to do is apply one good curve to the data (including black point and white point, of course), and you are done, and happy.
Not quite. Then we'd need the output device that could display it.
As I understand it, RAW isn't really fully RAW from the sensor. The data coming off the sensor is analog and goes through some signal processing hardware before it's actually turned into a digital value that ends up in the RAW file. The ISO setting is applied as part of this signal processing. As you turn up the ISO, the analog value in the senosr is amplified before turning it into the digital value that ends up in the RAW file. The problem with that amplification process is that it also amplifies noise making noise more visible the more you amplify it.
Correct. But there is noise in the photosites themselves that generates the analog signal to begin with. And then the problem of creating a linear analog-to-digital converter of high bit depth. What good is an 18 bit RAW file if your converter is only linear to 11 or 12 bits?
RAW files also have some amount of anti-aliasing applied to them. This whole notion that RAW is 100% unprocessed sensor data just is not accurate.
The orignal RAW picture is fact the first one: A
Now, I ask you those who voted 75 % :
How did you guess ?
How did you know it was that one ?
So, there is a difference after all ...
The orignal RAW picture is fact the first one: A
Now, I ask you those who voted 75 % :
How did you guess ?
How did you know it was that one ?
So, there is a difference after all ...
I guessed that the RAW would have been saved as a better quality JPEG than the one straight out of the camera and A looks (to me) to have a slightly less blocky (compressed) sky.
Suddenly I ask myself:
Is is worth, for me, to shoot RAW all the time ?
Antonio,
With the price of Compact Flash continuing to fall, and with large Gb CF available for under $100, why not shoot RAW + Large or Medium jpgs and have the best of both worlds??
You can immediately compare your jpgs to those the camera creates and decide which you prefer. Or you can use the comparison to help fine tune the parameters used in your camera setup to create the direct from camera jpgs.
These are two reasons that I have been shooting RAW + Medium jpgs like rutt suggested. The jpgs are available instantly, and can be helpful in RAW processing also, and gives me insight into exposure accuaracy and WB.
With the price of Compact Flash continuing to fall, and with large Gb CF available for under $100, why not shoot RAW + Large or Medium jpgs and have the best of both worlds??
You can immediately compare your jpgs to those the camera creates and decide which you prefer. Of you can use the comparison to help you fine tune the parameters used in your camera setup to create the direct from camera jpgs.
These are two reasons that I have been shooting RAW + Medium jpgs like rutt suggested. The jpgs are available instantly, and can be helpful in RAW processing also, and gives me insight into exposure accuracy and WB.
I voted for A also.
Pathfinder,
Thank you. I can remember Ruth said he was shooting RAW and JPG at the same time. This procedure is space consuming in the cards.
But today, and because of this very thread I shot pictures in RAW and in JPG and here they are.
I came to the conclusion that sometimes RAW is better and sometimes the other way around.
When in sports I shot JPG. At least that time. I'll try RAW next time to see what happens ...
May be my CF (Sandisk Extreme III) is not fast enough for the action. May be not. May be it behaves correctly because I shoot by small amounts of pictures, say 4 - 5 - 6 each time. If so, I think the CF eats the files just fine. Unless it takes too long to write in the card ...
Well, I'll try.
But let us come back to the experience I did today.
1.st picture, 1.st situation (inside a "house" - a working photo)
Strait from RAW with the automatic adjustments in Photoshop
2.ed picture
JPG direct, no adjustments
3.ed picture
From RAW with treatment
QUESTION
Would I achieve this same result using the JPG file ?
ANSWER
Certainly not. A RAW file is much richer, even more if one works in 16 bits, but the computer complains because it is too busy.
__AGAIN but in another place__
1.st picture, 2.st situation
Strait from RAW with the automatic adjustments
2.ed picture
JPG direct, no adjustments
3.ed picture
From RAW with treatment
In this photo I stacked 2 pictures with different treatments: one for the highlights and the other to the low lights. Highlight in the wall at left.
One last thing: My card is only 1 Gb. I want to get a 4 Gb Sandisk Extreme IV one of these days.
