The end of the Neutral-Density Filter

By Jean-Pierre Beauviala, consultant member of the AFC

La Lettre AFC n°262

[English] [français]

After a three-year wait, in San Francisco on 3 February 2016, Panasonic announced the upcoming release of image sensors with complete separation of the function of photon-to-electron conversion performed by an organic OPF* layer, and the function of accumulation and drawing-away of the electrons performed by the silicon substrate.


This announcement, following the announcement of Quantum InVisage**, leads us to expect the arrival of “global shutter” digital cameras that will accept high-angle rays of light (less peripheral illumination and no false colours in the corners), with a dynamic range increased by three apertures (which is a considerable amount), and most of all, with adjustable sensitivity. We will finally be able to choose our ISO like we used to be able to do in the time of silver-process film stock ; it will be even better because we won’t have to change film stock : a simple electronic command – manual or automatic – will do the job. There will no longer be any risk of causing the silicon reservoirs to overload under strong lighting.

Little reminder
In full sunlight, a photographer controls the sensor by playing on the aperture and the exposure time. Because closing the aperture isn’t always a good thing (the depth of field might change in an undesired way, and diffraction beyond f/11 harms the sharpness of the image), the shutter is there and can climb to up to 1/3000th of a second.
A film-maker, on the other hand, cannot play with the shutter, which produces photograms that are so sharp that they no longer allow the brain to reconstitute the fluidity of movements. In order for each photogram to bear the necessary flow, the shutter timing at 24 images-per-second must be 20 milliseconds.

Stuck as he is between these 20 milliseconds, the f/4 +/-2, and the intangible sensitivity of silicon sensors***, he is left with no choice but to insert a neutral-density filter (ND) into the optical path.
Installed in front of the lens of a reflex camera, the ND will darken the image on the target screen to the point that the viewer can see almost nothing ; interposed behind on a variable neutral-density filter support or on a sheet of electro-variable-transparency glass, the ND filter harms the FTM, harms the balance between visible and infrared rays, and harms the image because of the dust that gets stuck on it.

In 1985, for an ethnologist who only wanted to bring two film magazines along with him to Niger, only 250 ISO and no grey filters, we had set up a 5-blade shutter on his XTR that was hidden under the reflex mirror. (See patent images below)
Once he had left his hiding place, he would shadow the shutter opening, which, during the 20 required milliseconds only received 5 expositions of 1 millisecond melted into one single 5ms shot ; the 250 ISO film looked like it had been only 64 ISO.

Thirty years later, the Penelope Delta with reflex visor and 800 ISO digital sensor suffered from the same problem, but was even darker. The same solution was found, and the only change was to increase the number of shutter blades to obtain a better fusion between fugitive images.
On the OPF sensors, it ought to be possible to simulate this principle by activating the photosensitive layer by using short activations equally separated in time over the entire 20ms length of the shot. Only the photos arriving during the activation will be converted, the others will be lost (for example, by applying twenty 0,1ms impulsions, the conversation only takes place for 2ms, and the sensitivity will be reduced tenfold). We still don’t know what method Panasonic has adopted : have they diminished the OPF layer’s polarization tension (thereby risking loss of stability and S/N ratio), or have they decided to apply a time division like the XTR ?

We will soon know. On that day, the sad ND filters will immediately disappear and there will be a clear division of responsibility : the aperture will be responsible for the depth of field, and the real ISOs will be responsible for the correct exposition time.

* Organic Photosensor Film invented by Fuji (carbon + arsenide or selenide from rare earth metals).
** InVisage is a Californian-Taiwanese startup that is working on the same idea (www.invisage.com/technology)
*** The so-called ISO adjustment of CMOS sensors plays on the gain of the preamplifiers, it doesn’t prevent silicon saturation in any way.