DSLRUsers.com

I went to a great workshop on photoshop Raw processing yesterday run by David Harradine http://www.whack.com.au While I can recommend his workshops to anyone one point stood out something I did not know and should have.

EXPOSE DIGITAL LIKE NEGATIVE FILM

Overexpose rather than underexpose

There is far more information stored in the highlight area of the exposure than in the shadow detail.

This is only for RAW processing but we should all be shooting and processing in RAW

What surprised me is there is a Highlight warning on the nikon digital cameras

A shadow detail warning would be far more useful.

Dug wrote:EXPOSE DIGITAL LIKE NEGATIVE FILM

Overexpose rather than underexpose

I cannot find the bookmark however I came across a tutorial recently regarding reading your histogram. One of the points they mentioned was the statement above and went into detail explaining how they came up with this.

Edit:

Google is your friend.

http://www.luminous-landscape.com/tutor ... ight.shtml

hey dug,

i would have thought the opposite would hold true and that overexposed highlights are the enemy of digital given personal experience and what i've read.

seems like the manufacturers think the same given the highlight warning and also that digital meters tend to bias towards slight under exposure.

actually is that what he means by overexpose ie compensate for the under exposure bias of the metering in dslrs ??

what does he base his info on ? really quite interested to hear any more thoughts on this.


sam

samester wrote:hey dug,

i would have thought the opposite would hold true and that overexposed highlights are the enemy of digital given personal experience and what i've read.

seems like the manufacturers think the same given the highlight warning and also that digital meters tend to bias towards slight under exposure.

actually is that what he means by overexpose ie compensate for the under exposure bias of the metering in dslrs ??

what does he base his info on ? really quite interested to hear any more thoughts on this.
sam

Sam - Yes and no. Looking at the excellent Luminous Landscape article referenced above the quote below puts it in a nutshell:

If you do not use the right-hand fifth of the histogram for recording some of your image you are in fact wasting fully half of the available encoding levels of your camera... (so) bias your exposures so that the histogram is snugged up to the right, but not to the point that the highlights are blown. This can usually be seen by the flashing alert on most camera review screens. Just back off so that the flashing stops.

I was under the impression that the blinking highlights are the blank areas... ie there is no data in them. So therefore you wouldn't be able to get anything back from them. However if the image isn't overexposed and you still expose to the right, you're reducing the shadow area as well as keeping info in the highlights.

stubbsy wrote:Sam - Yes and no. Looking at the excellent Luminous Landscape article referenced above the quote below puts it in a nutshell:

If you do not use the right-hand fifth of the histogram for recording some of your image you are in fact wasting fully half of the available encoding levels of your camera... (so) bias your exposures so that the histogram is snugged up to the right, but not to the point that the highlights are blown. This can usually be seen by the flashing alert on most camera review screens. Just back off so that the flashing stops.

cool - thanks for the link myarhidia and for the article text stubbsy.

makes perfect sense, almost commonsense to be honest in that the ideal exposure is just prior to the onset of blown highlights.

Bruce Fraser, in his excellent book, "Real World Camera Raw with Adobe Photoshop CS2" says this:

"... the linear nature of digital captures makes it preferable to err on the side of slight overexposure rather than underexposure, because underexposing to hold the highlights will make your shadows noisier than they need be."

I was told the same thing at a Phase One: Capture One session and it was intriguing. Takes a little to get your head around but to seriously simplify it;

The top third of the histogram is capable of holding as much data as the lower half.

I know, i know..that's not exactly the correct maths but somebody mixed my grape juice with some old grape juice that had gone off and I can't be bothered working out the number things!

Besides, the whole "blown out" thing is so 2004! It's in vogue to blow your highlights and drench you images with light this year!

Liam =]

Also....

Something I noticed at my brother's wedding was that the Photog (shooting with a D100 and 80-200) was quite happy for the overcast bright white sky to be blinking on his shots as this WAS the correct exposure for the situation. If he exposed for the highlights the girls' salmon coloured dresses would have looked awful and the best man (an Indian fella) would have completely disappeared into his suit!

In the days of film, photogs were more keen to get the focus of the image correctly exposed rather than the white expanse of sky (which would look awful whether you exposed for it correctly or not!).