The 20 D says the CF only eats
RAW = 108 files
JPG Hight = 255 files
JPG Hight + RAW = 75 files
JPG Medium + RAW = 86 files
JPG Small + RAW = 94 files
Well I've done my home work today haven't I ?
(While writing this I was listening to Chicago and Blood Sweat and Tears in pandora.com )
All the best Pathfinder. Please comment !! Thank you.:D With CPFilter !
I think your post is a great demonstration of the advantages of shooting in RAW. Your edited RAW images have better highlight detail and better shadow detail than either the straight out of the camera jpgs, or even the straight from RAW conversions. Nice work!
Jpgs ARE faster and smaller and quicker, no doubt. Just not really better almost all of the time. Sometimes, rarely, faster, smaller, quicker, trumps better. Just not usually, for me, at least.
1 Gb cards are rather small for a 20D these days. I rec 4 Gb for substantially more breathing room. I prefer to be able to shoot for an entire day, and never have to open the door to my CF card on my camera until I return home for the evening.
I think your post is a great demonstration of the advantages of shooting in RAW. Your edited RAW images have better highlight detail and better shadow detail than either the straight out of the camera jpgs, or even the straight from RAW conversions. Nice work!
Jpgs ARE faster and smaller and quicker, no doubt. Just not really better almost all of the time. Sometimes, rarely, faster, smaller, quicker, trumps better. Just not usually, for me, at least.
1 Gb cards are rather small for a 20D these days. I rec 4 Gb for substantially more breathing room. I prefer to be able to shoot for an entire day, and never have to open the door to my CF card on my camera until I return home for the evening.
Pathfinder.
Thank you for commenting.
I hope this is usefull for others.:D
I'm hopping to buy a better CF in the Emirates on the way to India.
Next year !
I decided I was not that lazy and found the original photos.
This time the result is here in case you would like to have acloser look.
But here are the photos anyway, smaller size.
A. This photo is the original JPG file without any treatment. B. This photo came from the RAW file, White Balance As Shot. C. This photo came from the RAW file, White Balance as Tungsten. A. B. C.
__
This photo came from the JPG file and Photoshoped.
A soft blue filter was used to correct the original yellow cast.
This photo came from the RAW file White Balance as Tungsten and Photoshoped.
OK, I stand corrected. Glad to know it. When I design a camera, it's raw format will use a representation which will allow ISO adjustment after exposure.
ISO is actually an analog gain applied to the sensor signal before it is converted to digital which is why it is not part of a RAW capture.
The D/A converters in camera are typically only 10-12 bits. If you shoot RAW all 12 bits end up in the RAW file. When you shoot JPEG there are acually two separate compression process that happen. First your camera applies an S curve to the luminace to compress the 12 bit dynamic range down to the 8 required for JPEG. Just like you would do with curves in Photoshop, the S curves are tweaked on each color channel according to the white balance setting. Then it uses the JPEG compression algorithm to reduce the file size before it stores it on your card. The primary motivation for shooting RAW is so you can contol how the dynamic range compression happens. Features like exposure adustments, hightlight/shadow recovery, white balanace and vibrance/saturation in your RAW converter give you more control over the S curves that are used to compress the 12 bit RAW file to 8 bit dynamic range. Noticable JPEG artifacts are rare on modern cameras when set for high quality JPEG.
However there are limits on what you can do even with the RAW file. The typical 12 bit converter does not have enough dynamic range for 4-6 stops of exposure compensation which is possible with yoru ISO setting. You would need a 16 bit D/A converter in the camera to get the extra dynamic range required to cover ISO 100-1600 and that would make the camera dramatically more expensive, add 33% to the size of RAW files and probably require a significant reduction the resolution of the sensor.
It is possible that in the day when 40MBit full frame sensors are a dime a dozen, the camera manufacturers will start competing on dynamic range by putting 16 bit converters in their cameras. We are not there yet in the technology curve.
Comments
Think of it this way, why is there no adjustment for ISO in any available RAW converter?
A former sports shooter
Follow me at: https://www.flickr.com/photos/bjurasz/
My Etsy store: https://www.etsy.com/shop/mercphoto?ref=hdr_shop_menu
I don't think you should.
