Blackmagic Forum

Hi all, I am new to this forum. Nice to meet you all! People (John Brawley?)have commented that the Voigtlander 17.5 0.95 is soft and bloomy when wide open. I own this lens myself and never find that it is the case on my old GH2. It is actually very usable and quite sharp at the centre on my GH2, and it is my go-to night lens. Last week I have got my new BMCC MFT so I decided to test it out. The result is shockingly different from my GH2. The BMCC MFT is a lot more bloomy and soft with the Voigtlander 17.5 wide open, than on the GH2. Almost make the lens unusable at wide open.

I have attached the test result. The left side is 2.5K RAW from BMCC + Voigtlander 17.5 wide open, 1:1 crop to the centre of the lens. Distance to subject is around 1m. The right side is photo RAW from GH2, crop and scale to match the BMCC frame. Position of the camera is the same, lens setting is the same, even shutter speed is the same too. Both opened in ACR and scaling is done in Photoshop. No post sharpening.

The BMCC picture (left) is a lot more bloomy and soft and lack of contrast, while on the GH2 you can clearly see the texture of the paper and even read the text. You may argue that the GH2 photo sensor is a lot more resolution and it preserves that resolution when it scales down. But if you look at the edge of the bottle and the contrast of the text, if I didn't do the test myself I would have thought it's a completely different lens!

If I stop down the lens to f1.4 the "bloominess" seems to be gone and match with how the lens looks on GH2. But at 0.95, looking at the edge of the bottle, it almost looks like there is another layer of reflection inside the camera. While I see the area around the BMCC sensor is kind of reflective and is not exactly all black (unlike GH2), I tried matting the rear of the lens and see if it affects the bloominess. What I find is it does, but by the time it kill that bloominess, the image appears darker. So it's not all that different from stoping down the aperture. So I can't help but to conclude it is a design flaw with the area around the sensor reflecting an extra layer of light. Maybe it's the sensor glass? Or maybe just the general non black reflective areas around the sensor?

Please chimp in your thoughts about this! What might be the problem and how might we fix this? I have the Voightlander 25mm 0.95 and it doesn't do that wide open. Have you guys experience something like that with other lens?

I love my voigt 25mm, but I think the bloominess (is that even a word?) is just a dynamic of the lens. Incredibly shallow DoF, and just plain soft. I find 1.4 and faster to be, for most applications, totally unusable. It's not a true .95 lens IMO. If there were some need for a hazy flashback, or a dream sequence, it could do the job, but not for the typical shot. It's sharp at f2, and razor sharp by f2.8 but the image quality gets soft and detail gets lost at its fastest aperture.

On the 25mm, yes it's not quite usable at 0.95 lol. But 17.5 is a different animal. It is quite a bit sharper wide open and doesn't have the same dream like quality - at least on my GH2.

it would be great to know what is going on
I also found strange when they said the voigtlander was soft as mine (25mm) looks great on my gh2.. even wide open!

Interesting. I'd love to know why there's a difference.

Considering the sensor is smaller on the BMCC any softness should have been more prevalent on the larger sensor GH2 so this need some explaining

Samjack wrote:Considering the sensor is smaller on the BMCC any softness should have been more prevalent on the larger sensor GH2 so this need some explaining

I'd think it's the other way round: exactly because the sensor is smaller any blur would become magnified on the BMCC.

But I am not magnifying anything. It's the same setting same lens same distance same shutter speed.

You're comparing a video to a photo. The video is 2.5k. The photo is of much higher resolution.

As I said, you don't need to pixel peep to see the difference. Look at the edge of the white bottle. It looks like there is an extra layer of reflection.

pacinoho wrote:But I am not magnifying anything. It's the same setting same lens same distance same shutter speed.

Yeah, but not the same sensor. The BMCC has a very different number of pixels on the same area than the GH2, so the image is sampled very differently. Either you had to downscale the GH2-image yourself, or the camera did it for you. Downscaling tends to sharpen the image, that _might_ explain the different results. And of course stills cameras also tend to sharpen the images.

Yes I scale down the GH2 photo to match the BMCC capture. But explain to me what kind of downscaling got rid of the extra blurry edge of the left side of the white bottle? And why does the different diminish when I stop down the lens? It is not a pure sharpness problem, which if it is, is relatively easy to solve. It's almost like the brighter part of the image gets away and blur out, while the dark part stays at where they belong, if you know what I mean.

Compressed vs uncompressed.
in 8bit you lose data, saturation, and other image qualities.

a RED asks more in the terms of IQ from a lens then a GH2 does, since it renders more details and better image quality, you will more likely see more flaws aswell on a RED with a lens then a GH2 which is compressed.

Case closed.

The GH2 is RAW photo, which is NOT compressed. Case is not closed. And how does that gonna do with the problem presented in my photo? Sometime I wonder if people actually look at the photo I posted...

