[dlacouture]

Hello all!

Well, I've been using old lenses for quite a time now, on the MZ-7, MZ-6, K10, K20, K7, Kx, and now K5...

I've stumbled on the infamous stop-down metering problem plaguing the K10/20, and still visible on the K7/5...

I've tried several split screens, but I quickly found that my keepers rate was the same with the stock screen, so I ditched them, as they are useless anyway when you recompose the shot and severely mess the metering... And their blank parts were limited to f/2, in the best cases...

I then took a leap of faith and ordered the Pentax ME-60 blank screen, hoping that it was optimized for fast lenses, but, alas, it has the same f/2 limit than the stock screens... But to my surprise, I found the utterly blank viewfinder really pleasing and it helped me much in composition, as there is no disturbing element (as the focus brackets)...

Then, after a little research, I've found out that the Canon ee-S screens were optimized for fast lenses, and were supposedly able to accurately show DoF right down to f/1.7... So I immediately bought one and took out my saw... I already made a DIY split from a MZ-M screen, so I was not adverse to a little plastic dust...

I ended with a nearly mint focus screen (got a little scratch on it, visible only at about f/8 ).

OK, time for the tests...

DoF rendering:
I took out my f/1.4 lenses (Pentax FA50 and Samyang 85) and played a little with the DoF Preview, and hallelujah, I can now see a difference between f/1.4 and f/2 (try it with the stock screen)!!! Even f/1.7 produce a slight darkening of the viewfinder, so the eeS screen actually work right down to f/1.4!

Focusing:
With now an accurate DoF, seeing what is actually in focus at f/1.4 became really easy! But there was a severe back-focus, so I had to remove the metal shim and replace it by two thin strips of post-it...
Now, focus is eerily accurate, with both lenses! Even a split screen could not give me such a precision...
I'll order the proper metallic shim now that I have the exact thickness needed...

Metering:
OK, here I was worried...
So, I put the FA50 out of its A position, and took a series going from f/1.4 to f/22 in 1 stop increments... Here is what I got (repeated several times): 3000, 2000, 1000, 500, 250, 125, 60, 30, 15...
Yep, you've read it right : a near-perfect linearity on the whole range, with only a slight +0.5 overexposure at f/1.4...

With the 85mm now... 800, 500, 320, 160, 80, 40, 20, 10, 5... Really good, too, with a little +0.5 overexposure at f/1.4 and f/2...

I simply never had a screen so accurate before!!! For reference, the K5 stock screen has nearly 1Ev of leeway around the proper value...

[glee46]

That's great. Not sure I would even know where to begin to cut it, & shim it correctly. Do you have any links, or do you plan on maybe a small tutorial? Thanks for the info.

[DogLover]

Hello all!

Well, I've been using old lenses for quite a time now, on the MZ-7, MZ-6, K10, K20, K7, Kx, and now K5...

I've stumbled on the infamous stop-down metering problem plaguing the K10/20, and still visible on the K7/5...

I've tried several split screens, but I quickly found that my keepers rate was the same with the stock screen, so I ditched them, as they are useless anyway when you recompose the shot and severely mess the metering... And their blank parts were limited to f/2, in the best cases...

I then took a leap of faith and ordered the Pentax ME-60 blank screen, hoping that it was optimized for fast lenses, but, alas, it has the same f/2 limit than the stock screens... But to my surprise, I found the utterly blank viewfinder really pleasing and it helped me much in composition, as there is no disturbing element (as the focus brackets)...

Then, after a little research, I've found out that the Canon ee-S screens were optimized for fast lenses, and were supposedly able to accurately show DoF right down to f/1.7... So I immediately bought one and took out my saw... I already made a DIY split from a MZ-M screen, so I was not adverse to a little plastic dust...

I ended with a nearly mint focus screen (got a little scratch on it, visible only at about f/8 ).

OK, time for the tests...

DoF rendering:
I took out my f/1.4 lenses (Pentax FA50 and Samyang 85) and played a little with the DoF Preview, and hallelujah, I can now see a difference between f/1.4 and f/2 (try it with the stock screen)!!! Even f/1.7 produce a slight darkening of the viewfinder, so the eeS screen actually work right down to f/1.4!

Focusing:
With now an accurate DoF, seeing what is actually in focus at f/1.4 became really easy! But there was a severe back-focus, so I had to remove the metal shim and replace it by two thin strips of post-it...
Now, focus is eerily accurate, with both lenses! Even a split screen could not give me such a precision...
I'll order the proper metallic shim now that I have the exact thickness needed...

Metering:
OK, here I was worried...
So, I put the FA50 out of its A position, and took a series going from f/1.4 to f/22 in 1 stop increments... Here is what I got (repeated several times): 3000, 2000, 1000, 500, 250, 125, 60, 30, 15...
Yep, you've read it right : a near-perfect linearity on the whole range, with only a slight +0.5 overexposure at f/1.4...

