Explain this to me

[Robin Usagani]

I messed around with my flash the other day and I am confused why this is happening. What is the reason the light streaks I get from moving the camera are bunch of dots? Why not straight lines? Is it cause by the way our sensor works? Will the same thing happen to a film camera? Is it caused by the light flickering (i didnt notice it with my eyes)? Can someone explain?

 

[pgriz]

If those were LED christmas lights in the background, then yes, they do flicker.

 

[480sparky]

LEDs are direct-current devices. They are plugged into an alternating-current source. So the LED can only light when the 'polarity' of the AC sine wave is moving one direction and not the other.

Since an AC source changes directions 120 times a second (60 times one way, 60 times another, hence 60-cycle or 60-Htz) the LED are only lit for half of each 1/60-sec cycle.

Using simple math, I could tell you your shutter speed was around 1/5 to 1/6 sec. by counting the pulses.

 

[Robin Usagani]

Very close! 1/13! Ok, that makes sense now. So if I had shot it more than 1/60 and move my hand a lot faster, it should be continuous streak?

 

[Joey_Ricard]

LEDs are direct-current devices. They are plugged into an alternating-current source. So the LED can only light when the 'polarity' of the AC sine wave is moving one direction and not the other.

Since an AC source changes directions 120 times a second (60 times one way, 60 times another, hence 60-cycle or 60-Htz) the LED are only lit for half of each 1/60-sec cycle.

Using simple math, I could tell you your shutter speed was around 1/5 to 1/6 sec. by counting the pulses.

Yeah, what he said! I just didn't want to look like I was showing off....lol

 

[480sparky]

Very close! 1/13! Ok, that makes sense now. So if I had shot it more than 1/60 and move my hand a lot faster, it should be continuous streak?

Theoretically, yes. But there's a chance that 1/60 sec could still bridge the time space between two pulses if your shutter mechanism isn't 100% (and none of them are!). You may really end up with a slightly longer exposure and have parts of two pulses show.

They may not have them show up at all, either...... if the shutter speed is actually 1/70, you may record the time when they aren't lit.



Or, you could build a Full-Wave bridge Rectifier for the lights to plug into and not have to worry about this issue.

 

[Tony S]

In North America our power grid supplies AC power at 60hz, which in simple terms means the power turns on and off 60 times per second. To the eye if the lights are stationary you don't notice the flicker rate ,sometimes if the lights are moving you can see the flicker rate with your eyes.
Taking photos of the lights in a long exposure (something longer than 1/60) of moving lights will show the flickering like your shot as separate light spots. Often times when shooting lighting inside with multiple shots your color balance and exposure can also be affected by the cycle rate of the lights from one shot to the next. So if you are shooting in manual mode and notice a change in exposure/color balance it's not the camera messing up, it's the cycle rate of the lights. It's an effect visible with most types of lighting. Some LED lights have a full wave rectifier on them so the now faster cycle rate is not visible to the eye but a camera may still catch them.

... but wait a second you say. What about those long exposure trail lights showing car taillights and headlights as a solid streak? Well, those show up solid because they are direct current and don't have a cycle or flicker rate, they are on constantly.

 

[Robin Usagani]

cool.. very good discussion. Here is another one

 

[DiskoJoe]

Good stuff here. i actually learned something useful!

 

[bratkinson]

I too, learned something here.

I never thought about slow shutter speeds and flashing LED lights at 60hz. I've had a few surprises with slow shutter speeds and lights in and out of sync. But LEDs? I've never run into them in my limited experience.

For what it's worth, I think the results are very interesting!

 

[Garbz]

The theory on the camera side is right but the reality on the electronics side is not. It's true 60Hz flicker exists, but in practice there are very few LEDs actually hooked up to AC in a way that would present this flicker (typically the most common example is an indicator light on a surge protector). The reason being is that nearly all consumer LEDs are driven by DC by using a rectifier and then filtering in their powersupply.

So why do Christmas lights still flicker? The answer lies in the controllers for light. We expect Christmas lights to fade in and out in wonderful patterns. This is often achieved by pulsing LEDs as you can then control their brightness by the length of the pulse. Those fancy red / green / blue LEDs generate that smooth colour range as well using this method.

What can also cause flickering when you wish to control many different LEDs but only have a few pins on your little circuit to control them we typically rely on persistence of vision to make them appear as being all on at the same time when in reality we are controlling them in batches. Pulse one set, move on to the next, pulse the next set, etc. This one is really easy to detect though. If you take a photo of a set of Christmas lights with a fast shutter speed and in the photo only a section of LEDs appears to be on at one time then this is the explanation. We do this for a variety of reasons but some really advanced designs outright need to pulse in this way. Charlieplexing for instance allows us to control a large number of LEDs individually with great efficiency from a design point of view (8 pins on a controller can control 56 individual LEDs), however the nature of the design means that we can only control one LED at a time. Persistence of vision and pulsing LEDs to the rescue again.


So it's not just slow shutter speeds that cause problems, but fast ones too. Also it's not just LEDs. Fluros also vary in colour depending on your shutterspeed (though this one really is because of your 60Hz mains frequency).