Bouncing flash

Barney

Suspended / Banned
Messages
3,043
Name
Wayne
Edit My Images
No
What white balance should I use when bouncing flash?

if I select flash on the camera WB then I assume I will get the colour of the wall in the output when bounced,

Will the colour introduced from the bounce have a conflict with the setting and be more difficult to correct in PP and with other light in the scene lead to the dreaded "mixed lighting"?
 
You've identified a real issue when using bounced flash, the colour of the bounce surface will always affect the colour temperature, sometimes by a little and sometimes by a lot.

If you are only using one light source, i.e. if all of the light is bounced, it doesn't matter because it can be corrected very easily. Personally I never use AWB, I just set the colour temperature to 5500K, shoot in .raw and then correct in PS. But, if the light sources are mixed, it becomes far more difficult (almost impossible) to correct in pp, and this applies to all mixed source shots, not just when bounced light is used.
 
If your aim is absolute colour accuracy, then bounced flash is a nightmare!!

However, I’ve rarely ever shot for absolute colour accuracy (and if I did, I understand I’d need a very controlled environment) so bouncing flash using auto WB and RAW as a safety net is a piece of cake.

Indeed, white balance is rarely a headache, coming from film where I never found it necessary to buy a multitude of correction filters (instead letting the lab adjust my prints), I now don’t use absolute WB and adjust in post (effectively the same as before)
 
I see your problem. I have a frame which mounts above my flash gun and hold a piece of white card to create the surface for the bounce and no colour change. I normally use only Raw and AWB but if working in s studio with all flash I will often set the WB to 5500. I am sure that these flash bounce frames are easy to find and think it was a standard Canon part,

Dave
 
I always shoot in daylight and correct in post - for critical applications in complicated mixed lighting, I have a small mid grey bounce card I also take a shot with the same settings and space I can use as a basis of the correction. In very rare cases where the cast can’t be easily removed, I do a mono conversion
 
I see your problem. I have a frame which mounts above my flash gun and hold a piece of white card to create the surface for the bounce and no colour change. I normally use only Raw and AWB but if working in s studio with all flash I will often set the WB to 5500. I am sure that these flash bounce frames are easy to find and think it was a standard Canon part,

Dave
I have since checked and it is called Vivitar Bounce Diffuser Card Holder.

Dave
 
Anytime you add a second light source that is not the same color as the first you end up with a mixed lighting scenario. And that is more common than not; flash/sunset, flash/blue hour, flash/incandescent, etc, etc. But as with Phil, it is seldom a problem. Just avoid bouncing off very strongly colored surfaces (e.g. bright red walls). The WB setting selected isn't going to make a big difference either; it can only make one global adjustment.
 
Last edited:
Best way to avoid obvious mixed lighting mismatch is to have 1 clearly dominant light whether it is flash or ambient, unless you can obviously match them up.

Color temperature can swing wildly even if bouncing of mildly colored surfaces. Remember that white is not necessarily a true white and may come loaded with notable blue component.
It is best to correct wb by sampling a true white or grey area on your subject, but best of all do a shot with colour chart. They cost a little, but are not that expensive
 
Our house is a 1960s build and has the then fashionable 'textured' ceilings. They are wonderful at scattering bounced flash and create a superb effect.

My wife has formulated a redecoration plan which includes skimming over the texture because she doesn't like it.

I think my softly lit indoor pics are a thing of the past!
 
There’s some things I know enough about to know I’m an idiot.

And white balance is one of them. If I say the other one is calculating bounced flash distance, then you might know where I’m heading.

If I’m bouncing flash off a ceiling, I might be able to guess that the flash distance to subject is about 8 ft, then I could use the GN of the flash at the set zoom angle to calculate the exact amount of manual flash I need for the aperture I’ve set (then recalculate as I move, my subject moves or I change aperture). But that’s tons of information to compute! If I use ETTL, the flash does it in less than the blink of an eye, and brilliantly. I might want to add a little FEC, but compared to the work done by the camera, I’m fiddling round the edges.

Likewise white balance. I could buy a colorimeter, an expensive piece of kit that will give me an exact colour temperature I could set in camera, I could use a colorchecker passport, take a test shot and set a WB in camera, or slightly more lazy, use that test shot to correct the WB in post. I could use a white card to take a reading off, not quite as accurate but also works.

Or I could use AWB, then check the temp on screen whilst processing and batch process my images in seconds for free and stress free.

Some of the automation in our cameras allows us to do really complicated things simply. And calculating colour temperature is definitely on that list for me.

