Banding Issues with Artificial LED Lights

Since two years I’ve been fighting banding issues on the Canon EOS R5 with artificial LED lights.

The basic theory is easy: Use a shutter speed 1/x for a frequency y of the PWM LEDs where you can divide y / x with no remainder.

Let’s take Philips Hue as an example where the light bulbs operate at 1 kHz. In this case shutter speeds of 50, 100, 125, 200, 250, 500 & 1000 should work without any banding at all. But I still get banding e. g. at 250, which shouldn’t be the case here as 1000 / 250 = 4. The banding gets better at 250, compared to let’s say 400, but it is still visible, usually in burst shots with uniform backgrounds.

It doesn’t matter if I have “Anti-Flicker Shooting” on or off (which accordingly to the manual only works for 100 & 120 Hz, so no wonder), and when I press the Info button when checking for flicker it doesn’t detect any.

The first thing that puzzles me is why the formula doesn’t work for the R5 body.

The second important thing that I want to mention is that I use several bodies from different manufacturers like Fujifilm, Ricoh, Olympus etc. but only the R5 exhibits this behavior.

How do you deal with uncontrollable LEDs on your camera bodies? Do you see any difference between manufacturers, or even different bodies by the same manufacturer?

Are you using electronic shutter? You have to use mechanical shutter to alleviate banding.

Yes, I use MS which gives me slightly better results than EFCS (where many people on the internet say that both shutter types should be affected by banding in the same way, something that I cannot confirm).

I’ve seen some posts here on the forum, some of them more than four years old, but nobody seemed to have a satisfying answer. I hope some smart photographer can help me out this time.

I’ve also read several times that Philips Hue modulates at about 600 Hz, but this is also not true in my case. I can just point a phone with the camera app open at the light source, and I immediately see the effect of different shutter speeds on the screen without even recording.

It would be also good to know why I observe this behavior mainly on the R5, and why all of my smaller sensor bodies don’t seem to be affected by that issue (slower sensor and / or slower sensor readout time?).

Shooters on My Squad wrote:

Since two years I’ve been fighting banding issues on the Canon EOS R5 with artificial LED lights.

Perhaps try genuine LED lights.

As I mentioned in my initial post, my question is about uncontrollable light, often in dimly lit environments (I gave Philips Hue just as an example to keep things simple, because it is quite common these days).

I just don’t get it how some bodies seem to be able to defy physics. What’s going on here?

Shooters on My Squad wrote:

As I mentioned in my initial post, my question is about uncontrollable light, often in dimly lit environments (I gave Philips Hue just as an example to keep things simple, because it is quite common these days).

I just don’t get it how some bodies seem to be able to defy physics. What’s going on here?

You didn't get the joke, did you? That out of the way I have never had that issue with any of my Sony cameras. I don't know what Sony does to prevent it.

I’m still contemplating until God will send me a truly genuine, fully natural light-emitting diode down to earth 

But let’s be a little bit more serious again. I don’t know what kind of information would help to mitigate this issue. MS works best for me, with EFCS a close second. Fully ES will give you that typical banding, and it gets worse the faster your shutter speeds are. I’m not sure which kind of settings besides Anti-Flicker, shutter speed, and shutter type could affect banding?

I think if this would be caused by power it would be 120 Hz, but PWM LEDs should run on their own cycle, with cheaper LEDs usually running at a lower frequency.

I think I should get massive banding at 1/1000 when pointing my phone at the LED if you were right, but instead I get uniform images in the camera app?

And to further underpin my argument that it should have nothing to do with the power cycle, I’ll refer to my first post where I described the “Anti-Flicker Shooting” setting behavior. I would assume that the camera would detect something if it would be dealing with a 120 Hz cycle, but it doesn’t do it. No matter if I point the camera directly at the light source, or slightly nearby, a press on the “Info” button in the corresponding Quick Control menu just gives me negative feedback.

Still wondering what is going on here with the Canons, and why other manufacturers seem to handle this issue better? This kind of makes a professional camera unusable in amateurish conditions, if you’re not OK with image degradation due to banding issues.

Shooters on My Squad wrote:

Since two years I’ve been fighting banding issues on the Canon EOS R5 with artificial LED lights.

The basic theory is easy: Use a shutter speed 1/x for a frequency y of the PWM LEDs where you can divide y / x with no remainder.

Let’s take Philips Hue as an example where the light bulbs operate at 1 kHz. In this case shutter speeds of 50, 100, 125, 200, 250, 500 & 1000 should work without any banding at all. But I still get banding e. g. at 250, which shouldn’t be the case here as 1000 / 250 = 4. The banding gets better at 250, compared to let’s say 400, but it is still visible, usually in burst shots with uniform backgrounds.

It doesn’t matter if I have “Anti-Flicker Shooting” on or off (which accordingly to the manual only works for 100 & 120 Hz, so no wonder), and when I press the Info button when checking for flicker it doesn’t detect any.

The first thing that puzzles me is why the formula doesn’t work for the R5 body.

The second important thing that I want to mention is that I use several bodies from different manufacturers like Fujifilm, Ricoh, Olympus etc. but only the R5 exhibits this behavior.

How do you deal with uncontrollable LEDs on your camera bodies? Do you see any difference between manufacturers, or even different bodies by the same manufacturer?

That formula should work if the camera's idea of frequency exactly matches the LED's idea of frequency. Especially for cheap LEDs, I bet there's a mismatch. And that's probably the main source of your issue. Unfortunately, your camera probably is not very tunable (there are discreet values you can choose from), and the LED lighting probably has only one frequency, which, if you're lucky, is constant. But there's likely to be a mismatch.

https://www.youtube.com/watch?v=DdToSNgPnO8

Notice in this video that when the frequencies match (60 Hz) the light gradually changes. That's because they're not both at exactly 60 Hz.

Then there are light sources like with some of the new M1 Macs that vary the frequency of the display to save on battery (maybe it's not the light source but the pixels that vary in frequency). Such is probably not the case for LED lighting, though.