Banding/strobing issues with LED lights (Hydra 26's)

[i_declare_bankruptcy]

I have a Sony A7rII and am experiencing some pretty sever shot-breaking strobing issues with my Hydra 26's. I know mirrorless cameras are very prone to this but I'm getting frustrated that my choice of lights is penalizing me for photography with my nice camera! As my shutter speed approaches ~1/400s I start getting strobing sever enough to ruin the shots. Sticking below 1/250s seems to keep it at bay but that makes macro photography and photographing fish nearly impossible.

I also wonder if my lighting intensity affects it (as in how are the LED's dimmed). I run my Hydra's at only like 10-20% because I have such a small tank, so I wonder if that makes the banding/strobing worse.

Has anyone experienced this and what did you do to overcome it? I know it's impossible to "fix" the issue but maybe there's something I'm not thinking of?

 

[TheEngineer]

This is fairly typical of LEDs. Since they are eminating light from a small area you get what most people call shimmer or the disco ball effect. This comes from the light I teracing with ripples in the water surface.

If you turn off all flow in the tank it should go away while you take pictures.

 

[i_declare_bankruptcy]

Ah I'm referring to a different effect actually, one that cause what's called banding or strobing (black lines or dark areas accross the image). Sorry for the confusion

 

[TheEngineer]

Can you post an example? If you were catching flickering LEDs I’d suspect it wouldn’t cause bands.

 

[i_declare_bankruptcy]

Yes, sorry, meant to put it on my previous post. This is from my cellphone camera but same happens on my mirrorless. You probably know this (TheEngineer) but it's caused by LED's operating at some frequency, let's say 60hz, that we can't distinguish but still affects the camera's sensor readout since it reads in some particular order (top to bottom, line by line usually), so those 'strobes' cause 'bands' in the final image due to interference in the readout. I tried selecting multiples of what I thought was the LED's operating frequency but that theory failed at higher shutter speeds anyways.

Not being able to shoot above 1/200s is making photography of moving subjects (not that I have much in my new tank) difficult.

I don't think it can be perfectly fixed, I'm more wondering if changing intensity on the Hydras will increase or decrease that frequency and potentially improve my shots (of course, this is a risk for the inhabitants). Or perhaps some other trick.

 

[TheEngineer]

LEDs don’t run on AC, rather DC voltage. The 60Hz flicker you are referring to happens in incandescent and fluorescent lamps. Some LEDs are dimmed using pulse width modulation (PWM) at a rate faster than our eyes can detect but that would be universal flickering, not banding.

What you seem to be picking up appears to be from polarization of the glass if I’d have to guess. Do you see this if you take a picture straight down through the water? (You will need to turn your pumps off or use a look down box)

 

[TheEngineer]

Considering there is surface movement of the water and those bands are evenly spaced, in almost certain this isn’t your light.

 

[BigJim]

There was another recent thread about this issue. I don't know if there is a solution, but maybe you can find something that helps.

https://www.reef2reef.com/threads/leds-banding.308898/

 

[TheEngineer]

I’m still skeptical. The shots in that thread are through glass. I’d want to see this happen through the water.

 

[pickupman66]

I have a pair of Hydra 26HDS and i have never experienced any banding with them thru either my Nikon D3000 or my Phone with leica cameras.

Leica camera

Nikon

 

[i_declare_bankruptcy]

Interesting. I don't know how the Hydra's work but since I'm running them so low, like 13%, do you think that if they did operate with PWM that could be part of it? I know you're skeptical as I am too -- I did try shooting with a polarizer and no luck at any angle. I'll try top-down and pumps-off when I get a chance. The banding is even worse on my a7rII where literally half the image will be completely black, the entire image, 1/3 of it, etc at shutter speeds over around 1/400s. Slower shutter speeds were the only thing that fixed it. Unfortunately I don't have those pictures on my development machine but I'll upload when I get a chance.

I also don't understand how the glass would be significantly or completely polarizing the light on its way out of the tank. I understand the light is refracted through the water and reflected off surfaces within the tank, but I don't see any difference in the tank when shooting with a polarizer at any orientation, although it obviously does a great job of cutting external reflections (like windows, etc).

