Build an Overexposure Detector for Video Cameras

[kolbep]

Hi.

I was wondering if anybody has an idea if this can be done.

I do part time videoing of Weddings / School Plays, etc.
I use multiple cameras, but only the cheap consumer ones, that do not have zebra stripes to allow me to see if there is overexposure.


What I was wondering is, if there is a way to extract the exposure levels from the Composite PAL output of the camera.
(I do not know much about Video Formats, but if there is a way to filter out everything except the levels, and then feed this into a PIC (or even Opamp if it is not too complicated), and then give an output as a percentage of the picture that is overexposed?
Or even maybe hijack the video signal after this detection, and mask areas that are overexposed, or that are exposed above 80%, and then combine this masking with the original signal, which is outputted to a composite LCD?

Any Tips

P

[free_electron]

nope. in the pal signal it is already clipped. you need the stream between camera sensor and video processor.

buy cameras with zebra markers.

[NiHaoMike]

You could just check if it's getting too close to saturation. You'll lose a slight amount of dynamic range but not enough to really affect video quality.

[tom66]

First, clamp the video at -0.3V using a diode-capacitor clamp or an active video clamp.

Use a sync separator like LM1881, LMH1981 to determine active video area (BPULSE.)

Set a comparator at something like 100 IRE (~715mV), gated by BPULSE, anything over that maybe latch an LED (d-type/jk-flipflop), use the ODD/EVEN pin to reset on each field.

Optional bonus points: average comparator output and feed into a bargraph driver LM3914/16 to get a "percentage" clip meter.

You might get some false detections at exactly 100% bright - so turn down camera exposure a couple percent to avoid this.

Not done video stuff for a while, studied it in detail a couple years back for another project; would this work?

[Someone]

nope. in the pal signal it is already clipped. you need the stream between camera sensor and video processor.

So all the professional monitoring solutions which apply zebras on the video stream are a waste of time? A waveform or histogram view might be more appropriate in precise situations but shooting with zebras is a good compromise when you only have attention or space for a single screen.

buy cameras with zebra markers.

Yes, probably the cheaper solution.

Google returns lots of options for the OP:
"lcd composite zebra exposure"
But then they need mounting and a power supply and cabling, which adds up if you are the sort of person concerned about the cost difference between a $300 camera and a $500 camera.

[vk6zgo]

A Waveform Monitor gives you all the information you need,but you need to look at it as well as the viewfinder---difficult unless you have two heads!

That said,with a little experience,you should be able to see the exposure level reasonably well,just looking at the picture on  your viewfinder,assuming it looks at the video output.

Broadcast cameras were fitted with automatic systems to do this,but they were usually too slow & were not used.
Instead,a Camera Control Unit operator watched the levels on a Waveform Monitor & adjusted as they went along.

A particular nightmare was when a football player ran through the line of shadow from the grandstand.

One moment--perfectly exposed,next---down in the blacks--adjust like mad,then he kicks the footy,& it goes to someone standing in bright sunlight---instant overexposure & white clipping.

Even HD Digital Broadcast cameras suffer from this problem----I suspect that they leave them on "auto".

[SL4P]

What tom66 said.

You should also filter for (Y) luminance only, or use Y in a component or SVHS output if you have them available.

If you leave the chroma on the signal, you'll be triggering on the colour modulation as well as the luminance (doing it all properly is a project on colourspace, gamma and gamuts in itself@!)