If there is any structural sharpness issue regarding the sensor or resolution difference, then it should show in all lens in all settings. IE if I compare the same lens at F8 it should still be much softer and blurrier than on GH2. But that is not the case. If the BMCC is making everything looks like 80's porns, we won't be having this conversation here. Now the case is that BMCC is making things looks like 80's porns ONLY when I use the Voigtlander 17.5 at f0.95, while on GH2 is still reasonably sharp. That comparative difference only exist when the lens is wide open. This suggests something difference on the optical side. But since this is the same lens, I can only conclude that the difference is deal to the internal design of area around the sensor.

There is another thread on another forum, called "BMCC flares more than other cams" (http://www.bmcuser.com/showthread.php?4 ... other-cams) While you can all assume the issue is non existence and pull up explanations like down scale difference, compress vs uncompress (not even true in this test), difference colour science etc, people are actually seeing this for real.

Christian Schmeer wrote:Interesting. I'd love to know why there's a difference.

My guess is that the BMCC sensor has better angular response than the GH2. At f/1 the marginal ray angle is 30 degrees from normal in the center of the image, and many sensors have a very hard time dealing with such a steep angle of incidence. What happens is you get a sensor-induced apodization effect, and in practice, stopping down from, say f/2 to f/1 does not give you the expected 2-stop increase in exposure. This has the side effect of not really showing everything that the lens is doing, including aberrations at really large apertures.

One way to check whether this is true is to do an exposure experiment while stopping down the lens from f/4 to f/1 in one-stop increments. If I am right, then the BMCC will maintain a better correlation of exposure with f-stop than the GH2 will.

pacinoho wrote:But explain to me what kind of downscaling got rid of the extra blurry edge of the left side of the white bottle?

Sorry, i didn't mean to explain what happened here. I just pointed out that the images, allthough seemingly taken the same way, are not easily comparable. The fact still remains that you GH2 produces sharper results in this particular situation. And of course you are right to look for the reason.

Or maybe because the glass on top of the BMCC sensor is not exactly right on top, but probably like 1 cm on top (is it that thick or is it just away from the sensor?), given that angle you talked about, will there be internal reflection in that piece of glass that reflected back to the sensor?

Would be indeed interesting to see the same test done with a frame grab from a gh2 shot video, not a raw file.

I am confident the extra "bloom" is due to light bouncing around the relatively large funnel shaped baffle surrounding the sensor. It's not even lens flare, it's really sensor whitewash from light reflecting off the baffle walls onto the sensor. I have seen this "whitewash" with other lenses and it only happens on the BMCC.

On both the EF and MFT mount, the lenses used produce a much larger image circle than what would be appropriate for a sensor of that size. This overspill of light is not fully absorbed by the baffle.

Benton wrote:.

On both the EF and MFT mount, the lenses used produce a much larger image circle than what would be appropriate for a sensor of that size. This overspill of light is not fully absorbed by the baffle.

Except in this particular case the 17.5mm is a m43 native lens and the sensor size is close to the ideal.

JB.

Could it be, that the sensor reflects more light back only the rear lens element on the BMCC ?
And the lower tolerance for IR compared to other sensors.

brianc1959 wrote:

Christian Schmeer wrote:Interesting. I'd love to know why there's a difference.

My guess is that the BMCC sensor has better angular response than the GH2. At f/1 the marginal ray angle is 30 degrees from normal in the center of the image, and many sensors have a very hard time dealing with such a steep angle of incidence. What happens is you get a sensor-induced apodization effect, and in practice, stopping down from, say f/2 to f/1 does not give you the expected 2-stop increase in exposure. This has the side effect of not really showing everything that the lens is doing, including aberrations at really large apertures.

One way to check whether this is true is to do an exposure experiment while stopping down the lens from f/4 to f/1 in one-stop increments. If I am right, then the BMCC will maintain a better correlation of exposure with f-stop than the GH2 will.

not a bad theory.

It is apparent that the blur is most pronounced at the left edge of the white label, much more so than the right edge, in the BMC image, and the bottom is much more blurred than the top. I wonder where in the frame this was (in both cases)? If it was purely light-bounce, I would think any area of strong light would bloom or cause extra light effects, but this is definitely strong only on the one side.

Also, noting that the background is much more blurred in the GH2 version, these two are not so optically comparable.

Both are from the center of frame. The light comes from the left, so the left is blurrier. Yes, the optics looks totally different even though it's the same lens and same settings. And that's the point of this thread.

I have infinity focus problems with this lens. At f0.95 it's impossible to focus (and do it fine on my GH3)

Mine got infinity ok. The lens go pass infinity anyway. If yours doesn't but does ok on GH3 then your BMCC must be way off?
So is it possible that the slight difference in flange distance between BMCC and GH2/3 affects the "bloom" and softness, even when it is not focusing infinity?

I don't see how anyone can blame the Voigtlander for this one considering it looks great on every other m43 platform.

Still, would be indeed interesting to see the same test done with a frame grab from a gh2 shot video, not a raw stills file.

pacinoho, can you take a video of the same thing with the GH2 instead of a still?

i am the one wrote:pacinoho, can you take a video of the same thing with the GH2 instead of a still?

Yes, comparing apples to apples would be nice and more appropriate for this discussion.

Sure. Probably will have time to do it end of week.