Do You Realistically Use f1.2 through f1.8?

[Peter Moretti]

I'm debating between getting a set of used primes (a few $100 a piece) or going with a new Nikon 35-70 f2.8 zoom (~$470).

My main ? is how often do you realistically use f stops below 2.8? Am I giving up too much speed for the ability to change focal length on the fly?

I will be able to control lighting to a large degree, so my only need for a very wide iris would be to limit DOF.

So after first falling in love with the shallow focus depth, are you still using stops wider than 2.8?

Thanks much for all your help!

-Peter

[Chris Barcellos]

Its just anecdotal ( I have Nikkors 100mm f2, 50mm F1.4, 35mm F2.0, and a Canon 24mm f2.8), but I think the real benefit of the lower F stop lenses is that they have a reduced probability of vignetting, due to the larger aperature of the rear elements that tend to be found in them. With lenses above, the biggest offender regarding vignetting is the 24mm, which indeed has a smaller rear lens element.

As far as F Stop selection on the lenses, the 1.4 for instance, actually looks to soft in a face shot because of the shallow depth of field. For some effects, that might be just fine, but in others, it will be a problem, and stopping down will work better.

[Bob Hart]

Some lenses, probably very many, go soft when used wide-open, so the logic behind selecting a f1.2 lens which went soft at f1.8, then using it in the range f2.8 - f5.6 would be to preserve more resolution at f2.8 aperture settings than could be achieved with a f2.8 lens which went soft at f3.5.

Once the initial novelty of shallow depth-of-field wears off and the real world kicks in, a shot of a human face which shows in immaculate detail the pores on the nose is useless if the eyes are out of focus and the ears totally diffuse blobs, unless of course the pores and nostril hairs are the subject of dramatic import.

Then there is the hotspot/vignette issue already raised by Chris above.

I would favour used f1.8 or wider primes over a new f2.8 zoom.

[Igor Babic]

There is also an issue of minimum distance from subject. Primes are focusable at 35cm. Zooms are mostly over 1,5m. Ther is a lot of situatios where you need to be realy close to a subject.

[Jon Wolding]

Lens sharpness really isn't much of an issue when you're not shooting with a full frame sensor.

I have a F1.2 50mm that I use wide open all the time. Though I can now stop down to F5.6 with the new Brevis imaging element (and still avoid grain and vignetting), I typically keep it at F2.8 to match the exposure with my other lenses. But if I'm shooting w/o much light, I'll use the F1.2 wide open.

I also have a F1.8 50mm... but I only use it for rail-less shooting because the focus ring is so narrow (too narrow for a FF).

[Bob Hart]

QUOTATION - "Lens sharpness really isn't much of an issue when you're not shooting with a full frame sensor."

Sorry John. I can't let this one pass without clarifying something. Maybe I am wrong.

My imagining of the groundglass based relay imaging process is that the system becomes a defacto full frame sensor otherwise I have been wasting about three years of my spare time hacking up bits of plastic and shopping for parts in hunting down this particular grail.

Groundglass based relay imaging does throw away a little resolution, in my case about 150 TV lines at claimed HDV resolution.

In a vaguely similar way that tube based nightvision also amplifies any softness in a lens, a softly projected sharp point in an image falling on the groundglass will be rendered furthur softened by the relay process, especially so if it is a bright highlight which should be seen as a pinpoint.

Projected image softness may not be directly observable because the relay resolution has limits. One could assume that lens sharpness which just comes up to that resolution limit will be fine.

However, you will observe a loss of contrast and apparent clarity. Your projected image needs to be as good as it can be onto the groundglass.

But, creatively, a lower contrast image can be a pleasing aesthetic in its own right.

It is is a matter of personal choice. My preference is for as much contrast as I can preserve. I can always take it away later.

[Peter Moretti]

One potential problem w/ the Zoom I'm looking at is it's a push/pull design, so the barrel telescopes in and out. That could cause a problem for balancing the tripod head and for positioning a matte box. But it's considered to be one kiss a__ zoom.

[Jon Wolding]

(pardon the scaling and compression... the forum wouldn't post the full res. images)

The first pic was shot with an F1.2 50mm wide open with a 100W clamp light.

The second pic was shot with an F2.8 135mm stopped to F4 with a 650W, two 300W's, and a 400W.

I honestly don't think you can get much sharper than this when you're projecting light onto a ground glass. The only extra detail you'd pick up is grain.

[Ben Winter]

I agree with Jon that the sharpness isn't affected when just using an adapter and not a higher resolution medium, however the ultra-shallow DOF of a wide-open shot makes focus pulling very difficult and the general look of a shot can be very soft because so little is in focus. I always make a habit to shoot under 2.8 and my shots have begun to look much sharper.

[Peter Moretti]

Originally Posted by Bob Hart

..
My imagining of the groundglass based relay imaging process is that the system becomes a defacto full frame sensor...

Bob, this is actually a very interesting observation, IMHO.

It seems that when using an adapter, camcorder sensor size becomes irrelevant and it's only mega pixles that matters. So 1/3rd and 2/3rd chip equal mega pixles cameras each fitted with identical adapters should yield the same results (ignoring other differences between the cameras).

[Bob Hart]

QUOTATION

"So 1/3rd and 2/3rd chip equal mega pixles cameras each fitted with identical adapters should yield the same results (ignoring other differences between the cameras)."

Provided the adaptor is custom matched to the camcorder's imaging area (what P+S Technik does)your description is a good fit.

[Jad Meouchy]

Originally Posted by Bob Hart

"So 1/3rd and 2/3rd chip equal mega pixles cameras each fitted with identical adapters should yield the same results (ignoring other differences between the cameras)."

Provided the adaptor is custom matched to the camcorder's imaging area (what P+S Technik does)your description is a good fit.

What about dynamic range? I have to disagree here. The pixel imagers on the 2/3 will be physically larger and thus will be more sensitive.

[Bob Hart]

Jad.

I get your point.

If you relay a fixed size groundglass image to 1/3" and 2/3" sensors by two different relay lenses designed to frame to scale, the fixed groundglass area identically, then you might expect the larger image relayed onto the 2/3" CCD to be dimmer than the image relayed onto the 1/3" sensor.

Extra low light performance of the 2/3" sensor might be offset partially by the dimmer image.

It is all a bit academic. If a DP is serious about production values, then attention will be paid to learning the limitations of the system then getting the lighting right, regardless of the sensitivity of the total system.