Think of it as of a car with the manual transmission. For each gear (ISO) you'll have a certain (overlapping) range of velocities (exposures). It's not wise/optimal/possibleto try start the car in 5th gear, it's not feasible to try to drive 90 mph in 1st.
Or like a lens line. There is no such thing as quality zoom lens 10-1,000mm. You have to narrow down your soom range to achieve certain engineering goals and the end image quality. Hence we have ultrawide, wide, medium, tele, and ultra tele zoom ranges - kinda like 100, 200, 400, 800, 1600 ISOs :-)
Anyway, I'm glad we brought this issue up and got some solid practical confirmation on how these things actually work..
This case is a little different. I want a high dynamic range camera which can capture faces with sunsets in the background. Using a better representation of the events that actually happened on the sensor would be a step in this direction (as your experiments show.)
You'd have to design your sensor with like 15 (or even 20) f-stops of the range to cover the usual 5 stops of the today's dslrs and +/-5 (or more) stops for possible ISO setting. Which means 2 to the 15 (or 20th) power...
I sure like you to do so, but I'm kinda suspicious if the technology today is capable of creating a sensor with such a dynamic range
But hey, if you find a VC for that - gimme a call, I'd be glad to work on something like this
A former sports shooter
Follow me at: https://www.flickr.com/photos/bjurasz/
My Etsy store: https://www.etsy.com/shop/mercphoto?ref=hdr_shop_menu
As I understand it, RAW isn't really fully RAW from the sensor. The data coming off the sensor is analog and goes through some signal processing hardware before it's actually turned into a digital value that ends up in the RAW file. The ISO setting is applied as part of this signal processing. As you turn up the ISO, the analog value in the senosr is amplified before turning it into the digital value that ends up in the RAW file. The problem with that amplification process is that it also amplifies noise making noise more visible the more you amplify it.
Homepage • Popular
JFriend's javascript customizations • Secrets for getting fast answers on Dgrin
Always include a link to your site when posting a question
If we ever get that, ISO would be unnecessary. All you would need to do is apply one good curve to the data (including black point and white point, of course), and you are done, and happy.
RAW files also have some amount of anti-aliasing applied to them. This whole notion that RAW is 100% unprocessed sensor data just is not accurate.
A former sports shooter
Follow me at: https://www.flickr.com/photos/bjurasz/
My Etsy store: https://www.etsy.com/shop/mercphoto?ref=hdr_shop_menu
Now, I ask you those who voted 75 % :
How did you guess ?
How did you know it was that one ?
So, there is a difference after all ...
I guessed that the RAW would have been saved as a better quality JPEG than the one straight out of the camera and A looks (to me) to have a slightly less blocky (compressed) sky.
With the price of Compact Flash continuing to fall, and with large Gb CF available for under $100, why not shoot RAW + Large or Medium jpgs and have the best of both worlds??
You can immediately compare your jpgs to those the camera creates and decide which you prefer. Or you can use the comparison to help fine tune the parameters used in your camera setup to create the direct from camera jpgs.
These are two reasons that I have been shooting RAW + Medium jpgs like rutt suggested. The jpgs are available instantly, and can be helpful in RAW processing also, and gives me insight into exposure accuaracy and WB.
I voted for A also.
Moderator of the Technique Forum and Finishing School on Dgrin
Thank you. I can remember Ruth said he was shooting RAW and JPG at the same time. This procedure is space consuming in the cards.
But today, and because of this very thread I shot pictures in RAW and in JPG and here they are.
I came to the conclusion that sometimes RAW is better and sometimes the other way around.
When in sports I shot JPG. At least that time. I'll try RAW next time to see what happens ...
May be my CF (Sandisk Extreme III) is not fast enough for the action. May be not. May be it behaves correctly because I shoot by small amounts of pictures, say 4 - 5 - 6 each time. If so, I think the CF eats the files just fine. Unless it takes too long to write in the card ...
Well, I'll try.
But let us come back to the experience I did today.