Add in the complication that my bounced flash on the subject is maybe not the same colour temp as whatever is lighting the rest of the scene, and we then get to the question of ‘what is a correct white balance’.
 
Last edited:
With some cameras you can use live view and scroll through the WB options, this will sometimes help get the room WB, this may give you an idea of how different your flash is likely to be. You can also get gels for the flash to "correct" it to the room lighting in critical cases.
Im my experience, in most cases I just sort it in Lightroom,
 
There’s some things I know enough about to know I’m an idiot.

And white balance is one of them. If I say the other one is calculating bounced flash distance, then you might know where I’m heading.

If I’m bouncing flash off a ceiling, I might be able to guess that the flash distance to subject is about 8 ft, then I could use the GN of the flash at the set zoom angle to calculate the exact amount of manual flash I need for the aperture I’ve set (then recalculate as I move, my subject moves or I change aperture). But that’s tons of information to compute! If I use ETTL, the flash does it in less than the blink of an eye, and brilliantly. I might want to add a little FEC, but compared to the work done by the camera, I’m fiddling round the edges.
Yes, let's not over-complicate a simple process.

ETTL is the obvious tool, for most situations, it will sometimes get it "right", it will nearly always be good enough and it will never be totally wrong. For pretty much all "quick" shots, where there's little time to experiment, it's a no-brainer. For something like still-life shots, where there are no time pressures, we can always set the flash power manually and should, usually, get better results - but casual viewers are unlikely to see much if any difference.
Likewise white balance. I could buy a colorimeter, an expensive piece of kit that will give me an exact colour temperature I could set in camera, I could use a colorchecker passport, take a test shot and set a WB in camera, or slightly more lazy, use that test shot to correct the WB in post. I could use a white card to take a reading off, not quite as accurate but also works.

Or I could use AWB, then check the temp on screen whilst processing and batch process my images in seconds for free and stress free.

Some of the automation in our cameras allows us to do really complicated things simply. And calculating colour temperature is definitely on that list for me.
When I did this for a living I used a colour temperature meter routinely, because when clients are paying a lot of money for the shots they need to be right, not somewhere near, but the world has changed, and although I still have one somewhere, I haven't seen it for years because it's no longer relevant as far as I'm concerned. Originally of course, I shot everything on film, if the colours were wrong there was nothing that could be done about it, and even when the trannies were scanned, allowing PP work, the software wasn't that good and computers were too slow to facilitate effective colour correction. When digital took over things got a bit easier but it was still essential to get the colours right, and the way that I did it was to filter individual lights with CC gels, to match them.

Back then, one of the problems was that colour temperature varied according to the power setting of the flash, pre IGBT. There are still variations with IGBT but they are far less, and the cheaper flash heads were pretty terrible in terms of consistency, and the bold claims made by most manufacturers were simply lies, they knew perfectly well that hardly any of their customers had the skills, equipment or interest to check for themselves. There are still good reasons to stick to a single make of flash, and to not mix IGBT with non-IGBT, and then there's the massive colour temperature variations caused by different modifiers, e.g. softboxes that have yellowed, and despite wild claims about CRI, continuous lighting causes problems too, although many of the LED lights aren't too bad today.

Add in the complication that my bounced flash on the subject is maybe not the same colour temp as whatever is lighting the rest of the scene, and we then get to the question of ‘what is a correct white balance’.
Good point, and especially when people create "fill" by using a longer shutter speed to include whatever ambient light happens to be present.

"Correct" white balance is neutral, but I once took a portrait shot and changed the colour temperature in PP, starting with "correct" and ending with 500K too warm, a massive difference, but everyone who saw those shots preferred the most warm version:)

And there's another factor too - everything used to be presented in print form, but today very few shots are printed and the vast majority are just viewed on a computer or phone. I have 3 monitors, 2 are calibrated and so should be "right", one is just used for documents and is a cheapie that isn't calibrated. If I look at a shot on both a calibrated and on the non-calibrated monitor then the differences are dramatic, but most people don't have a calibrated monitor, so most of the effort that can go into getting correct white balance is just wasted.
 
but most people don't have a calibrated monitor, so most of the effort that can go into getting correct white balance is just wasted.
Sad truth... when I sell a digital image usage right I put a disclaimer; but at least I know any resulting issue is not my fault. If I were selling finished prints regularly I think I would have to offer a viewing room option as well (or small sample prints).
 
How many also use a camera calibration profile. I have profiles for all my cameras in LR. Usually the colour differences are fairly subtle compared to the standard profile.

