Same white balance but inconsistent tones

[Est2131]

Hi all,

I recently did some corporate headshots as a favour for a friend. I used a single soft box and reflector setup which stayed the same throughout. I set the WB to flash but shot in raw so that I could adjust the WB afterwards anyway.

Now that I’m processing the images in lightroom, I can’t get a consistent tone at all - even though I’ve applied the same WB settings for all the images. Some have a bluer tone, others more magenta or yellow. How can this happen given that the same conditions were used on the day and the WB settings in lightroom are the same for each image?

I’m stuck and need to deliver the images soon - though I don’t have a proper explanation to give them!

Any info or guidance much appreciated.

Thanks

 

[TimHughes]

what was the light source? has it given you trouble before?

 

[sk66]

What profile are the images using in LR? If the camera was set to something variable (like auto picture style) and LR is set to "camera settings" you can get default edits (imports) that are variable...

 

[Est2131]

what was the light source? has it given you trouble before?

Hi, it was a continuous light soft box - with a silver reflector on the other side. I’ve used it a few times but not for this kind of shoot with multiple portraits over a couple of hours.

 

[Est2131]

What profile are the images using in LR? If the camera was set to something variable (like auto picture style) and LR is set to "camera settings" you can get default edits (imports) that are variable...

I set the picture control to portrait in camera (Nikon) and the LR profile is showing as ‘adobe colour’. I’ve just tried changing this to adobe portrait for a few of the images but it doesn’t seem to change much.

 

[ecoleman]

I’m not a portrait photographer so have very little experience with artificial lighting but as these were corporate head shots where they shot in an office? Could other office lighting be having an effect on the lighting, perhaps a mix of LED and fluorescent where the LED lights will flicker creating a cooler tone when on and a warmer tone when off?

 

[TimHughes]

Hi, it was a continuous light soft box - with a silver reflector on the other side. I’ve used it a few times but not for this kind of shoot with multiple portraits over a couple of hours.

ok - if it's a steady drift over the shoot then that could be the light heating up, if it's all over the place maybe your shutter speed was too high and capturing only part of the cycle.

 

[Est2131]

I’m not a portrait photographer so have very little experience with artificial lighting but as these were corporate head shots where they shot in an office? Could other office lighting be having an effect on the lighting, perhaps a mix of LED and fluorescent where the LED lights will flicker creating a cooler tone when on and a warmer tone when off?

That’s right, they were shot in a room in an office but I switched the room lights off so the only source was the soft box light and I suppose the light from the main office through a door window though I can’t imagine that changed much and it stayed consistent.

 

[Est2131]

ok - if it's a steady drift over the shoot then that could be the light heating up, if it's all over the place maybe your shutter speed was too high and capturing only part of the cycle.

I didn’t think about the change with the lights heating up so thanks for that point.
Shutter speed was 1/160 so probably not quick enough to have an impact?

 

[Dave Goodier]

I didn’t think about the change with the lights heating up so thanks for that point.
Shutter speed was 1/160 so probably not quick enough to have an impact?

There are various types of continuous light sources and for some 1/160th could be an issue

I'd run a series of tests at home varying the shutter speed from 1/50th to 1/250th, and shoot maybe 10 images a time in burst mode so covering just a few seconds each burst; just make sure to have a grey card or just white sheet of paper in the shot so its more obvious if they is any shift in WB; and do it at night so there no chance of any ambient recording

HTH solve the mystery, though it doesn't help with sorting your current set of shots

 

[Garry Edwards]

I can't help with a solution, but there will be some experts who can, pretty much anything can be done in PP, with knowledge and a lot of time.

But as for the cause, you already have the answers. Continuous lighting has its problems, which is why most of us use flash.
There's a basic problem of poor colour rendition index. Daylight and flash have a perfect CRI of 100, which means that all colours are "read" and reproduced correctly. So do the very old tungsten lights, with a filament, but nobody uses them anymore. Both LED and fluorescent are incomplete, they lack magenta, so (to some extent) reds photograph as orange, orange as yellow, yellow as white. Sellers tend to quote a high CRI value, often around 95 or so, and if that claim is true then the problems won't be too pronounced, but it requires very specialised equipment and technical knowledge to actually measure the true figures, and some sellers lie about their own figures, knowing that they will get away with it. For example, many LED lights made cheaply for security or general illumination use only have a CRI of 60 or so and the results are terrible, but it's easy to use those components in "photographic" lighting, and they are a fraction of the cost of high CRI lights.