Liam =]

owen wrote:I was under the impression that the blinking highlights are the blank areas... ie there is no data in them. So therefore you wouldn't be able to get anything back from them. However if the image isn't overexposed and you still expose to the right, you're reducing the shadow area as well as keeping info in the highlights.

I would have thought so too whole correct exposure is still better I had noticed there is still a lot of detail in those blinking highlight

Seeing the recovery of an image from a very overexposed digital image was amazing basically there is 32 times the information in the brightest part of the image compared to the darkest one.

So Dug how does one recover from the 'blown' highlights in CS2. Is curve manipulation the answer. I also read somewhere (perhaps in iNova) to set your D70 to +0.3 in exposure. Confusing but important issue, where are all the techies to nail this one down.

Strangely no curves, using PSCS2 In bridge as long as you keep the image as a raw file and work with the histogram you can adjust the image without clipping any information.

The explanation went above my skill to write about have a look at the tutorials on the site.
very basically you use the exposure, brightness, contrast and saturation sliders to move the histogram to a centered curve and a good on screen exposure. they showed us some amazing uses for curves but only as fine adjustments not as processing the main image.

Keep your histogram in the center and you are not losing any information.

I had always been taught to use curves never touch the Brightness contrast sliders and now I am told don't use curves use the sliders.

I am glad I am not the only person having a problem coming to grips with this concept

I was worried people would go "Well Duh! everyone knows that!"

I wish my explaining skills were better but it is worth reading the notes on the links shown here to get a lot more out of your shooting.

I am not advertising the course but for me, I paid $250 for the day and it started at 9;30 am by 10;30 am I was going "Well, If I leave now I already have got my moneys worth " It was very valuable information and will save me time tears and Money...... Not really most importantly it will make my images look better
but it will save me money too


cheers doug

Idealy I try and get the bulk of the information in the upper part of the histogram, and then dial it down in PS if I need to, but blown areas have no information. For my long exposures, I can't afford to spend the 5-10 minutes getting the shot again if the highlights are blown, so I tend to play it a little safer. Small blown bits aren't usually too bad, and sometimes you can clone them out.

Hi Dug,

I had come across this article a while back. Found it pretty useful. Considering the Histogram is the most accurate light meter you can find, everyone should know what it means and does.

http://www.bayphoto.com/Instructions/Histogram.htm

Nosh.

Dug wrote:... I had always been taught to use curves never touch the Brightness contrast sliders and now I am told don't use curves use the sliders.

I am glad I am not the only person having a problem coming to grips with this concept ... cheers doug

G'day Doug,

I think the "end points" of the curve operate basically like the top and bottom sliders in the levels command, so I think you can do both levels and curves by using a curve. Horses for courses though with them I think...

Haven't played yet, but curves look really cool in LAB when using color curves... Hmmm, I'll have to go play with that

Dargan wrote:So Dug how does one recover from the 'blown' highlights in CS2. Is curve manipulation the answer. I also read somewhere (perhaps in iNova) to set your D70 to +0.3 in exposure. Confusing but important issue, where are all the techies to nail this one down.

G'day,

CS2 (and CS and the new version of NC) can recover some highlights within RAW by dialing down your exposure level (as in -EV), I'd guess at .3EV or a little more maybe? Anymore than that and I reckon you're pushing it though...

Also, changing the WB can alter your exposure by "pushing" some colors up, you can end up clipping some colors where a different WB didn't clip it, just check your histogram in RAW (having the highlights checkbox in Camera Raw checked is handy) when doing the initial convert...

The recent photos that I took from Werribee Treatment Plant were all processed using adobe camera raw. Using a workflow explained by a fellow on NSN. I also decided to try some of my own experimenting and found that using the exposure slider (not too much), contrast, brightness and saturation as mentioned above worked quite well. Then to bring out some of the detail in the black areas of my subject I would create a lasso around my subject, contract and feather it by by 2 an 1.5 respectively. Then I would use the shadow/highlight recovery tool as required until the image looked good. Be careful with the shadow/highlight sliders not to push the image too much as you'll end up with noise in the shadow areas. I think I'll be doing away with Nikon Capture in my workflow, I used to love it and when ever I tried somebody elses workflow I would immediately run back to NC. However now I think I've found a good method on using it.

I've been doing the "expose-to-the-right" thing since I started on RAW a few years ago. There are some very important points to remember:

The histogram displayed by the camera is derived from the processed image, not from the RAW. That is, the choice of WB/etc will affect the displayed histogram (as has been noted, by playing with WB in a RAW converter you'll see it can change dramatically). The camera always produces a processed image (even when in RAW-only modes) even if it's just for display on the LCD.