Dave
 
How many also use a camera calibration profile. I have profiles for all my cameras in LR. Usually the colour differences are fairly subtle compared to the standard profile.

Dave
You need one for each different lighting condition... IMO it is too much effort.

Plus, it's not really "correct" either. I.e. it degrades some colors which were more accurate in order to correct colors that were less accurate... it's a better average accuracy overall. And other things also cause similar effects. I.e. if you underexpose an image and then recover it in post it will shift the color accuracy; even with raw files.
 
You need one for each different lighting condition... IMO it is too much effort.

Plus, it's not really "correct" either. I.e. it degrades some colors which were more accurate in order to correct colors that were less accurate... it's a better average accuracy overall. And other things also cause similar effects. I.e. if you underexpose an image and then recover it in post it will shift the color accuracy; even with raw files.
Not so. I use Adobe's DNG profile Editor and I create a Duel Illuminent Profile for which you effectively create a profile for sunlight (6500k) and tungsten (2850k) which it combines to give this Duel Illuminent Profile for the specific camera. LR then interpolates the data according to the White Balance. It probably takes 15 -20 mins to produce such a camera profile but then you have that for the rest of the cameras life so very little effort. I got the information on this from Martin Evening the LR/PS Guru who wrote books for Adobe. To be fair, Martin suggested that this amount of precision is only required by a Fashion Photographer (which he was at the time) or a high end wedding Photographer. At my club the few who do calibrate in this way can produce almost identical images (Colour wise) at the same model shoot but using completely different cameras. Even more interesting for some is the fact that the DNG editor has an infrared section. I rarely need this amount of precision for me but it was a challenge originally because a shot of a girl with red hair looked yellow/orange from my Canon and dark red on a friends Nikon and both were wrong. We both calibrated and resulting images were then almost identical so I have continued with calibration.

Dave
 
At my club the few who do calibrate in this way can produce almost identical images (Colour wise) at the same model shoot but using completely different cameras.
Unfortunately, consistency does not mean accuracy. I.e. if you take a picture of a color checker with known RGB values for the patches, in daylight using the daylight WB setting; and you apply the calibration in post, some colors will be more accurate and some colors will be less accurate. And I don't know of any solution for that problem other than individual/selective corrections in post. Luckily, critical color accuracy is almost never a requirement; and if it is, it is limited to a very few specific colors (i.e. the product).

And if critical accuracy is not required, then what is "correct colors" for a given situation? Is it the colors that were apparent at the time? Or is it the colors corrected by removing the lighting contribution that existed at the time? That really is just a subjective choice...

I will add the caveat that I haven't messed with camera calibration in any great detail for the last couple of years. but I don't think the root issue is actually correctable.
 
Unfortunately, consistency does not mean accuracy. I.e. if you take a picture of a color checker with known RGB values for the patches, in daylight using the daylight WB setting; and you apply the calibration in post, some colors will be more accurate and some colors will be less accurate. And I don't know of any solution for that problem other than individual/selective corrections in post. Luckily, critical color accuracy is almost never a requirement; and if it is, it is limited to a very few specific colors (i.e. the product).

And if critical accuracy is not required, then what is "correct colors" for a given situation? Is it the colors that were apparent at the time? Or is it the colors corrected by removing the lighting contribution that existed at the time? That really is just a subjective choice...

I will add the caveat that I haven't messed with camera calibration in any great detail for the last couple of years. but I don't think the root issue is actually correctable.
I suggest you argue this with people like Martin Evening. The consistency aspect can be very important. For example a colleague who is a wedding photographer takes two cameras with different lenses as there is no time to keep changing lenses. Before calibration the colours varied considerably between the cameras but after calibration they looked identical. The cameras were identical Nikon models. The customers did not expect to see different colours between shots but were not carrying out scientific tests to identify the precise colour. I do not need such precision for most of my photography but my camera specific profiles consistently provides me with pleasing results whereas the default Adobe profiles do not.

Dave
 
The consistency aspect can be very important.
Of course it can. That's why using auto white balance for a shoot/event with constant/controlled lighting can be more of a problem. But consistency is not accuracy, and camera calibration does not generate accurate colors. It only results in a better average accuracy. And in terms of consistency, using a calibrated WB setting isn't really any better than using any other fixed WB setting (except maybe in the case of attempting to match different cameras).

I do not need such precision for most of my photography but my camera specific profiles consistently provides me with pleasing results whereas the default Adobe profiles do not.
I can't argue with that, I don't use Adobe profiles either. I don't even just use a profile, I have lightroom automatically apply a preset which includes a profile along with other settings.
 
Last edited:
Back
Top