Is RAW format the answer?

[PaulB98]

Hello,

I recently took some photos of a concert by the local choral society. The pictures of the rehearsal in the afternoon were no problem as there was plenty of soft natural light from the skylights in the hall. The pictures of the concert in the evening were more of a problem, with spotlights creating areas that were 'blown out' in the results while other areas were too dark. Also there was a mixture of lighting with the spotlights, some fluorescent lights by the exits and some remaining natural evening light through the roof, so I took pictures with different white balances but no one shot looked ok everywhere. I did the best I could editing the jpegs afterwards, but I have since read up about RAW format which I have never used. Would it have been useful in this situation to shoot in RAW? What would be the steps? It looks like I could 'develop' a RAW shot several times with different exposure compensation and white balances and produce a set of jpegs (or presumably TIFF would be better?) - and then import them all as layers in GIMP and combine the best bits of each. Would that be feasible?

Your advice would be welcome!

Paul

[RustBeltRaw]

Welcome to Cambridge in Colour, Paul. Donald or another mod should be along in a minute to ask you to add your approximate location to your user profile. Helps us make local connections.

In terms of image quality, shooting RAW is generally better than shooting JPEG. Essentially, a RAW file contains finer color detail (12 or 14 bits per R, G, B channel at each pixel instead of JPEG's 8). While the image's dynamic range (ratio between lightest and darkest part of the shot) is basically determined by the camera, RAW smooths the transitions between differently-colored areas.

The other advantages are mainly related to the algorithms applied by post-processing software. When you apply a white balance change to a JPEG, you are applying a change to the output of the camera's post-processing software. When you apply a white balance change to a RAW, you are applying the change to the actual sensor data. Since the RAW contains more data than the JPEG, there's more for the algorithms to work on, which frequently manifests in things like improved shadow detail, the appearance of greater dynamic range thanks to additional details in bright and dark areas, more predictable responses to edits, and greater freedom to make edits without losing details.

Essentially, a RAW file gives the computer a lot more data to work with. The bugger is that most monitors only have 8 bits per color channel, so an unprocessed RAW will usually look identical to a JPEG made from it. This is part of the reason prints frequently look spiffier than the same image viewed on a monitor. Some special monitors can display 12 or 14 bits per pixel.

With respect to post-processing, there's nothing particularly special about RAW's non-destructive editing. It's nice that edits are non-destructive by default, but you can easily set up a work flow which does the same thing for a JPEG (just work from a copy of the original).

Regarding your on-site problems, yes, RAW probably would have helped a bit. Transitions into and out of dark and light areas would have been finer, and slightly less blown out. But mixed light is always a bugger. Best you can usually do (if possible) is always keep your back to one set of lights, and set your white balance to make people's faces look correct.

It's certainly possible to composite ("use bits of each") a better shot. However, with the right software (Adobe Camera Raw, Photoshop, and possibly Lightroom), you can apply targeted adjustments (including white balance, among many others) to selected areas. That's probably an easier approach. Seeing a sample of your work will help us provide more concrete advice.

[Manfred M]

Lex has provied a very good overview of RAW versus jpeg; so I won't add any more to his very good answer. I would of course be great to see a sample image to confirm our suspicions here, otherwise we are more or less guessing at what you are trying to deal with.

The issue is still going to be your camera's sensor performance; if the hightlights are too bright (the bright lights for instance) can still be blown out and having a RAW file is not going to get those back. The same comment goes for shadow detail; it sounds like you are shooting a high contrast scene, and if your sensor registers shadow detail as pure black, RAW is not going to help there either. Dynamic range is related to the ISO setting you are shooting at; the lower the ISO setting, the higher the dynamic range your camera can capture.

The solution to mixed lighting is to try to avoid shooting in it. You might be able to do something in post, but I wouldn't count on it, especially in the complex lighting situation that you describe. Not only do you have multiple colour temperatures of the various light sources, but they run into each other coming up with unique lighting gradients. You might consider a B&W conversion, as this largely negates the impact of the different light sources.

Can you take different images and mask / combine different layers; yes of course, but understand that you will have to get fairly good registration between the layers for this to work; multiple images taken using a tripod where neither the camera nor the subject move would be ideal. Again, just because it can be done, doesn't mean you will get results that are better than you already have.

Bottom line is, you were shooting under challanging conditions, so neither RAW nor post-processing are likely to be a magic bullet that can fix all (or any) of the problems you have noted.

[PaulB98]

Thanks to both of you for the detailed replies! I think I have understood how developing RAW files works with regards to colour temperature, but I still have a question about exposure compensation if it wouldn't try your patience.

As I understand it, the RAW file contains the data from the light that fell on the sensor, plus a set of instructions on what to do with it in 'Development'. So if I have set the camera to 'Daylight' and leave that setting in the computer software it makes the colour balance of the exported jpeg look correct assuming I had indeed taken a picture outside on a sunny day. If it had actually been a cloudy day but I had left the setting on Daylight, I could correct my mistake by changing the setting for colour balance to 'Cloudy' before exporting. My understanding is that this correction would produce an identical result to having got the setting right in the first place.

The setting on the camera for Exposure Compensation is surely different. If I set +2EV on the camera it changes the shutter speed or aperture so that more light falls on the sensor and therefore records more detail in the shadows. So this is a physical change rather than a logical change. Failing to set +2EV on the camera and then setting +2EV in the RAW file development process is not going to produce an identical result is it? I think what you are saying is that it would produce a better result than brightening a jpeg. Presumably the converse is true and setting -2EV in the RAW file development will restore a blown highlight better than darkening a jpeg?