If your camera only has a "luminance" histogram it may show that nothing is overexposed, even if the red or blue channels are overexposed. The luminance histogram on some cameras is actually the green histogram (lazy programmers) whereas on others I think it's derived as something like R=0.29/G=0.60/B=0.11. I find the R/G/B histograms on my EOS 30D very helpful (in the Nikon range the D200 and D2X also have this - any others?) even keeping in mind that they're still affected by the in-camera WB choice.
Take a photo including the blue sky and push the exposure so that a luminance histogram has a peak towards the right but with no flashing highlights, and then check the image in a RAW converter that has a R/G/B histogram (e.g. Camera Raw) and you'll probably find that the blue channel has well and truly clipped. Similarly it's easy to blow the red channel when photographing sunsets...

Considering the above two issues, it's worth exposing towards the right rather than "right up to the right edge"! As you gain experience you'll get a feel for what is reasonable in different situations.
Overexposing a scene in-camera and then darkening it in the RAW conversion or through adjustment layers in Photoshop can indeed produce cleaner images with more detail, but the whole effect can be ruined if you blow out your highlight detail/colours!


Lastly, a "bell-curve" is not the best histogram for all situations. Each scene has different tonality. For example a high-key image will have lots of bright areas with some interesting detail in the darker portions. At the other xtreme a photo of a cat's eyes reflecting light in the darkness is going to have a large "lump" on the left of the histogram, with the eyes hardly showing up as a blip on the graph.
It is important to consider the histogram in the context of the image. Incidentally, this is related to my whinging in the Nikon forum last week about some cameras not displaying blinking highlights and the histogram on the same screen...

There was an amazing section on LAB colour it went over my head but will have a play with it when I get time.

you will find that the quality of your colour pix's will improve using LAB colour......... you just have greater control .........

If you only shoot in jpg, you will obtain more data capture using RAW. 8bit jpg Vs 12bit RAW.

Thom Hogan best explains this when talking about the Fuji S3 and it's ability to produce 14bit RAW files;

"Take a normal outdoor scene. I often encounter 10 stops of dynamic range (or more) and the usual impact is this: I have to expose for the highlights (i.e. make sure they all get measurable exposure), lest they all blow out to 255,255,255. Since most digital cameras only have about seven stops of dynamic range, at best, that means that the bottom three stops in the scene is all effectively rendered as black, and even the midrange is pushed down towards the shadows. Perhaps a post processing curve can pull out a bit more from the shadows and fix the midrange position, but more often than not that simply increases the noise properties of the image.

With the S3 Pro everything flips (though you still have to watch the histogram to make sure you're not blowing out highlights). Expose to keep the shadows above black. This would normally blow out highlights, but we have that second set of photosites that is more light resistant, which means that they retain highlight detail that is lost by the main photosites. The trick is this: how the heck do you integrate the secondary sensor data with the primary? In camera, the S3 Pro can do exactly that with JPEGs, with settings for Auto, Wide1, and Wide2 in a new function called D-RNG (for dynamic range). Fujifilm's documentation and explanation will confuse you (what does "[set] the dynamic range...to 400%" mean?), so let me try: in theory you can get up to two stops more dynamic range (we're going to have an argument over the exact amount when we get to the performance section, below). No other current DSLR has anything like this ability, so, if it works, the S3 Pro gets bragging rights in at least one imaging area."


This is very true of the S3 and I find it handy when taking photos of white wedding dresses in full sun.

The problem will always be, once you blow out a shot (blinking highlights) there is no data to be retrieved from this area in any format.

samester wrote:i would have thought the opposite would hold true and that overexposed highlights are the enemy of digital given personal experience and what i've read.

especially given the 10 or 11 stops of sensitivity that DSLRs have, compared to 8 or 9 for colour negative film (and the 8 that jpeg lets you use, and the 5 or 6 that most monitors are good for :), you can afford to err somewhat on the side of caution I think. But then, if you use something the the Breezebrowser combined conversion (linear plus exponential) you can get quite a lot of detail out of the highlights... 6 of one sort of thing.

I tend to err on the side of slight underexposure, but that's with the 300D that tends to slightly over-expose...