OT: Backup camera glare

Yep, but I made a mistake and this is not going to be as simple as I originally suspected.

The old vidicon tube was b&w only as is today's b&w composite video. However, backup cameras are now color, which is not quite the same waveform. That's why I mentioned the color burst, which appears just after the sync pulse and is centered around 0.0v. If a simple comparator is used to trigger on something near +0.075v, it may also trigger on the color burst and mangle the waveform and picture. I'm not quite sure how to avoid that problem. Maybe a one-shot or timer that's triggered by the sync pulse, and times out after the color burst, when it's safe to switch the video to +1.0v (black).

Unfortunately, methinks that your 2017(?) Chevy Volt does not use composite video. It might use the GM-LVDS interface: So, if you want to add inputs, outputs, or additional circuitry to the interface, you need an interface box: There are two for the 2017 Chevy Volt: You won't like the prices.

The problem is the dynamic range of the camera isn't enough. Your solution is rather like gluing sunglasses to your eyeballs. It's going to decrease brightness in daylight and increase noise at night. ===========================================================

Exactly like sunglasses. I don't know that his camera doesn't have the dynamic range for this to work. I have a cheap backup camera and it seems to have a reasonable dynamic range, so there might be a happy medium where just enough of a neutral density filter cuts the glare at night but still leaves a useable picture day and night, and if so, that's the cheapest, easiest solution and only takes a few minutes of testing to evaluate.

How would this help? It would replace the white flare with black, still blotting out the image. Flare is an optical effect where the light bounces off the black surfaces of the camera or reflects off the lens surfaces enough to blot out the image. Once it is in the signal path there is nothing you can do to recover the image. A better camera is needed or at least a better lens or filter.

Yes, they're mixed signal chips. The outputs are some sort of standardized digital interface (MIPI, parallel BT.656, WiSPi, or some such). Some are even adding the video processing DSPs to the sensor chip.

Sounds like you'd basically have to use some dedicated IC to separate out the sync and luminance for processing and then composite them back. ICs like that aren't hard to come by but yeah, not so simple.

It got me thinking about "video compression" though, not in the sense of data compression but dynamic range compression, like an audio compressor. Like a feedback compressor with a filter in the control signal such that it has a different "knee" for low and high amplitude signal, but doesn't chop the high amplitude signals completely, then followed by an expander to restore the dynamic range of the low amplitude stuff.

There are a few articles online about stuff like that, apparently dynamic range compressing video while maintaining other bandwidth-related qualities like sharpness is a non-trivial problem and there are algorithms (probably patented) implemented in FPGAs and uPs to handle it.

I don't know if there are any commercial analog ICs that have the bandwidth to do audio-like compression of composite video signals.

Blech! Cheaper to just try to find a better camera that already incorporates some kind of dynamic range processing.

Use the mirror's on you car? Or just turn your head and look. :^) I'm guessing any 'fix' will screw up the daylight performance. (I use cheap back-up monitors and CCD camera's to look at physics stuff in the NIR, (~800nm) to get a good image I often have to screw around with the various light levels.)

George H.

Good ones have a pretty good dynamic range (new ones are pretty amazing). Cheap cameras are just that.

I think my trick of blocking any clamping until after the sync pulse and chroma burst might work. If not, you're right. One would need to seperate out the various video signal components, massage the luminance, and put everything back together again. Not fun.

You just described HDR (high dynamic range) photogrphy:

Just about every CCTV camera has built in AGC (automagic gain control), a form of compression, usually with adjustable threshold and gain. For example: (Yes, I know the chip is obsolete).

Yep. Buy a cheap 3rd party backup camera and monitor kit. The camera will probably use composite video, which should make the previously mentioned video clamping techniques work.

I'm trying to determine what's inside the GM or Volt camera box. LVDS (low voltage differential signaling or TIA/EIA-644) is just the interface specification.

More:

"LVDS Offers Robust Video Interface for Automotive Applications"

"MT9V032 LVDS camera board" Beyond raw sensitivity, the MT9V032 supports a form of high-dynamic-range exposures. By progressively reducing the sensitivity of its pixels over the course of an exposure, the sensor is able to approximate a non-linear response. This significantly increases the dynamic range that the sensor can capture in a single exposure - making it much easier to operate in environments with wildly varying lighting conditions (e.g. outside on a sunny day).

Disclaimer: I have a few fundamental incompetences. Unfortunately, video is one of these.

Crappy ones are composite video. The better/newer ones are HD and use some flavor of LVDS interface. There is a lot of work being done in this area these days.

Daylight-cut IR filter?

Might need to supplement with IR illumination...

Not quite like sunglasses. I can take sunglasses off when I walk in the house.

No matter how slight the filtering is, it will increase dark noise, this impeding night vision. I wouldn't want to be the owner of the car who modified the system and than ran someone over because (or not) of the modification.

Like vacuum tubes and bellbottoms analog video mostly falls under "before my time"

Cars aren't made to see out of anymore, it'd actually be nice to have a camera on the front too so I can actually see how far in front of me the curb is when parking.

Google the term "Gamma"

mark

It only clamped signals that shot over 1V to 0V*, so white was OK, whiter than white was clamped to black.

• or was that 0.3V?

Mikek

Why would it matter what level it was clamped to? If it's whiter than white, I would think clamping it to white would be fine.

I'm sure it depends on the system and where in the system you measure, But from '73' to '94' all the video outputs I looked at were 1Vpp, the sync was 0.3V and the video information was 0.7V. I will give you that Black sat at 0V, white at 0.7V and sync was minus 0.3V. Like this. >

But, I can back up what you say with this.

That's not my experience though.

Mikek

Ya, maybe with today's sets, but I seem to recall our B/W monitors would also flare on bright objects. All I can say is it turned headlight images and parking lot lighting black and made the rest of the picture viewable. Back then this was a big deal to security guards setting in their shack watching the factory. Even had windshield wipers on the camera housings, and a fan for hot days. Mikek

I wasn't sure of the voltage levels so I looked it up with a Google image search before posting. I don't recall which image I used, but here's a few that agree with my

-0.4v sync voltage levels:

and here are some more that agree with your -0.3v sync voltage:

or, perhaps you might like a -0.5v sync pulse?

When in doubt, there's Wikipedia: which shows something quite different from any of the above contenders. So much for Wikipedia.

Well, maybe EIA RS-170 (for b&w video): See Detail YY on Page 13 of 15. -0.4v sync pulse.

It this is where I stole my numbers: White: +1.000 V Black: +0.075 V Blank: (0V reference) Sync: -0.400 V

My guess(tm) is that we might be looking at the difference between composite video as delivered over a CCTV coax cable, versus composite video as delivered by a video broadcast transmitter, where the negative voltage sync pulse cannot go below zero. Or, everyong doing video is derranged.