Here is an example of the problem. The spotlights were up in the roof and there was no way of moving them. They hadn't been set up specially for this event so the choir was stuck with this random pattern of illumination, where a couple of people on the front row are highlighted, along with someone on the back row whose hair and beard are very white anyway, while the soloist is in the shade at the front! The fluorescent lights are off to the left (out of the picture because I cropped it) but are giving the wall on the left what looks like a greenish tinge to me. I normally use my camera just to take nice landscapes when I'm out for a walk so I really didn't know how best to cope with this setup. Feel free to suggest anything I could have done to get a better result!

[HenkB]

Paul,

You have had two very good answers already and there's only a little bit left to add. I've been shooting raw only for quite a while and have been in circumstances much like those you experienced. My experience has been is that extreme differences in color cast (from gelled spots, perhaps) can be managed with modest desaturation. It will still be evident that the image was taken in cross lighting but this may be taken as part of the 'charm' of concert lighting.

The real killer is highlight clipping which is worsened by the gelled spots' ability to totally blow out single channels. In some cases, the only answer is to trend moody and take the image to black and white. In others, your proposal to process a single raw image for each might hold some promise but may never produce totally normal looking color renditions without more work than the image might be worth. And, yes, TIFF is the better choice for layered post-processing as long as your computer can deal with the memory issues.

I find the best strategy under these conditions is to meter for skin tones only (I use spot), shoot raw, chimp the rgb histograms (because the meter isn't necessarily prepared to compensate for gelled spots and the luminance histogram is likely to miss single-channel mega-clipping), shoot in manual (so I can nail down an exposure that doesn't clip and keep it) and post-process until the result works for me.

[NorthernFocus]

The simple and direct answer to your original question is, no, RAW is not the answer. Some excellent information has been provided by the prior posters. And as pointed out, RAW definately gives you more to work with in post processing than JPEG. But as touched on by the last post, the real answer to this particular problem is for the photographer to 1)understand the limits of the camera 2)understand what can/can't be done in post processing 3) perhaps most important, with the first two in mind, develop realistic expectation of what can be achieved under the shooting conditions and properly expose the important parts of the scene, recognizing that some highlights/shadows won't be recoverable. In other words, decide what you're going to capture and conciously decide what you aren't going to worry about.

Once you do understand the equipment limits, you can take advantage of the known properties of your camera system to achieve a desired effect. It may not look like what the eye would see at the moment in time, but it may be as or more interesting.

[davidedric]

The responses you have give excellent advice about how to manage the conditions, which could be summarised as "it's (very) hard", but not a lot has been said about how to improve the image you have in pp, which I think is quite a lot. You mention Gimp, which I don't know or use though many here do, but as a (fairly novice) Lightroom user it would be quite straightforward to tone down the guy with the white hair, select and bring out the exposure and the crispness of the soloist, do something about evening out the light on the choir, and so on. It wouldn't solve the blown highlights, nothing can, but I'm pretty sure you could get a much more pleasing image.

[dje]

The setting on the camera for Exposure Compensation is surely different. If I set +2EV on the camera it changes the shutter speed or aperture so that more light falls on the sensor and therefore records more detail in the shadows. So this is a physical change rather than a logical change. Failing to set +2EV on the camera and then setting +2EV in the RAW file development process is not going to produce an identical result is it? I think what you are saying is that it would produce a better result than brightening a jpeg. Presumably the converse is true and setting -2EV in the RAW file development will restore a blown highlight better than darkening a jpeg?

Hi Paul

Just a bit of clarification if I may.

The main issue here is electrical noise generated by the sensor which shows up as "graininess" in the darker areas of the image. There is a component of this which is independent of the amount of light the sensor is exposed to. As a result, if you under-expose, then the level of the actual image data is lower in relation to the noise and the noise will appear more noticeable. Correcting the exposure in software will not improve the situation whether you shoot raw or jpeg. However when the camera processes the jpeg image, it usually applies some noise reduction. With raw, noise reduction is applied in pp. You can also apply extra noise reduction to a jpeg in pp but applying it once only is probably a better process.

For blown highlights, there is usually a bit of safety margin built in to selection of the white point by the camera processing software. With raw processing, you can often recover a bit of that margin so yes raw does usually give some chance of recovering blown highlights but only to a certain extent.

Dave

[jcuknz]

As for the photo you supplied I think you could have spent a little more time in editing to correct the minor problems. With PSP I used the 'lighten/darken' tool to adjust .. I think I took the gentleman just behind the singer back too much but that could be considered appropriate as he was distracting from the singer. For the dark lady in the back row I wasn't sure what to do.
I also think the singer was a little on the dark side. I cannot see anything wrong with the colour for a shot in these conditions, I think one can be overly precise in some matters at times. I do not see any great need to have shot in RAW ...it is probably worthwhile if one has doubts to shoot raw+jpg to give oneself the option.

[Glenn NK]

OK, I'm going to add some more confusion to the mix.

Last week while shooting in strong mid-afternoon daylight, at times one of the faces of my grand-daughter or her friend were in shadow, while other was in direct sun. Even with strong shadow reduction in LR, I couldn't fully correct it.

I found a solution (albeit time-consuming): In Lightroom using the adjustment brush, I carefully brushed the areas that were in shadow (I do this with magnification set to 3 or 4). Then I increased the exposure in these areas. Of course, the WB was not the same as the sunlit areas and the colour was off, so I adjusted the WB sliders in the brushed areas.

It is fussy, but it can be done (but not for 100 images); but it does work better with RAW, so I now have another reason to shoot RAW.

In Paul's image of the singer and the choir, one could brush the area encompassing the choir's legs (total dark shadow), to bring out some detail (if desired).

I would think that this can be done in PS too.

Glenn