GG and Bokeh
 

So recently there have been a lot of posts talking about acheiving good bokeh (nature of out of focus point sources of light) and which ground glass is best suited for the purpose in 35mm adapters.
I'm of the opinion the bokeh is a property of the slr lens used. When using fast lenses as is the necessity with these adapters, out of focus point sources of light can generate quite a large image on the gg. Surely, if the diffusive properties of the ground glass could in anyway distort this image, then it would useless for its primary purpose which is to relay the 35mm image as accurately as possible enabling the video camera to capture it.


So, the question is: has anyone actually tested varieties of ground glass and compared bokeh in like for like circumstances and if so what was the result?

Hi David,

I think its mainly down the the level of diffussion across different types of GG. I have been testing this alot and it has been posted many times that the more the GG diffuses, the better the bokeh, but there is more lightloss.

Testing a 1000 grit GG against a 400 grit GG, for example. The 400 grit would render the sharp disks from the out of focus points of light, which would be the same shape as the aperture when the lens is stopped down (pentagon shape). With the 1000 grit glass, this would not happen, the bokeh would be much more hazy.

So yes, the Bokeh is down to the 35mm lens, but also in the GG's ability to capture it.

So, which one gave better bokeh? 1000 grit GG? When you say hazey do you mean smoother?

(o/t:

Wayne Kinney
Inner Circle

Join Date: Jun 2005
Location: Brighton, East Sussex, UK
Posts: 666 <--)

I had better reply to make it 667 ;)

The 400 grit glass gave better bokeh, but more lightloss. Hazy due to the ghosting mainly (aerial image).

Hi Wayne,


But if the gg disffuses the unwanted sharp halos (the dreaded bad bokeh)
would it also not diffuse any other detail in the 35mm image as well?


cheers

Dave

Love the new look website by the way!

Well, the points of light turning into disks when out of focus is what we want (good bokeh).

When they have a distinct edge or halo that is bad right.
So if the gg softens away this effect then surely any detail in the rest of the image that we want to preserves will have to be diffused away.

I'm just not convinced that the bokeh can be positively influenced by a good gg.

This is good bokeh: http://www.the-digital-picture.com/I...keh-Sample.jpg

This is bad bokeh:
http://www.frozenphoenixproductions..../LetusTest.bmp

The G35 has a focusing screen that produces results very similar to example 1. Good GG means good bokeh.

You see bokeh like example 1 in movies all the time. If you think that is bad bokeh, you are in disagreement with the majority of the industry.

Ben,
The first image says what I mean, thanks. If shooting straight to film, this is what we should get.

In an adapter, the lower grit GG renders this better then say a 1000 grit glass.

Well, i guess we are comparing good ground glass and bad ground glass, and the bokeh produced between the 2 will be different.

I agree the first image looks great.
But what I don't get is this:
If we were shooting straight to film that is what we'd get right so the bokeh is a product of the slr lens.The gg should be such that it diffuses this image and what we get on video is identical to what would have been on the film cell.
If the gg modifies this then surely it is distorting what would have been on the film cell thus degrading the image we want to capture.

Beyond merely softening the image (which isn't really a good thing) and hence the edges of the out of focus lights I can't see what the gg can acheive.

So what you're saying is if the out-of-focus areas don't look right, the in-focus areas might not either? It's possible, but the issue is usually that the gg isn't diffusing the image enough, not distorting it. As a result it's usually that more things are in-focus than they should be. I'd personally like to believe that a focusing screen that diffuses properly lends itself to a better image, but there are lots of people on these boards who cry themselves to sleep over light loss.

That's it. Beyond merely softening the image (which isn't really a good thing) and hence the edges of the out of focus lights I can't see what the gg can acheive.

here's what I mean: on this site about 1/2 way down are 3 images of bokeh as a product of the lens
http://www.kenrockwell.com/tech/bokeh.htm
so if the gg can turn fig 1 to fig 3 then it must be doing something weird to the light all over the cell, which isn't quite what we're after.

David,

Imagine this: We take an adapter and imagine using a clear peice of glass as the GG. Your shooting complete aerial image. DOF is very wide. Now, we begin to grind the glass, ground a little, we are diffussing very little, and still seeing the 'aerial' image. grind more, diffuse more. Thats 1 factor. Then the grit size will also influence diffussion.

This is just an example of how the GG can influence the Bokeh.

The 3 images in your example is good, I have had a GG which gives bokeh like the third image, more like a gaussian blur.

I think that example is dealing with good and bad bokeh between lenses, our job is to match the bokeh of a given lens to what it would produce if attached to a film camera.

