HMC40 - First Look

[JONJON]

very exciting!!!!

[Barry_Green]

Well... I'll be pimp-slapped by a monkey...

Okay, so here's what the iris actually does: from fully open to f/2.8, it performs exactly as I expected -- each tiny motion of the wheel equates to a very tiny closing of the iris, in 1/6-stop increments.

Up until f/2.8.

Then, when you get to f/3.0, the iris itself stops changing. It stays at an actual f/2.8, and instead, a little ND filter comes gliding in from below and covers up the very bottommost section of the open iris window. And each tick you move the iris wheel, causes the filter to slide more into position, until it's all the way in, which is at f/6.4. So at f/6.4 the physical iris is still open to an f/2.8, but the neutral density filter is fully in place, so the actual light transmitted is the equivalent of f/6.4. Then, if you stop the "iris" down any further, it commences actually closing again.

So -- score one for Professor Bob and Cranky.

Not 100% sure how I feel about this. I mean, I know exactly why they've done it, they've done it for the sake of video quality because diffraction is a problem enough on a 1/3" camera, packing 1920x1080 pixels on a 1/4" camera can cause serious resolution differences due to diffraction. So they're attempting to keep the video as clean and sharp as possible for as long as possible throughout the iris range. Using the ND filter instead of actually closing the iris ... well, I guess it does give the user what they want, while preserving the highest video quality... it's just that it doesn't feel right that closing the iris doesn't necessarily actually close the iris, y'know?

I guess we just have to think about the display readout as T-stops instead of physical f-stops...

More disclosure would have made this perfect -- as Bob said, f/2.8 + ND.01, then f/2.8+ND.02, etc... but then again that might make things confusing. Technically it's doing what you want, it's controlling exposure the way you'd expect, it's just that the physical iris isn't closing. And with 1/4" chips, you're going to have deep DOF no matter what you do, and - frankly, - most of us spend our time trying to get the iris as OPEN as possible, and here the camera does it by default...

I guess I just chafe at any auto feature where I don't have 100% control over what the camera is doing, but at this price point, with the limited space for buttons, and the very serious image quality impairment that happens due to diffraction, it's probably a perfectly reasonable compromise. Of course, I'd prefer an HMC150-style three-position physical ND filter wheel that I can swap in at any time, but at this price point that just doesn't happen.

What this means from a practical aspect is that we should be diffraction-free from (display readout of) iris:OPEN to iris:f/6.4. Any deeper than f/6.4 and you're probably giving up some res to diffraction. Which may mean keeping a 3-stop ND filter on you would be a good idea if you're going to be shooting outdoors.

[Cranky]

So, Panasonic decided to keep a wider aperture at f/2.8 instead of f/4.0 or f/5.6 like Sony and Canon are doing. AFAIK, values chosen by Sony and Canon provide the least distortion, while the value chosen by Panasonic veers to the side of shallower DOF, given that the chips are only 1/4-inch. Works for me. BTW, f/2.8 is aperture that is accessible both at full zoom and full wide. Any larger aperture will not be accessible at full zoom, so Panasonic chose the widest aperture that is available at all focal lengths. It wanted to provide as wide aperture as possible. Geez, why did not Pana just use larger chips?

On the HMC40 the ND filter absorbs about 2-1/3 of stops, less than 4 stops on the Canons, but this seems to be enough, the lens is not that fast.

> it's just that it doesn't feel right that closing the iris
> doesn't necessarily actually close the iris, y'know?

You are not closing the iris, you are darkening the image

> I guess we just have to think about the display
> readout as T-stops instead of physical f-stops...

Bingo! Panasonic should have put exactly that in the operation manual.

> Of course, I'd prefer an HMC150-style three-position physical ND filter
> wheel that I can swap in at any time.

The old-school ND filter with fixed position is worse, because when you engage it you darken/brighten image for how many? 4 stops? ND16 ==> 2^4. Then you have to immediately adjust aperture to correct brightness. With auto ND you get smooth change of brightness in the whole exposure range. For small chips auto ND is a much better approach.

Now the bonus question: does the consumer TM300/SD300/HS300 have the same ND filter inside?

[Barry_Green]

The HMC150 has three two-stop filters.

As for what the consumer cams have, I have no idea, but I would be surprised if they engineered it differently as they'd have to face the same problem -- diffraction.

[Cranky]

As for what the consumer cams have, I have no idea, but I would be surprised if they engineered it differently as they'd have to face the same problem -- diffraction.

They might just not care. The TM300 closes down to F16. Those who want wider iris can use screw-in ND filters like in the olden times.

[shphotovideo]

Barry, thank you for taking the time to post all that information. I just got my HMC40 and really look forward to learning all that it can do. Will be using it this weekend for the first time at a car meet and drive.

[BobDiaz]

Barry,

First of all thank you for all the work in testing the camera.

Second, given the higher pixel count and the smaller image sensors, diffraction comes into play a lot sooner than on the HMC-150. So, I'm happy to see the ND filter kick into place BEFORE the diffraction appears. As long as we know what is going on, I'm happy.

It would be interesting to see what happens to the resolution at F1.8, F2.8, F8, F11, and F16. Or at least when the camera says F8, F11, and F16.


(2.8 x 2.8) / (6.4 x 6.4) = 7.84 / 40.96 = about 1/5.22 or about 19.1%

So the ND drops the light by a bit more than 2 F-Stops; roughly 2.38 F-Stops

Now, if F6.4 = F2.8 with maximum ND filter
F9.1 is really F4 with maximum ND
F12.7 is really F5.6 with maximum ND
F18 is really F8 with maximum ND (F16 is the maximum, right?)


Bob Diaz

[Barry_Green]

I'll shoot a 12233 chart at various apertures, so we'll know what the diffraction issues are.

[studio1972]

What this means from a practical aspect is that we should be diffraction-free from (display readout of) iris:OPEN to iris:f/6.4. Any deeper than f/6.4 and you're probably giving up some res to diffraction. Which may mean keeping a 3-stop ND filter on you would be a good idea if you're going to be shooting outdoors.

Thanks Barry, loads of great info there. I actually quite like the sound of this as it makes the controls simple, and the ability to have greater depth of field is not something I ever use on the HMC151 anyway. This actually keeps the DOF relatively shallow which is nice.

Just one thing though, if the ND filter is partially applied (lets say 80%), don't you effectively have a small aperture 20% of f/2.8 with no ND filter?

In case what I am asking is confusing. What I mean is that wouldn't you get diffusion at the edge of the ND filter? It would be like having 2 apertures at the same time, one small, one large.

[Superfresa66]

Just one thing though, if the ND filter is partially applied (lets say 80%), don't you effectively have a small aperture 20% of f/2.8 with no ND filter?

for my understanding there are multiple ND filters with different values, you're not actually covering only a percentage of the area with only one. It keeps applying the next higher value!

So at no moment you would have a filter covering only part of the lens/sensor. At least that's how i understand it!