Where's the sweet spot?

[Frank Granovski]

You got 10X or 12X of optical zoom, and whether you're at .75X to 10X there's no degradation of the footage, right? Hmmm. It's supposed to be like that we would think but it's most likely not the case. Most zoom lenses have a sweet spot or 2, or within a certain range. Never mind the "Leica" stamped on the barrel. Even Leica makes some not so perfect lenses. So where's the sweet spot with your cam? At 10X, mine's not so good; anywhere from 1X to 4X looks pretty good. (DVL9500.) I should play around a little more with my MX300, but I know it's got a bit of pin-cushion, that's why I like to sit down to be comfy while I shoot. What do you think?

http://www.dvinfo.net/conf/showthrea...ght=pincushion
http://www.dvinfo.net/conf/showthrea...ght=pincushion

[Jeff Donald]

I would think think the pin cushion would interfere with you sitting down. Let's face it, the lens is made to Panasonic specs. and is not a 'real' Leica lens. If it were it would cost 1/2 the value of the camera and would leave me wondering how many corners were cut on the camera to get it to market at it's price. The Leica stamp is only a marketing ploy to keep up with the licensing agreement Zeiss and Sony have. Sort of a keeping up with the Jones'. In fact most of the Leica lenses I've encountered on video cameras show pronounced barrel distortion. I wouldn't rank them as good as the Sony Zeiss lenses or even the Canon lenses.

[Bruce Simpson]

I agree, I'm quite disappointed with the amount of barrel distortion produced by the MX500's optics.

[George Beck]

Bruce, can you post a pic as an example of your claim?
I would like to compare it with the GS100 or GS200 if possible.

Jeff, I'm sure the reason of using fake Lecia and only stamping the logo is to fool all those professional that buy those consumer cams, and know that there's any difference between one lens and another.
But.. c'est la vie.

I liked the wild-life still pics of yours tho. What did you use to take them? (not the barrel distorted pana-cam i bet) =)

Frank, let me know how (if) you can find your sweet spot. since the lens of the GS400 will be whole new story we'll have to go fishing for the sweet spot again ;-)

here's something related I found while surfing on the net. Do you think this may help? It's Excell file and recalculates depending on the setup values...
(I have the USAF chart too, in case you don't have it)



Resolving power (lines per mm) as a function of film resolution, aperture and defocus (V0.1)
Frank Loeffel, Version 0.1, 17may02, 28may02
film resolution (l/mm) 147 This is for Velvia. Velvia resolves 80 l/mm at contrast 1.6, 160 l/mm at contrast 1000. I logarithmically interpolated the l/mm at contrast 30 because my USAF resolving power chart shows this contrast.
lens base resolution (l/mm) (for model 1) 80 A reasonable default value, for the center of the lens' image circle.
diffraction constant 1500 Used to characterize diffraction as a function of fstop.
lens aberration base (l/mm) (for model 3) 50 Used in a heuristic formula for lens aberrations.
lens aberration factor (for model 3) 3.5 Used in a heuristic formula for lens aberrations.

stop, f/ Model 1: with lens base resolution (lines/mm):
defocus (mm) 0.0001 0.33 0.5 0.66 1 1.33 2
5.6 68 30 21 16 11 8 6
8 66 39 29 23 16 12 8
11 62 46 36 29 21 16 11
16 56 49 42 37 28 22 15
22 49 46 43 39 33 27 20
32 39 38 37 36 33 30 25
45 30 30 30 29 29 28 25
64 22 22 22 22 22 22 21

stop, f/ Model 2: without lens base resolution (lines/mm):
defocus (mm) 0.0001 0.33 0.5 0.66 1 1.33 2
5.6 129 33 22 17 11 8 6
8 116 45 31 24 16 12 8
11 100 55 40 32 21 16 11
16 79 61 50 41 30 23 16
22 62 56 51 45 36 29 21
32 45 44 42 41 37 33 26
45 33 32 32 32 31 29 26
64 23 23 23 23 23 23 22

stop, f/ Model 3: with lens aberration base + lens aberration factor x fstop (lines/mm):
defocus (mm) 0.0001 0.33 0.5 0.66 1 1.33 2
5.6 61 30 21 16 11 8 6
8 65 39 29 23 16 12 8
11 66 47 37 30 21 16 11
16 63 53 45 39 29 22 16
22 56 51 47 43 35 28 20
32 43 42 41 39 36 32 26
45 32 32 32 31 30 29 26
64 23 23 23 23 23 22 22

