PD: Take a look to the link posted by Rodolphe, I found it very interesting.

My french is a little rusty, but it illustrates very well what I was able to see using a flashlight - 4 blade diamond shaped iris with 2 NDs at various positions in the f4 range. I attempted to photograph it this weekend but there are so many reflections from the flashlight in the various lens elements that it's difficult to get a shot which shows the NDs - their method is certainly much more effective.

It also explains something I'd noticed occasionally in a few shots, especially one long focal length shot of a bird with water behind it - the bokeh has a noticeable geometric texture to it when there are a lot of out-of-focus highlights.

My french is a little rusty, but it illustrates very well what I was able to see using a flashlight - 4 blade diamond shaped iris with 2 NDs at various positions in the f4 range. I attempted to photograph it this weekend but there are so many reflections from the flashlight in the various lens elements that it's difficult to get a shot which shows the NDs - their method is certainly much more effective.

It also explains something I'd noticed occasionally in a few shots, especially one long focal length shot of a bird with water behind it - the bokeh has a noticeable geometric texture to it when there are a lot of out-of-focus highlights.

Sony's US patent no. 6,533,473 describes a diamond-shaped iris with two ND filters. Figures 3 and 6 in the patent illustrate the progression as it stops down. Is this what you were seeing?

http://www.google.com/patents?vid=USPAT6533473&id=U1UNAAAAEBAJ&dq=6533473

Good Work, Mike.

Are you a patent attorney : )

No, but for both the Sony camcorders and the Canon HV10 (thread titled 'HV10 - ND Gradation feature'), the patent descriptions provide more detail than is available elsewhere.

Sony and Canon have pursued different physical implementations of the iris / ND filter assembly. But the design rationale is similar -- apparently the iris can't be stopped down beyond f8 without causing diffraction problems on the sensor chip. So steps of ND filtration are added to provide the effective aperture control range that otherwise would be obtained simply by stopping down the iris to f16, without using internal ND filters.

Excellent find, Mike. I hadn't seen that document; it provides very useful information. It's an oldie, but a goodie!

Using the Google link below, you can search for terms in the text of patents -- something that isn't possible on the US Patent Office site.

Search terms such as "ND filter," "ND filter Sony" and "ND filter Canon" turn up many interesting results, both old and new, which give insight into the manufacturers' thinking, and quite a few specifics of the physical mechanisms and control schemes.

http://www.google.com/patents

Evan - I think you're seeing a two bladed iris inside the Sony - the top blade is an upside-down V and the bottom blade is a rightway-up V. This is much simpler and cheaper to produce than the nice 6 bladed iris as used on the VX, FX etc, but it does the job, albeit with less attractive highlights and flare when used into the light. Out of focus highlights are noticeably less attractive too.

The internal ND filtration of the PDX10, the HC1 etc is designed to stop the cameras working at any stop smaller than f/4, regardless of the focal length. Of course if it gets brighter than even the internal ND can soak, then smaller apertures are used - but this is seen as an emergency getUhome, as diffraction with 1"/3 chips becomes quite noticeable at f/5.6 and smaller - and especially so at the wide-angle end of the zoom.

Cameras with this undocumented internal ND often 'make up something the punter will believe' when 'display' is hit on tape replay. Panasonic are also guilty of this, displaying silly f/22 readouts on many of their cameras.

tom.

Hello again after some years,

I am using a HVR-A1 and would like to know, if anybody knows how the three (?) internal ND filkters work togehter with external ones. Has anybody used them in combination?

In addition, am I correct with the following assumption regarding the exposure steps of the A1?

f4: 0db/None ND
f4: 0db/ND2
f4: 0db/ND4
f4: 0db/ND8

Thanks,
Peter

I'm not sure if the A1 has separate ND filters - I would expect it to have a varying density filter so that the more it moves into the light path the more light is absorbed. I don't know though _ I haven't looked down into its lens.

So what happens is this. You're filming at f/2 (say) and it gets brighter. The iris closes to f/4. It gets brighter still and now the ND absorbs the light. If it gets brighter than a theoretical f/11 (with max ND in place) the iris will start to close again, or in some cases the shutter speed will shorten.

It's far better to do this yourself (film at 1/100th rather than the default 1/50th) if it's very bright and avoid the use of external ND filters altogether. They cause flare spots with such short focal lengths and reduce the hood's efficiency. Not good.

tom.

Thanks a lot Thom for you explanation and advice!

Another question is what will happen when the varying density filter work together with an IR filter or circular Pol filter I am also considering to use.

You plan to use an IR filter? Maybe you mean a UV? Whatever - the internal ND will make no difference at all - it's neutral as the name suggests. If you attach a polarisor that absorbs 1.5 stops, the internal ND will simply move to allow 1.5 more stops of light to the chips.

Before you buy the circular polarisor it's worth trying a linear polarisor. Much cheaper and will probably work just as well.