1.st picture, 1.st situation (inside a "house" - a working photo)
Strait from RAW with the automatic adjustments in Photoshop
2.ed picture
JPG direct, no adjustments
3.ed picture
From RAW with treatment
QUESTION
Would I achieve this same result using the JPG file ?
ANSWER
Certainly not. A RAW file is much richer, even more if one works in 16 bits, but the computer complains because it is too busy.
__AGAIN but in another place__
1.st picture, 2.st situation
Strait from RAW with the automatic adjustments
2.ed picture
JPG direct, no adjustments
3.ed picture
From RAW with treatment
In this photo I stacked 2 pictures with different treatments: one for the highlights and the other to the low lights. Highlight in the wall at left.
One last thing: My card is only 1 Gb. I want to get a 4 Gb Sandisk Extreme IV one of these days.
The 20 D says the CF only eats
RAW = 108 files
JPG Hight = 255 files
JPG Hight + RAW = 75 files
JPG Medium + RAW = 86 files
JPG Small + RAW = 94 files
Well I've done my home work today haven't I ?
(While writing this I was listening to Chicago and Blood Sweat and Tears in pandora.com
All the best Pathfinder.
I think your post is a great demonstration of the advantages of shooting in RAW. Your edited RAW images have better highlight detail and better shadow detail than either the straight out of the camera jpgs, or even the straight from RAW conversions.
Jpgs ARE faster and smaller and quicker, no doubt. Just not really better almost all of the time. Sometimes, rarely, faster, smaller, quicker, trumps better. Just not usually, for me, at least.
1 Gb cards are rather small for a 20D these days. I rec 4 Gb for substantially more breathing room. I prefer to be able to shoot for an entire day, and never have to open the door to my CF card on my camera until I return home for the evening.
Moderator of the Technique Forum and Finishing School on Dgrin
Pathfinder.
Thank you for commenting.
I hope this is usefull for others.:D
I'm hopping to buy a better CF in the Emirates on the way to India.
Next year !
All the best !
http://www.chrislaudermilkphoto.com/
It was a good idea wasn't it ?
Have a nice week end.
In it there is a small pieces of iron to get the pigeons away.
In fact I do remember to see those pieces of iron when I looked at the statue during the day.
I can´t - I could but I would have much trouble - find the photo to re-treat it, but it is obvious that the RAW file has more detail.
Long live RAW !
This time the result is here in case you would like to have acloser look.
But here are the photos anyway, smaller size.
A. This photo is the original JPG file without any treatment.
B. This photo came from the RAW file, White Balance As Shot.
C. This photo came from the RAW file, White Balance as Tungsten.
A.
__
This photo came from the JPG file and Photoshoped.
A soft blue filter was used to correct the original yellow cast.
ISO is actually an analog gain applied to the sensor signal before it is converted to digital which is why it is not part of a RAW capture.
The D/A converters in camera are typically only 10-12 bits. If you shoot RAW all 12 bits end up in the RAW file. When you shoot JPEG there are acually two separate compression process that happen. First your camera applies an S curve to the luminace to compress the 12 bit dynamic range down to the 8 required for JPEG. Just like you would do with curves in Photoshop, the S curves are tweaked on each color channel according to the white balance setting. Then it uses the JPEG compression algorithm to reduce the file size before it stores it on your card. The primary motivation for shooting RAW is so you can contol how the dynamic range compression happens. Features like exposure adustments, hightlight/shadow recovery, white balanace and vibrance/saturation in your RAW converter give you more control over the S curves that are used to compress the 12 bit RAW file to 8 bit dynamic range. Noticable JPEG artifacts are rare on modern cameras when set for high quality JPEG.
However there are limits on what you can do even with the RAW file. The typical 12 bit converter does not have enough dynamic range for 4-6 stops of exposure compensation which is possible with yoru ISO setting. You would need a 16 bit D/A converter in the camera to get the extra dynamic range required to cover ISO 100-1600 and that would make the camera dramatically more expensive, add 33% to the size of RAW files and probably require a significant reduction the resolution of the sensor.
It is possible that in the day when 40MBit full frame sensors are a dime a dozen, the camera manufacturers will start competing on dynamic range by putting 16 bit converters in their cameras. We are not there yet in the technology curve.