I actually find what they call "bad" bokeh pleasant. I like how the G35 has switchable focusing screens that give you "hi-lux" and "lo-lux" type bokeh, which is basically figure two and figure three of those examples. At the bottom he talks about bad bokeh examples, and how one picture had bad bokeh yet the photographer knew what he was getting. So while they may classify bokeh as technically "good" or "bad", it's still artistically subjective.

I was actually thinking the same myself. I guess thats subjective.

I guess to conclude, the GG will effect the resulting bokeh, and we need to match it the best we can to what it is on a film camera, for any given lens.

i agree with David:
http://www.dvinfo.net/conf/showpost....03&postcount=9

I spent a small fortune on a 50mm f1.2 Nikkor. I took some shots on film and it produces bokeh like fig 1 - bad!

Another thing with GG and diffusion - a less diffusing gg can produce good bokeh and solid discs with a longer focal length and/or smaller aperture. If you look on Dan's site he has an example of solid discs taken with a 200mm f3.5 on a beattie.

The tricky thing is to get a gg that can produce solid discs on a 50mm at f1.2 or f1.4.

The tricky thing is to get exactly that image that drew your photolens. If gg diffuses too little then the image will be mixed with non-processed(by lens) rays. Which will be part of the image on camcorders ccd but it's not part of the image on the gg.

That's physically impossible due to the nature of what is essentially a rear projection screen. With full diffusion no light would pass through the gg, so you have to compromise if you want an image that is going to be bright enough to film.

So it seems that we have a good understanding of the tradeoff between transmittance and diffusion. But back to one of Dave's main points:

The "in focus" parts of the image that are being projected onto the gg is also being diffused. Even though it does not have the circle of confusion that the out-of-focus spots have, they are unavoidably diffused when they hit the gg. This is probably why the "in-focus" part of the image never looks as sharp when using an adapter vs. when using your camera alone. We have to componsate by using sharpening filters etc.

I think gg technology is going to advance quickly in the near future. But then again, affordable full frame 1080 24p 4:2:2 detachable lens cameras are problably not that far off either. What will I do then? Maybe go out and shoot something rather than reading this forum and working on my adapter to no end.

id agree the bokeh is much nicer on the second one! in photography people preffer smooth, circular booky - not pentagons.

Andy, i agree with You 99%. But i'm actually, maybe too indirectly, trying to say that direct unprocessed rays are cause of image artifacts which sometimes rise up in forums as topics of bokeh. Those have no connection to bokeh really. Those just smear image what is drawn by lens. Besides those rays are illusory on gg. So whole bokeh thing from point of gg is always conversation about diffusion.

About retained 1%. It's that i'm just not too sure physical impossibilities :)

Tim, the shape of the out of focus lights reflects the form of the 35mm lens diaphragm. When the iris is wide open, you get circles; as you stop down, the blades of the diaphragm form a polygon, which is reflected in the bokeh.

You also get polygons (the number of sides depends on the number of blades in the diaphragm) in photography and, depending on the lens used (and the bokeh it produces), they may or may not be distracting.

i know im just saying, the edges are very defined - i wouldnt consider that good bokeh personally.

Interesting thread...and after a lot of experimentation, there's little question in my mind that diffussion/bokeh/sharpness are inherently related. I don't believe we'll ever see true film emulation with respect to diffusion because we are not dealing with a film emulsion layer and it's subtle inter-layer reflection properties, as Andy alluded. That's why belabouring the point of perfect film diffusion is simply beating a dead horse (apologies to horse lovers there). Getting film-like diffusion at higher f-stops is fairly easy, but a match at F1.2 is likely impossible...we can only approximate it. At that level of diffusion, I don't believe the focal plane can be kept as sharp. Hence our swappable diffuser setup. Focused film images are not soft, nor are they all that grainy at 100 ASA. Try a test, however, at 400 ASA and you'll see that the bokeh is extremely grainy. The look would likely be dismissed outright by the adapter community.

From what I can gather, very high diffusion screens have grain issues at fairly low f stop values which limit their usefullness. Unless someone can design an adaptive GG, you simply cannot have a GG that does it all. You can, however, chose the GG for the look you want, or vary your focal length/distance/f stop values to achieve shallower DOF. What you cannot do is take a highly diffuse adapter and make it sharper or grainless at higher f stops...you're stuck with its properties.

Bokeh is an entirely subjective topic. Most agree though that soft geometric shapes, absent of internal distractions constitute "good" bokeh. I find growing halation on specular highlights unacceptable..however many adapters show this.

What's good for SD is of course a whole other issue when HD enters the fray.

