Member Discussion

True Innovation? Sharp Electronics Bold New Security Camera

Color / Infrared. Pick one, right?

Sharp electronics now says you can have both. Recently they announced the LZ0P420A 720P, SDI bullet camera, with Infrared illumination and real color image. Price: around $1350. From the lavish marketing brochure comes this:

and also this real-world cartoon example to show the difference between Sharp's camera and a conventional security camera. (Hint: Sharp is on the left).

For any of you damn intellectuals out there, here's the theory:

First of all, just what is color? When visible rays of light strike an object, the light reflects at the wavelength of each particular color, and our vision recognizes this as the color of the object. In a setting of 0 (zero) lux, in other words an environment of absolutely no light, it is necessary to produce light rays that reflect the color of an object without being bright. Such light rays are called near infrared light rays.

But because near infrared light has a completely different composition than visible light, these must be reconstructed as colors close to those that we can see. In other words, in the theory of capturing color at night, an object is irradiated with invisible near infrared rays—not the visible light we normally see. Light reflects back off the object and the colors of this light’s various wavelengths are analysed and subsequently reproduced as the colors we can see.

Our solution was to try developing a single-CCD chip system that would read the three primary colors simultaneously. In order to bring the colors of near infrared rays close to the actual visible colors, we developed a special pixel structure for the CCD. While enabling a single CCD to read all three primary colors (RGB), this also solved the practical problem of creating a sensor that would contribute to more compact cameras.

What do you think, Great Innovation or Great Imagination?

One thing curious is that, AFAIK, none of the security trade rags have reported on this model yet. Maybe that's because Sharp, being new to security, isn't familiar with the marketing model in this industry.

This is what we discussed here.

I did not know it was shipping, but if we can figure out how to get one, we'd do a test.

The device won't be available until November 28, and Sharp says monthly production sits at 4,000 units. Nanolux Japan will be serving as the device's distributor.

...but possibly all going to Japanese government.

Sharp is new to security, right? Because this PR statement

Further, this camera is also compatible with standard HD-SDI output, making it an easy replacement for existing camera systems.

would lead one to think so. Either that or they hired Todd on as VP of marketing.

Is this division of Sharp under the Matsushita umbrella too?

No, I don't think so. I always thought that Sharp and Panasonic (née Matsushita) were rivals.

Oops I was thinking of Sanyo. My B.

In depth article about Sharp's IR color technology. And more importantly the first non-cartoon pictures that I could find from this camera:

Left: no infra-red. Right: infra-red magic. Curiously absent : Full light shot, so we can see just how purple that wall is (or not).

Do they say what the lux level was for that scene? Do they mention what the other camera was in the 'bad' left side images?

I am generally distrustful of vague vendor low light comparisons, like this piece of gamesmanship:

The report leads you to believe that they are the same camera, just with IR (Right) and without IR. The pictures are not so much a demonstration of the overall low-light performance of the camera, but instead show how adding IR to a dark scene renders colors visible, even though there is little visible light present in the scene. They don't indicate lux.

All other cameras require at least some visible light to render colors that are somewhat 'true'. This is the essence of the innovation and they are still just apparently enamored with showing that off! because it doesn't seem like there is anyone selling them yet...

But it matters how good or bad that camera is before adding IR. Maybe that camera simply sucks in any low light without IR.

I don't know one way or the other but without knowing the lux level of the scene, you cannot reasonably make a claim whether it is impressive or not.

Maybe that camera simply sucks in any low light without IR.

Well maybe it does simply suck in low light without IR. Who said it didn't?

John, the breakthru in this camera is only when the IR is on, not off!

As I have said already, the set of pictures on the left are NOT intended to show how good or bad it's non-IR LL performance is. They are only demonstrating the ability of the camera to go from not having enough light to make an image, to the ability to have a color image, by just adding IR. Name another camera that can go from too dark to color by adding IR.

Don't agree? How about a shoot-out? The winner is whoever can produce a color image in 0 lux visible illumination. What's your camera? Lightfinder? Lose. Image would be dark. Standard IR camera? Lose. Image is b+w.

Before you object and say "these pictures are not taken at 0 lux.", understand that they are just illustrating what the effect is of IR on a scene that is already too dark. The IR illumination is clearly the reason for the the picture getting better on the right and would therefore make a 0 lux image at least somewhat visible, and in color, by using the IR illumination.

Anyway, choose your camera!

Honestly, those images suck.

They're not really "in color", they are more like "grainy with some hints of color".

The color accuracy looks very off (as you also noted), so you couldn't really say "the suspect was wearing a red jacket" with high certainly. In fact you might actually get LESS usable data compared to more traditional low-light cameras. Higher detail in B/W is frequently better than low detail with highly inaccurate colors.

And there is still the point about shutter speed/motion blur. None of those images show any activity in the scene.

The issues you mention are valid, and that is what will rightly make or break the camera, not its non-IR LL performance.

"The winner is whoever can produce a color image in 0 lux visible illumination."

We don't know if the Sharp camera is producing a color image in 0 lux. In this example, the light level is clearly over 0, probably over 1 lux, which is easily shown in scene B.

We can't make attempt to make any meaningful conclusions without knowing the lux level of the scenes in those 'test' images.

I agree with you that with 1-2 lux of illumination falling on the image on the left, the cameras low-light performance is sub-standard, certainly for today's cameras.

I mean, I can't make out anything at all in the extreme right side of image (a), can you?

After IR is turned on I can read the whitish letters on the greenish storage box.

Since we agree the non-IR performance of the camera is poor, what other explanation is there except that the letters are illuminated by the added IR? Do you think the 1-2 ambient lux is somehow illuminating the scene all of a sudden?

So yes, IMHO, it can be reasonably concluded that the camera can illuminate a scene with IR and interpret that IR to render a color picture. Whether it's 0 lux or 1 lux ambient is irrelevant, since the lux from the IR is orders of magnitude greater.

You can object, as I did in when introducing the images, that we don't know if the colors are faithful to real-life, this the primary difficulty addressed by the images.

Do you think Sharp is actually lying and switching the poorly performing LL camera for a Axis Lightfinder and not turning on IR?

If that scene is 3 lux, than many 'conventional' high end IP cameras can deliver a similar or better color image.

If that scene is .3 lux, than it might be a significant improvement over conventional cameras.

Since we do not know the lux level of the scene, I cannot reasonably take a position one way or the other.

Well said, Thanks!

Yes, I agree with your logic that we can't say if a Lightfinder would have rendered a better picture, in the scene shown, without knowing the lux. Totally Agree.

My point is different, these are my premises

  1. The pictures intend to show that the Sharp camera is at least somewhat sensitive to IR illumination; and, indeed new detail is visible when IR is turned on, while retaining a colorish image. (IR Off: Left, IR On:Right)
  2. If it is sensitive to IR to the degree shown, then it will create at least some color image, even at 0 lux, since it
  3. Lightfinder will have no image at 0 lux
  4. Therefore, At 0 lux Sharp should have a better picture

If you wouldn't mind edifying me one last time by indicating the step number which you find unlikely to be true, I'll won't argue it.

I am not taking a further position on this, given the lack of information. You may be right. I have no idea. If and when it ships in production and is being marketed for use in surveillance, we will test it out.

Those test shots are suspiciously devoid of any motion or other activity.

I bet this is more Intensifier 2.0.

At 500% more expensive than Speco, I hope for Sharp's sake it is a lot better...

Since we don't have any idea when the Sharp camera might be available, here's an even more impressive video of an even more unavailable camera: Full color at 1/4 moon.