It does depend on how the sensor is read. There is a lot of variability in cellphone camera hardware as well as DSLR's and mirrorless. It doesn't surprise me that your Nikon doesn't have issues due to its construction/operation. Using LED strobes for professional photography can also cause issues with mirrorless cameras so that's originally where I got the idea.

However, surface agitation also is a great theory and worth testing. Here are some things I'll try to see if we can narrow down a cause:

Shoot with pumps off
Shoot top down with pumps off
Shoot with pumps on
Shoot top down with pumps on
Shoot at different intensities to test PWM theory while maintaining constant shutter speed
Shoot at different shutter speeds; try to find 'sweet spot' at higher shutter speeds and constant light intensity

Any other things I should test?

Here's another example of strobing from my phone:

Ooh, if you look closely you can actually see the breakup of strobing lines and you can also see them stretch to the edge of the frame. Interesting breadcrumbs.

I'm not understand how surface agitation causes this because then wouldn't you just see 'hot spots' in the image where there's constructive interference?

 

[Chris Wells]

Wow, that is interesting.
I haven't seen those bands either. Samsung S9 I have 2-Hydra 26HDs on 90G Tank

 

[Chris Wells]

I work heavily in the glass industry and that looks like glass draw lines from float process. I would presume it is tempered after that.
If so, you would only being seeing them if you were using some sort of polarized filter.

Just my 2 cents.

Chris

 

[i_declare_bankruptcy]

Look at the picture I posted just now though. The lines extend beyond the tank. Shooting with or without a polarizer, and using either the cellphone or a7rII produce banding.

 

[ajm83]

I wouldn't be surprised if this effect wasn't from PWM to be honest. It's a bit like how car LED lights flicker on car shows like Top Gear, except you're seeing it happen multiple times during a single frame (guess your camera scans horizontally?)

Why don't you whack the light up to 100% very briefly and see if the effect changes/disappears? In that case it should be on a lot more time than it is off.

 

[Chris Wells]

Look at the picture I posted just now though. The lines extend beyond the tank. Shooting with or without a polarizer, and using either the cellphone or a7rII produce banding.

I see what you are talking about. I went and tried to generate bands with mine... uv and violet effect light bands, but nothing like what you are seeing. You are seeing vertical and horizontal lines...the light bands that I was effecting were normal and illuminating directly from the LEDs.

Do all colors create same effect?

 

[i_declare_bankruptcy]

Yes, that's the plan for tonight (if I can get a chance). I think turning them up to 100% on one channel should reduce PWM (if that's how they work?) and maintaining the same shutter speed as in the banded photos at lower intensities to keep things consistent. I'll also introduce a different light source (cheap incandescent) to see if that causes the same issue when viewed through the glass. Or perhaps a different LED source that doesn't have adjustable intensity.

I'm excited to get to the bottom of this, there are a lot of interesting ideas on this thread!

Chris, I will check color spectrum if I cannot fix the bands via intensity. Thank you!

 

[Chris Wells]

I wouldn't be surprised if this effect wasn't from PWM to be honest. It's a bit like how car LED lights flicker on car shows like Top Gear, except you're seeing it happen multiple times during a single frame (guess your camera scans horizontally?)

Why don't you whack the light up to 100% very briefly and see if the effect changes/disappears? In that case it should be on a lot more time than it is off.

I thought of that too, I am just hung up on that they are vertical and horizontal. I am definitely stumped.

C

 

[ajm83]

I thought of that too, I am just hung up on that they are vertical and horizontal. I am definitely stumped.

C

Depends which way round he physically had the camera during the shot* (unless I'm missing a shot which is showing banding in both directions?)

*hope I'm not teaching granny to suck eggs here but think about how a digital camera captures the image from the sensor. It reads a column of data, next column of data, next column...etc for the whole frame, left to right.

If the LED is off when the data in a particular column is read then you get a dark line at that point in the image.
Flip the camera on it's side, and instead it reads row, row, row...
So you get a dark line the other way (horizontally)

 

[i_declare_bankruptcy]

I thought of that too, I am just hung up on that they are vertical and horizontal. I am definitely stumped.

C

Vertical and horizontal is just due to how the sensor reads. They're actually always 'vertical' on the cellphone (only images I supplied in this thread) -- one is short portrait and then one is shot landscape, hence the difference in orientation. The banding itself is not actually changing orientation -- I just changed the camera's orientation.

Edit: Dang you beat me ajm83