The biggest problems tend to come with the cheapest lights, i.e. if you're lucky, you get what you pay for but you won't get more than you pay for. The problems go away pretty much completely with the pro lights used for movies, but most of us don't spend new Tesla money on a single light for still photography . . . and we don't need to, when flash is so cheap and so good.

Then you have the low power problem. They look bright but they aren't, so unless you're shooting in complete darkness there will be other colours of light contributed from other sources, and the brighter these other lights are, the more colour contamination there will be. Again, the answer is flash because flash has so much more power, so it grossly overpowers any colour contamination from other light sources.

Then there's the shutter speed. With LED or fluorescent lighting, 1/30th second will record at least one cycle, and so will be about the best option. 1/60th will hopefully record one, anything faster is inadequate and you may or may not get a meaningful contribution from the artificial light. Flicker will be a real problem and as photographers, we need to rely on physics, luck simply doesn't work

Lights do heat up, leading to a change in colour temperature, but that change is pretty minor.

 

[Box Brownie]

I can't help with a solution, but there will be some experts who can, pretty much anything can be done in PP, with knowledge and a lot of time.

But as for the cause, you already have the answers. Continuous lighting has its problems, which is why most of us use flash.
There's a basic problem of poor colour rendition index. Daylight and flash have a perfect CRI of 100, which means that all colours are "read" and reproduced correctly. So do the very old tungsten lights, with a filament, but nobody uses them anymore. Both LED and fluorescent are incomplete, they lack magenta, so (to some extent) reds photograph as orange, orange as yellow, yellow as white. Sellers tend to quote a high CRI value, often around 95 or so, and if that claim is true then the problems won't be too pronounced, but it requires very specialised equipment and technical knowledge to actually measure the true figures, and some sellers lie about their own figures, knowing that they will get away with it. For example, many LED lights made cheaply for security or general illumination use only have a CRI of 60 or so and the results are terrible, but it's easy to use those components in "photographic" lighting, and they are a fraction of the cost of high CRI lights.

The biggest problems tend to come with the cheapest lights, i.e. if you're lucky, you get what you pay for but you won't get more than you pay for. The problems go away pretty much completely with the pro lights used for movies, but most of us don't spend new Tesla money on a single light for still photography . . . and we don't need to, when flash is so cheap and so good.

Then you have the low power problem. They look bright but they aren't, so unless you're shooting in complete darkness there will be other colours of light contributed from other sources, and the brighter these other lights are, the more colour contamination there will be. Again, the answer is flash because flash has so much more power, so it grossly overpowers any colour contamination from other light sources.

Then there's the shutter speed. With LED or fluorescent lighting, 1/30th second will record at least one cycle, and so will be about the best option. 1/60th will hopefully record one, anything faster is inadequate and you may or may not get a meaningful contribution from the artificial light. Flicker will be a real problem and as photographers, we need to rely on physics, luck simply doesn't work

Lights do heat up, leading to a change in colour temperature, but that change is pretty minor.

Purely out of curiosity and as the OP has done the shoot.

In what you cover re: issues, would the use of a WB card or "colour checker" in each shot allow for ease of PP mitigation for what the OP says.

PS a goodly while ago I used my WhiBal card on the Hindhead Tunnel walk in anticipation of lighting issues....worked a treat for corrections in post.

 

[Garry Edwards]

Purely out of curiosity and as the OP has done the shoot.

In what you cover re: issues, would the use of a WB card or "colour checker" in each shot allow for ease of PP mitigation for what the OP says.

PS a goodly while ago I used my WhiBal card on the Hindhead Tunnel walk in anticipation of lighting issues....worked a treat for corrections in post.

The answer is always to avoid the problem by getting the lighting right but yes, the inclusion of a grey card may have helped.

I'm just wondering - but I'm far from expert in this - whether there is something pure black in every image that can be used to set the correct colour balance, allowing each image to be corrected individually?

 

[Dave Canon]

I use a bespoke calibrated profile in LR which I created using the x-rite colour checker. When a shot is very critical I also take a shot of the colour checker in the same light as used for the subject. This enables me to get a precise value for WB as well.