With the 85mm now... 800, 500, 320, 160, 80, 40, 20, 10, 5... Really good, too, with a little +0.5 overexposure at f/1.4 and f/2...

I simply never had a screen so accurate before!!! For reference, the K5 stock screen has nearly 1Ev of leeway around the proper value...

I think I see a great little side business for you of buying and modifying these screens for Pentax use. If you can work out stricter methods so as to avoid scratching them, you're all set. I'd definitely be interested.

[stevebrot]

Yep, you've read it right : a near-perfect linearity on the whole range, with only a slight +0.5 overexposure at f/1.4...

Which mode...Av or M? The K-5 uses different metering logic in M as opposed to Av mode. If the Canon screen works equally well in both modes, this indeed is good news for M42 lenses users.


Steve

[macTak]

I too would love to know with what exactly (you say a saw) you used to cut down the screen (I plan on doing this myself). Can you compare how the ee-S performs to the MZ-M screen--I already have a MZ-M screen I was planning to cut down, but now you may be tempting me to try the ee-S. Also, can you say how the ee-S screen performs with slower lenses or lenses stopped down to the middle apertures?

[dlacouture]

I used M mode, with the aperture ring set to actual apertures instead of A. So the K5 stopped down to meter. In Av mode, the metering is done wide open anyway, so there is no problem here.

This is indeed the ultimate m42 screen, as it is nearly perfectly linear in response to the aperture. The K10/K20 screens (and a lot of split screens around) has a non-linear response to stopped-down metering : -1Ev underexposure at f/2, +2Ev overexposure around f/8... The K5/7 has a slightly better screen, but it's still -0.5/+1...

But this eeS screen, man, it's a real metering pleasure...

I used a small metal saw (I'll try to post some pics).
I've made a split screen from a MZ-m, and it's not better than the stock screen, focus and metering-wise... Not bad (better than the K10/20 screen), but those FF split screens are really too big on APS-C... And metering was still suffering from the over/underexposure at varying apertures...

[macTak]

This is indeed the ultimate m42 screen, as it is nearly perfectly linear in response to the aperture.

But, as I recall, the downside with this screen is loss of brightness, so I would worry about the brightness of this screen with one of my Takumars at say f8. Can you give some assessment of the screen brightness in the slower (~f4-f11) range.

[rhodopsin]

The Ee-S in my 5D gets much darker than standard screen past f/5.6 -- no problem focusing or dof preview given enough light, can be a problem otherwise. Ee-S screen is essential for accurate focus with lenses faster than f/2.8.

The problem of which screen to use arises when I use Tamron SP 90/2.5 -- if I intend using it wide open only or need absolute critical focus, I'll choose Ee-S, otherwise I'll use stock screen because DOF preview at f/8-f/16 is much brighter.

[Rupert]

I like to see and read of you adventurous guys that try different things with you camera gear. It is interesting and informative. For me, I just leave it alone and use it how it is. I once worked on a lens because I had the right little screwdrivers and such. That was several years ago and I still find those teeny-tiny ball bearings on the floor every so often. I think they "exploded" out to my surprise.......the lens still works.....manual only and is a little loose, but it was my last time to become a technician on my camera gear.
Regards!

[DogLover]

The Ee-S in my 5D gets much darker than standard screen past f/5.6 -- no problem focusing or dof preview given enough light, can be a problem otherwise. Ee-S screen is essential for accurate focus with lenses faster than f/2.8.

The problem of which screen to use arises when I use Tamron SP 90/2.5 -- if I intend using it wide open only or need absolute critical focus, I'll choose Ee-S, otherwise I'll use stock screen because DOF preview at f/8-f/16 is much brighter.

Yeah, I read that on the Adorama site. Seems to be the Achilles heel of these screens. I don't often shoot past f/5.6, but I do sometimes, and sounds like that would be nigh impossible. Still tempting but I'm definitely not as gung-ho as I was at first. I'm thinking now that maybe I would only do this if I could devote that body to MF lenses only.

[dlacouture]

Honestly, it does not strike me as being so dark at f/5.6-8 as what we can read around...

Thinking about it, it seems logical for me that for apertures smaller than f/5.6, it will actually be brighter than the stock screen...

My logic? Having charted the stock screen's stop-down behavior, I know that below f/5.6 it tends to severely overexpose (as much as +2Ev around f/11), indicating that the metering sensor sees less light than what is really passing through the lens.

Whereas with the EE-S screen, the metering cell keeps doing its job in a linear fashion, so the amount of light it sees is really in accord with what goes through the lens.

