Just a Simple Question About Dynamic Range

John from PA wrote in post #15049735
Some interesting tips on photographing bold colors, specifically yellow, at http://www.slrphotogra​phyguide.com/tips/vivi​d-colors.shtml

So they seem to be proposing shooting JPEG and underexposing 2/3rd stop. Oh and just wind the saturation up in camera. I followed the link for the "exposure" guide and all of the images used as examples on the page looked rather underexposed to me. But then I like those sort of well saturated images to be nice and bright too!

Yes yellow and orange can be prone to ending up over saturated, especially when you ETTR. I much prefer to pull back the sturation in the yellow/orange channels in the RAW converter and compensate with a touch of -ve lightness. I see this with "Dayglow" pinks and reds as well, but not so much with "normal" darker reds.

Alan

John from PA wrote in post #15049735
Some interesting tips on photographing bold colors, specifically yellow, at http://www.slrphotogra​phyguide.com/tips/vivi​d-colors.shtml

Method 1 is basically what im doing i think.. Except i go with 1 stop, Not 2/3...

I guess the questions i was trying to ask were 1. Is this something that would be improved by upgrading my camera and 2. is there a better way to go about this.. The answer to both questions seems to be "Not really"

Purple and blues are no problem, Yes the petals actually -do- look like this and lack any fine detail in them lol... they're also out of focus since they lack any detail to draw attention to the center of the flower...

IMAGE LINK: http://www.flickr.com …unetsukiphoto/8​028858957/

IMAGE LINK: http://www.flickr.com …unetsukiphoto/8​028859145/

Are you using RGB histogram on your picture review?
That might help figuring out which color is saturated and you can dial it back a bit to get that color's full dynamic range.

KenjiS wrote in post #15051209
Method 1 is basically what im doing i think.. Except i go with 1 stop, Not 2/3...

I guess the questions i was trying to ask were 1. Is this something that would be improved by upgrading my camera and 2. is there a better way to go about this.. The answer to both questions seems to be "Not really"

Purple and blues are no problem, Yes the petals actually -do- look like this and lack any fine detail in them lol... they're also out of focus since they lack any detail to draw attention to the center of the flower...

IMAGE LINK: http://www.flickr.com …unetsukiphoto/8​028858957/
Purple Spiral by Kenjis9965, on Flickr

Hmn, think maybe i shoulda gone more centered with the composition though... given what i was trying to accentuate... ah well

More relevent:

IMAGE LINK: http://www.flickr.com …unetsukiphoto/8​028859145/
A Better Mum by Kenjis9965, on Flickr

The challenge in this was to get a non-distracting background, Throwing the other flowers out of focus doesnt work because the color blur is the same and the flower im trying to highlight simply doesnt stand out enough compared to the background...

Oh also, Look at all that banding in the shadows, Its dreadful /s

Well if you want to get rid of the banding shoot RAW, ETTR pull the general exposure back in post. At the same time you may have to pull the saturation of the oversaturated colour channel (yellows and oranges in these cases) back a bit. As well as adjusting the saturation value for the channel you may have to reduce the Lightness channel as well. I'm not sure about other processors but Lightroom / ACR allow you to adjust Hue Saturation and Lightness for Orange, Yellow, Cyan, Purple and Magenta as well as Red, Green and Blue colour channels.

If you choose the nutral picture style and reduce the saturation and sharpness controls to minimum in camera the RGB histograms display on the camera should show you reasonably closely if you have blown any of the in camera RGB channels. Even with these settings there is normally a little headroom left in the RAW data before any of the RGB channels actually clip.

Alan

Psst, i was being sarcastic, there is no banding

KenjiS wrote in post #15051390
Psst, i was being sarcastic, there is no banding

Reading this on my phone so it's a bit hard to see the images. It scales everything wierdly. I never thought a 280 plus ppi screen could be such a pain (Galaxy Note Mk1). Plus it's getting late over here now.

Alan

BigAl007 wrote in post #15051615
Reading this on my phone so it's a bit hard to see the images. It scales everything wierdly. I never thought a 280 plus ppi screen could be such a pain (Galaxy Note Mk1). Plus it's getting late over here now.

Alan

Lol its alright mate We have our moments... I dont push things hard enough to get banding..

Lowner wrote in post #15049579
You are talking about the actual DR of a scene. Thats got nothing at all to do with the the DR that the camera captures. And you and I both know how to manipulate captured tones in camera so don't get cute.

You have shown me an image that looks VERY low contrast. Thats NOT high DR to me.

Your last comment is unworthy of you. But I will admit that I am absolutely certain that my 5DII captures no more than 5 stops. A proper 11 or 12 stop range would be massive, an amazing image to see and if Canon are claiming that they have sensors capable of that, then you need to show me one. Because I've not seen one yet - EVER.

This is just bizarre. Do this test. Meter on a grey card. Take a shot. You have a grey photo. Now underexpose by 2.5 stops. You should get pitch black - 0,0,0. Now overexpose by 2.5 stops. You should get pure white - 255,255,255. In RAW there should be no information on either the overexposed or underexposed images. If you have any data at all at + or - 2.5 stops, then you are wrong. Your camera is capturing information outside a 5 stop range.

I suspect what you are getting at is not 5 stops of DR, but +/- 5 stops which is 10 total stops. From that perspective you might have some bit of an argument.