Dave

 

[TimHughes]

It's fairly easy to spot the dodgy videos, just pay no attention to those that are
1. About the alleged benefits of a product
2. Are made by people who promote their own paid content (usually online courses)
3. Illustrate their videos with beautiful photos allegedly taken with the promoted equipment, i.e. heavily retouched photos
4. Are famous for being famous, not for their actual photography.
5. Fast talkers, clearly reading from an autocue.

My own videos were always honest, but still deceptive - something that takes an hour to do ends up as 6 minutes, because the attention span on social media is short, so everything got edited down. For example, I might move a light 6 times, an inch or 2 at a time, or adjust the power setting a few times, but in the video this might be shown as just one adjustment or none, and I always took the view that retouching photos for instructional videos is the wrong thing to do, let people see what you actually get SOOC, so they won't try to get something that is actually impossible to get, real-world.

The one I posted above, on lighting a white background, is a case in point. except for (I think) the opening shot, I lit ONLY the background, because lighting the background was what it was all about, and all the light on the model was unwanted reflected light from the background, to show faults, but there were still adverse comments about the lighting on the model, which just goes to show that people don't watch or listen properly.

All of my videos were terrible, because I'm not a happy smiley chappy by nature, so they didn't suit the YouTube algorithm and didn't get the view numbers that justify the cost, but that's just me. My very early ones, filmed with a camcorder entirely without help, were even worse than the later ones but, in a way, were more complete.

I didn’t think about the change with the lights heating up so thanks for that point.
Shutter speed was 1/160 so probably not quick enough to have an impact?

There are various types of continuous light sources and for some 1/160th could be an issue

I'd run a series of tests at home varying the shutter speed from 1/50th to 1/250th, and shoot maybe 10 images a time in burst mode so covering just a few seconds each burst; just make sure to have a grey card or just white sheet of paper in the shot so its more obvious if they is any shift in WB; and do it at night so there no chance of any ambient recording

HTH solve the mystery, though it doesn't help with sorting your current set of shots

A couple of experiments to try at home to validate the cause and if you can mitigate:
- range of shutter speed
- with and without any anti-flicker feature if your camera has it

 

[sk66]

I use a bespoke calibrated profile in LR which I created using the x-rite colour checker. When a shot is very critical I also take a shot of the colour checker in the same light as used for the subject. This enables me to get a precise value for WB as well.

Dave

It allows you to get about as close as possible; but it isn't actually "correct." And it is "incorrect" if the conditions are any different from those used to calibrate under. Colorchecker calibration is more like "averaged error" than actual accuracy.

I'm just wondering - but I'm far from expert in this - whether there is something pure black in every image that can be used to set the correct colour balance, allowing each image to be corrected individually?

Unfortunately, LR does some pretty odd things when setting WB... especially if referencing an area of black or white (I cannot explain what it is doing, or why). It used to throw an error if you selected an area w/o enough variation (e.g. black, RGB 0/0/0)... now it seems to ignore the sampled area instead and does something else.

Curves and levels adjustments in Photoshop (ACR) allow you to reference and set black or white (or both), and curves will allow you to reference middle grey as well. But each sample overwrites/refines the previous one. E.g. I usually sample black first, and then white, so that any residual color cast exists primarily in shadows/blacks where it is less apparent. IME, most images will have something that is (should be) white or black that will allow you to set colors close to true... that is often a pupil or catch light.

 

[Dave Canon]

It allows you to get about as close as possible; but it isn't actually "correct." And it is "incorrect" if the conditions are any different from those used to calibrate under. Colorchecker calibration is more like "averaged error" than actual accuracy.

What I described is what Martin Evening recommended and he is a professional fashion photographer. A member of my club is a wedding photographer and was using two identical Nikon DSLR's but with different focal length lenses. However, he found that colours could be noticeable different between the cameras but generally only relevant for critical clothing colours. By calibrating both cameras he found he could get very consistent results.