Comments:
For actual use, Model 3 is best. Model 1 and Model 2 were used to develop Model 3.
The formulas used are models of the real world and only approximate the behavior of a photographic system.
These models are valid for the center of the lens' image circle. With additional work, models could be devised that give results dependent on image height.
Remember, resolving power is dependent on contrast.
Resolving powers of system components are assumed to accumulate according to 1/Rresult**2 = 1/R1**2 + 1/R2**2 + ...
Resolving power based on diffraction is asumed to be 1500/fstop.
A heuristic formula to characterize lens aberrations other than diffraction was used in Model 3: lens_aberration_base + fstop * lens_aberration_factor.
Resolving powers are not well suited for comparisons between analog and digital photographic systems. This is because film's resolving power is very higly dependent on the contrast whereas digital sensors' are not.
"Defocus" means the focus displacement of the ground glass to bring any chosen point in the scene into focus, relative to the focus position chosen to take the photograph.
If you have measured the resolving power of your lens, you can tweak the models' parameters to make the results match your measurements.
The defocus of 0.0001 mm is given instead of zero to avoid division by zero complications. The results are the same.
When lines / mm is referred, I mean black lines, so my numbers are the same as line pairs / mm.

[Allan Rejoso]

Interesting topic.

I've also noticed that in terms of video quality consistency over the entire zoom range, my Optura100 (with its awesome f1.6 -f1.8 (is there any other consumer cam with an f rating this tight???) zoom lens) outperforms the Black Mamba.

[Jeff Donald]

All those shots are with a Canon 10D and Canon L series lenses. I've own my share of Lecia's over the years (M3, M5, R3) and the better Japanese glass (Nikon ED, Canon L, some Pentax, some Olympus, some Fujinon) produced in the past 5 to 10 years has really caught up to the Leica, Schneider and Zeiss glass.

[George Beck]

I was going to ask for the cam, but looks like you answered above =) canon 10D

[Jeff Donald]

The charts demonstrate that the sweet spot is about two stops down from maximum aperture. After that the lens start to become diffraction limited. The only reason to shot the smaller apertures (f16, f22 etc) is the need for DOF, which video doesn't normally need.

[Tommy Haupfear]

What did you use to take them?

If I recall Jeff uses the Canon EOS 10D with a Canon EF 300mm f/4L IS USM lens. My 10D would really appreciate some of that quality Canon L-series glass.

[Jeff Donald]

The Canon 10D.

[George Beck]

Got it =)
sorry, i'm a little slow at 2 am.. time to hit the pillows
g'nite

[Jeff Donald]

Your correct Tommy. I've shot the 300mm non IS side by side with the IS version and the non IS is noticeably sharper in the samples I tested. Go figure.

[Tommy Haupfear]

Jeff, how odd. Did it make a difference whether the IS lens was on or off and were your shots low light (sans flash)?

I've only got the one Canon 28-135mm IS USM lens (my starter lens) but I appreciate the image stabilization since I rarely shoot on tripod/monopod.

[George Beck]

oh.. in the model above you can change the values for:

film resolution (l/mm)
lens base resolution (l/mm) (for model 1)
diffraction constant
lens aberration base (l/mm) (for model 3)
lens aberration factor (for model 3)

should I change those and how so to get the GS400 properties (matrix-model)?


while you are on it, can someone translate in english what that means?
35mm Film - 45~450 (tape mode) / 37.6~451mm (card mode)

[Jeff Donald]

I always shoot on a tripod (at least the nature stuff) and it didn't make a difference with the IS on or off.