The stock screens seem to be optimized to increase apparent brightness with lens between f/1.4 and f/5.6 (which makes sense, as it's the most common aperture range). This, in turn, tricks the metering sensor, resulting in an underexposure (right down to -1Ev around f/2).
That's why fast lenses underexpose compared to slower ones.

And the downside of this optimization is that with slower apertures, these screens darken faster than the amount of light being made available.

And this leads me to think that the Canon EE-S screen will be indeed darker than the stock screen, but only at apertures faster than f/5.6! Below that, you'll be maybe better with it than with a standard screen.
The only unknown in this is the initial offset given by the EE-S screen, against the one given by standard screens. It seems to me that the EE-S produces a slight overexposure with my usual lenses, so it seems to be slightly darker at wide apertures than the stock screens.
If the EE-S is 2Ev darker on average, then by f/11 both screens will be about the same brightness.
But if it's only 1Ev darker, the EE-S will actually be brighter in the f/8-f/16 range...

I'll be doing comparative tests next week, I should be able to get hold of a K7 with a LL60 screen.

[DogLover]

And this leads me to think that the Canon EE-S screen will be indeed darker than the stock screen, but only at apertures faster than f/5.6! Below that, and you'll be better with it than with a standard screen.

Don't you mean slower (numerically higher) than 5.6? That was my understanding.

[dlacouture]

Nope, I did mean faster...
If you take the factory screen and chart its behavior in stop-down mode, you'll see that it underexposes from f/1.4 to f/4, then overexposes from f/5.6 to f/22 (roughly, it can depend on the lens used)...

This means that the metering sensor sees to much light between f/1.4-4, and not enough under f/5.6, thus proving that the Pentax screens darken faster once below f/5.6 than a linear screen such as the EE-S...

This points to an optimization with lenses faster than f/5.6, which is consistent with the very name of the Pentax screens (labeled "Bright matte" IIRC).

EDIT:
I've just thought about an easy test to check exposure against a reliable reference, and used LV... Using a defocused target (to lessen the differences due to the higher accuracy of LV against small highlights), I get identical values between LV and the EE-S screen on the whole aperture range (M mode, Aperture ring manually set)!

So, for me, it's really the ultimate screen :
- accurate DoF
- linear response to stop-down metering
- accurate metering compared to LV
- blank!

[DogLover]

Nope, I did mean faster...
If you take the factory screen and chart its behavior in stop-down mode, you'll see that it underexposes from f/1.4 to f/4, then overexposes from f/5.6 to f/22 (roughly, it can depend on the lens used)...

This means that the metering sensor sees to much light between f/1.4-4, and not enough under f/5.6, thus proving that the Pentax screens darken faster once below f/5.6 than a linear screen such as the EE-S...

This points to an optimization with lenses faster than f/5.6, which is consistent with the very name of the Pentax screens (labeled "Bright matte" IIRC).

EDIT:
I've just thought about an easy test to check exposure against a reliable reference, and used LV... Using a defocused target (to lessen the differences due to the higher accuracy of LV against small highlights), I get identical values between LV and the EE-S screen on the whole aperture range (M mode, Aperture ring manually set)!

So, for me, it's really the ultimate screen :
- accurate DoF
- linear response to stop-down metering
- accurate metering compared to LV
- blank!

Huh! I guess only in trying it will I be able to tell if it would work for me. I think the only other thing giving me pause would be the fact that I have come to rely on a split screen/microprism collar to focus and I'm having trouble getting my head around being able to do that with "only" a matte screen. Again, I guess I won't know 'til I try. But if extremely shallow depths of field can be detected, then it shouldn't be a problem, right? I just hope that it wouldn't be for me.

Also, if there are no features on the screen, then trimming it down becomes that much easier, right? I mean, there's no split screen, focus brackets or anything to center. I'm thinking that a Dremel rotary tool might be better for trimming it down. Do you or does anyone know if there's a reason that would not be a good idea? I've got a loose stock screen that I could use as a template. Also, those screens seem to be kinda hard to come by right now, at least at the usual sources here in the states.

Thanks for the info!

[dlacouture]

Also, if there are no features on the screen, then trimming it down becomes that much easier, right? I mean, there's no split screen, focus brackets or anything to center. I'm thinking that a Dremel rotary tool might be better for trimming it down. Do you or does anyone know if there's a reason that would not be a good idea

Well, there is a definite center on the EE-S, you can see concentric "circles" on the fresnel face. But simply laying your screen on the EE-S and centering it is good enough, I think...

As for the Dremel, yep, should be faster, but you'll have to securely fasten it in place so you won't rip and scratch it...

Frankly, I think a good saw with very fine teeth and a protective screen (maybe a Nintendo DS protective screen, those working with static forces) are sufficient for the job...

I went the hurried route and did not protect the screen, and it ended with only a really small scratch (say, a dot) invisible at working apertures (and I'm not even sure it's not some saw dust gunk)...