Also when I calibrate I create more than one calibration profile say for daylight, flash and tungsten light. In practice I have found that creating a dual illuminent profile when it is applied LR checks the measured WB and interpolates between the the two calibration profiles which turns out to be pretty accurate. This is using the Adobe DNG editor rather than the X-rite software. In reality, most the images I take (a very wide range) are for myself and competitions so do not need to be that accurate as the viewer will not normally know exactly how a colour should look. For fashion shots and weddings, I would always wish to use a calibrated profile. The profile is used automatically so. having set this up, it is no trouble. Incidentally I have found by observation that (at least for my Canon 5D4, "camera standard" is good enough for all but the most critical situations. I have calibration profiles for my other cameras as well, including a Sony A6600 and a infrared converted EOS 20D. Sadly Martin Evening has passed away but his detailed advice on calibration is published in several of his Adobe books.

Dave

 

[sk66]

In practice I have found that creating a dual illuminent profile when it is applied LR checks the measured WB and interpolates between the the two calibration profiles which turns out to be pretty accurate.

I hadn't heard of dual illuminant profiles before; I'll have to look into that.... (I do know about dual ISO presets). I use a calibrated profile frequently, but if you check the RGB values of the color checker patches they won't be matches (all/any).

Most things are not color critical and white balance becomes a preference thing (e.g. leaving the warm sunset tint); there's no right or wrong really.

 

[the_accidental]

The suggestion to test your lights above at different shutter speeds is a good one, but you must test them at different brightness levels (or the level you had for your shoot).

The reason is many LED (esp cheaper ones) use PWM (pulse width modulation) to control brightness - that means they actually flicker.

I’ve had the issues you describe when shooting dslrs on a robot with continuous lighting, colour shift at different parts of the PWM cycle. Though often you’d see shifts in exposure too - though if you were on auto iso you might see that as changes in ISO with changes in colour rather than changes in exposure?

If you can somehow sort by colour cast (maybe on iso) then you have a chance of applying different WB automatically otherwise you’re stuck per photo…


They wouldn’t happen to want b&w headshots? Solves the problem and was trendy for a while ;-)

 

[Garry Edwards]

The suggestion to test your lights above at different shutter speeds is a good one, but you must test them at different brightness levels (or the level you had for your shoot).

The reason is many LED (esp cheaper ones) use PWM (pulse width modulation) to control brightness - that means they actually flicker.

I’ve had the issues you describe when shooting dslrs on a robot with continuous lighting, colour shift at different parts of the PWM cycle. Though often you’d see shifts in exposure too - though if you were on auto iso you might see that as changes in ISO with changes in colour rather than changes in exposure?

If you can somehow sort by colour cast (maybe on iso) then you have a chance of applying different WB automatically otherwise you’re stuck per photo…


They wouldn’t happen to want b&w headshots? Solves the problem and was trendy for a while ;-)

Good thought about the possible B&W solution, and also very useful extra info on the flicker problem . . .

This is really all about the unsuitability of continuous lighting for still photography. It’s perfect for video, where faults either don’t show at all or where they don’t matter in the least, and where there’s no economical choice anyway, but the required quality standard for still photography is in a different league entirely.

Chinese manufacturers started making continuous lighting for still photography because they had no choice. Aggressive and very successful marketing of flash products by one Company (Godox) completely killed off nearly all other flash manufacturers, who simply couldn’t compete, so the others simply moved to continuous lighting, which Godox wasn’t bothering with at that time.

Also, the profit margins on continuous lighting, a few years ago at least, were extremely high – just buy in cheap components designed for non-critical applications such as vehicle lighting or security lighting, stick them into a fancy box, add a couple of almost zero-cost electronic controls, make false claims about the quality of the light and they then have a product that sells into our specialist market at a very high price.

And the other big advantage, for the manufacturers, is that continuous lighting has a very low entry bar. Just about any backstreet factory can make them, they don’t need skilled personnel or expensive machinery.

I’m going to leave this thread where it is in PP and image editing for now, in the hope that someone can help with correcting the issues, but it’s really a lighting matter, so I will eventually move it to lighting.

 

[the_accidental]

Just to update this, I’ve actually seen exactly the thing in the OP today, and I think it was what I described.

I was shooting this video, so had LED lamps on for a bit of light:

https://www.talkphotography.co.uk/t...-else-welcome-mega-thread.511885/post-9563625

Afterward I checked the images I had on the still camera - subtle WB shift between frames.

They were dimmable LED room lights that will be PWM for sure, so I’m fairly certain that’s what it was. Though I did also notice that was despite fairly long exposures (I was seeing it at 1/6th) - it was pretty subtle changes though.