KenjiS wrote in post #15043427
In regards to DR/Color/Curves/Etc...​.

First yes i have a calibrated screen lol

Basically I just have noticed sometimes with some images of flowers that I end up in a situation where exposing it "normally" ends up with details in the petals/colored bits are completely blown out, Despite the histogram and the image itself being "correct" in regards to exposure

Usually i can even everything out with a whole ton of levels adjustment and such and a lot of tweaking, but sometimes the image is just "beyond" help and i cant bring it back or i just have to do a lot of work playing to get it to look nice

But generally i have more luck slightly underexposing the image and pushing things a bit as opposed to the other way around, Which is against HAMSTTR but it works in these situations...

If it helps it happens a lot with Red and yellow flowers if that helps and its just me thinking and trying to figure out a way to improve things..


1. Would having more dynamic range help reduce the amount of adjusting i have to do and reduce the number of times i get something completely unfixable (No im not suggesting i switch to Nikon, I could simply go up to a FF body instead which would buy me quite a bit of DR)

2. I have pondered the idea of using a circular polarizer in these situations and seeing what happens as it might help... Any other suggestions would be good

Just some thoughts recently, Not complaining, Love my 7D just think i might have found a limitation there...

Here, example btw

RAW:

PPd:

IMAGE LINK: http://www.flickr.com …unetsukiphoto/8​025231982/
Warm Sunflower in Fall by Kenjis9965, on Flickr

When i exposed the image closer to the final product it lacked a lot of the detail (The petals were just bright yellow and just didnt look "right", Underexposing it and pushing it up resulted in that which i think looks great personally)

1. the original exposure probably blew out one of the channels, with something that deeply saturated it can happen easily
2. some of those colors won't show properly unless you use a wide gamut monitor and view the image in adobergb or even prophotorgb and srgb will clip them and make them all look the same

wombatHorror wrote in post #15052551
1. the original exposure probably blew out one of the channels, with something that deeply saturated it can happen easily
2. some of those colors won't show properly unless you use a wide gamut monitor and view the image in adobergb or even prophotorgb and srgb will clip them and make them all look the same

Yeah I actually got to thinking on number 2, My monitor is certified 100% sRGB, but its only 96% AdobeRGB... Could be hitting that 4%!

I mean my monitor is good, but we're not talking a $2500 Eizo or something..lol..

kfreels wrote in post #15052509
This is just bizarre. Do this test. Meter on a grey card. Take a shot. You have a grey photo. Now underexpose by 2.5 stops. You should get pitch black - 0,0,0. Now overexpose by 2.5 stops. You should get pure white - 255,255,255. In RAW there should be no information on either the overexposed or underexposed images. If you have any data at all at + or - 2.5 stops, then you are wrong. Your camera is capturing information outside a 5 stop range.

I suspect what you are getting at is not 5 stops of DR, but +/- 5 stops which is 10 total stops. From that perspective you might have some bit of an argument.

Thanks for the reply. But thats basically what the pro landscape 'tog and I were playing with when we agreed we were seeing 5 stops. This was on a very bright day with hard shadows and white cloud. What software can achieve subsequently is an entirely different story, my complaint is with the manufacturers (not just Canon) using artificial lab tests to claim completely artificial DR figures. But you are falling into the same trap that Elie fell into. Just because Photoshop does not read 0,0,0 or 255,255,255 does not mean that whats left is usable in the field, and thats all I'm interested in. So it reads 8,10,6 or 250,252,249 simply because we guessed at a hilight or shadow does not disprove what we found.

I did see some superb results shown here from the new Nikon D800E. Still not 11/12 stops, and after processing I seem to remember so a bit suspect. But at first glance it was the best dynamic range I've seen to date.

Richard,
In the photo from Blenheim Palace, http://rcb4344.zenfoli​o.com …40446/h440f7d78​#h440f7d78, do you really think that the range from the shaded side of the tree to the clouds is only 5 stops?

Elie,

Yes, I most certainly do.

Edited to add: Don't forget that I've been playing in Photoshop, this is not out of camera. The shot is actually 2 exposures carefully blended and even then a real 11/12 stopper would have far more detail in the tree trunk and I could have let the sky stay lighter and still retain detail without it blowing out.

Not relevent to this discussion but I thought I'd mention it anyway. Its not my favourite image, its too "pretty pretty", I prefer a less "structured", more "came across by accident" style (even if in reality I have had to work long and hard to achieve it).

The tip in the link I provided reads

Picking this apart, especially if you are familiar with the zone system, you will find that healthy green grass is a near perfect equivalent to an 18% gray card. The slight underexposure then assures the details don't suffer by being blown out (once gone, not recoverable). The OP if he is really into macro flower photography should also consider an ExpoDisk (instructions attached). It works well although I find it best on very small objects.

John from PA wrote in post #15053593
The tip in the link I provided reads

Picking this apart, especially if you are familiar with the zone system, you will find that healthy green grass is a near perfect equivalent to an 18% gray card. The slight underexposure then assures the details don't suffer by being blown out (once gone, not recoverable). The OP if he is really into macro flower photography should also consider an ExpoDisk (instructions attached). It works well although I find it best on very small objects.

John,

That advice is unchanged from the days of Slide film and probably for exactly the same reasons. Over exposure was the big no-no if we wanted saturated "wow factor" projected images.