Jump to content

Dynamic Range Test Chart?


Max Field

Recommended Posts

Does anyone have a link to a printable dynamic range test chart?

 

Been trying to find one but they're just photographed results from other cameras. And can any printer handle it or would I need to purchase something special?

 

Thanks for any input.

Link to comment
Share on other sites

  • Premium Member

Not really doable. Printers will struggle to create something with enough range to meaningfully test modern digital cinema cameras, and linearity issues make it very difficult to achieve any sort of accuracy.

 

Bear in mind, if you want a 15-stop chart, and in reality you probably want a bit more than a 15-stop chart, the density range is 215 or over 32,000:1 (which is why it is difficult to make high dynamic range cameras.)

 

At that point it is very, very difficult to create a blacker black or a whiter white, to the point where charts like that tend to be backlit and use "cavity black" where the darkest chip is actually a cutout box lined with black velvet (or you could find some vantablack paint, it does exist). The best approach is probably to get some ND gel, hope it's reasonably accurate (you can measure it, of course), and start layering it up on a lightbox. Done carefully with the right values you can avoid having to use more than four layers, and avoid having the error in the gel add up too much.

 

You will have noticed that DSC Labs' Xyla backlit chart is extremely expensive. Note the chips reduce in size as they get brighter to avoid the inevitable small veiling that all lenses have from making the darkest chips invisible in the glow from the brightest ones. There's also a sliding window that allows a small range to be visualised all at once.

 

The actual physics involved in dynamic range in the modern world of thirteen or fourteen stop cameras is becoming quite extreme, which makes this difficult.

Link to comment
Share on other sites

  • Premium Member

The only real way to do this would be to have say, 16 rectangular evenly illuminated LED modules, with individually adjusted power levels so that each one was twice as bright as say the one to the left of it, and half as bright as the one to the right .
In practice that would mean that in theory, if the brightest LED was drawing 1 Amp, the dimmest one would be drawing about 15 microamps, but it could never produce an accurate output at that tiny a current.
I did do some preliminary experiments with the above some years back, but the results tended to tell camera owners what they didn't want to hear, so I promptly lost my access to high-end cameras.
With the much better and cheaper white LED modules available these days, I may revisit this subject.

Link to comment
Share on other sites

  • Premium Member

Doing that with DC-driven LEDs would be a nightmare of linearity problems; to calibrate the extremely low output patches, probably a photomultiplier-based instrument would be required.

 

I think the best way to do it would be to use suitably high frequency PWM. That's much easier, although you'd still have to be careful about the behaviour of the brighter lights changing as they warmed up. At a sufficiently high PWM rate, interactions with the camera timing could be acceptably minimised and very high dynamic range (especially with modern high power LED arrays!) could be achieved with good linearity. It would be possible to implement this as a small, handheld device with sequential display of a high dynamic range test series in a single test patch. In fact, it could even be clamped directly to the lens mount to illuminate the sensor directly, excluding external sources of light and removing the lens as a source of inaccuracy. The desirability of doing that would depend on whether the goal was to test the sensor or the entire system dynamic range.

 

Properly-chosen emitters or filters could also test dynamic range per colour channel.

 

It's not even very hard to do; a weekend project, perhaps?

 

I can only reiterate your dismay at the public reaction to unpalatable truths.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...