I'm not sure this is quite correct. If you imagine a very very thin gg surface and consider a very very small element which scatters incident photons in all directions with uniform angular intensity, then you would have 100% scattering and 1/2 of the incident light would be scattered one side of the gg and the other towards the slr lens (with a very little in the plane of the gg). As such, the image recorded by the camcorder would be exactly that as captured by a film cell. Ie there woud be no aerial image (unscattered rays) which is what we are after.

The point was though, if you constructed an adapter such that the camcorder recorded off the front surface (as would be perfectly possible but perhaps a little silly) then one could just use any white surface (something that scatters incident rays in all directions). If this was the norm, would we be discussing the properties of the surface used in conjunction with bokeh. I think not.

i had a reply but it was somewhat foolish - but i dont think the idea of recording off a front surface is silly at all

I did test a variety of GGs. Here are some results showing the difference between a less diffusive GG and a more diffusive GG in the exact same lighting conditions. The diffusers used in this case are engineered diffusers which have a 10° difference in angle FWHM.

Camcorder Panasonic PV-GS400
Vibrating adapter
Canon 50mm F1.4

Less diffusive GG, camcorder at F4

More diffusive GG, camcorder at F3.4

By the way, both these setups are unusable because of grain showing in good lighting.

It was discussed at length earlier on the forums. It was concluded that the technical aspects involved were unrealistic.

I also did some focusing screen tests, this time with varying types of diffusion gels for theatrical lighting. I will post the clip of the highest diffusor soon. You can see the bokeh in the video, and it is bokeh I find very desirable.

if you were just taling about lenses for a minute then a lense that gives you good bokeh (say a 58mm f1.4 minolta rokkor ) will generally give a less sharp image on the other hand a sharper lens has slightly poorer bokeh , however the difference is slight. This might have something to do with the fact that to get a shallow DOF and nice bokeh out of a lense you generally have to have the appeture wide open , when a lenses focal sweet spot as it were is somewhere around f5.6 . Again when taling just about a lense ( before complicating things with diffussion) there is an artistic choice between sharpness and nicer bokeh. I agree that the best focusing screen would be able to best emulate what happens on the film plane in an slr camera ( I say best because I also agree that it would be near impossible to exactly replicate what happens on film)
rob

I've compared my adapter footage to film shots on SLR with the same lens (50mm f1.2 Nikkor). On 100ASA and 400ASA film (surprisingly to me) the dynamic range is absolute crap by comparison, bright areas are blown out, dark areas are too dark in the same shot, bokeh is grainy and highlights have a bright edge classed as bad bokeh.

However on the adapter footage the dynamic range is good, and because the gg doesn't fully diffuse it actually softens the bokeh somewhat and the result is it looks better than actual film haha!

hey andy
sorry I havent replied yet to the last email you sent me lifes a little hectic at the moment , how did you go with wax , if you are using wax ,I noticed that depending on exposure I still get slight hotspotting and occasionally the grain is evident but mostly not, haven got any decent footage to show as everything ive got has been deinterlaced and colour and gamma corrected and would probably not accurately demonstrate the ability of the adapter
rob

I'm oscillating 40 micron gg at the mo, wouldn't go back to wax unless you tell me bees wax can give light loss around 1-1.5 stops total loss for the system.

I always correct my footage anyway so I'd still like to see "doctored" grabs.

This was shot on the Letus35 with a theatrical lighting gel diffuser switched out with the focusing screen. The footage is great, but there was too much grain even with the vibration. I'm sorry I can't find my sorenson serial number so I was stuck with the trial, hence the watermark. But you still get the idea.

http://www.frozenphoenixproductions....ffusertest.mov

Ben, the link does not work for me (Page not found).

Sadly ever since Powweb went under new management they moved everybody to a new system and with it, new rules. You'll notice you can no longer browse my exposed directories, they're blocked. I think another dumb rule they put in is no direct file linking--which is going to be a serious problem for me since I link to files all the time. I'll see if I can fix the problem.

EDIT: problem solved. I was uploading to the old server. Link should work when the upload finishes.

I should also mention that the R8 leica focusing screen is a fantastic gg but has a dumb split-prism in the middle. What a bummer.

The Nikon D100 replacement screen I ordered is exactly what I hoped for. Small, but a fantastic image, very similar in construction to the Beattie Intenscreen for the Contax. It doesn't really diffuse completely to my satisfaction, but the bokeh is different than the Beattie from what I can tell, hopefully better. There's practically nil grain. I'm installing it in my Letus and I'll have test footage shortly.

Ben,
Sounds great, be sure to show us some footage!