4/3rds and movement

[Vivek]

Would a tilt lens be going against the telecentric "requirements" of the 4/3rds (and /or the M4/3rds)?

 

[Robert Campbell]

The simple answer would seem to be yes.

However, if an image is to be focussed on the sensor, presumably some of the rays can't be telecentric - if this means all rays are parallel and at right angles to the sensor.

I'm not too sure just how absolute the telecentricity is, and if all lens makers adhere strictly to the guidelines.

So, it might be possible.

 

[Vivek]

This whole argument advanced by the 4/3rds on telecentricity, in practical terms, seems bogus to me.

All those who are using movements with (whatever) lenses to capture their images should not be doing that at all with a digital sensor if 4/3rds and telecentricity are to be believed.

 

[jonoslack]

This whole argument advanced by the 4/3rds on telecentricity, in practical terms, seems bogus to me.

HI Vivek - I don't think it's bogus, and in practical terms I think the quality of the Zuiko 4:3 lenses speak for themselves - even the cheap ones can be shot wide open at all focal lengths without concern . . .not something that you can say about any other manufacturer's lenses.

Worth noting that although Sigma and others make lenses for 4:3, I don't think they conform to the telecentricity rules that Olympus set themselves - I think that explains why, for instance, Sigma and Leica have produced 4:3 lenses faster than f2, whereas Olympus haven't.

All those who are using movements with (whatever) lenses to capture their images should not be doing that at all with a digital sensor if 4/3rds and telecentricity are to be believed.

Just because obeying the laws of telecentricity would seem to produce great lenses, doesn't imply the opposite (that not obeying the rules produces bad ones).

 

[Riley]

with the current state of microlens technology i dont think there is any doubt that the attributes of 'telecentric lenses' are less of a requirement.

Problem is if they changed something tomorrow, not all FT sensors are new, so it will take some time for older designs to be phased out allowing newer smaller lenses.

 

[Vivek]

Well, I think what Jono said is quite valid. The telecentric requirement would bring out the best in all digital sensors.

 

[Riley]

well here is what Ogawa Haruo says in an interview with DCW, a translation can be found at the link below

http://fourthirdsphoto.com/micro43/04.php

"OGAWA:“Certainly, it may turn out in future that telecentricity is not required so much as it is today. But at present, technological development hasn’t progressed to that point, so we continue to believe that telecentricity is important. Telecentricity in the 4/3 system does not exist merely as numbers sought in lens design, but represent a common consciousness among the manufacturers participating in the 4/3 consortium. We don’t know what the future may hold, but it doesn’t make sense to make lenses for use only in cameras that use a certain kind of sensor, so for the present we continue to design lenses with an important emphasis placed on telecentricity."

 

[Vivek]

Thanks for the additional info, Riley.

I am thinking about making a T/S adapter to use a few lenses on the M4/3rds.
I have to wait till I get the G1 first.

 

[Riley]

welcome Vivek
I tend to read more into these interviews than is actually said, and I got the feeling coming away from this one, that Olympus is beholding to other FT participants and older technology. On reflection I think the telecentric requirement will just become less, more in some lenses than in others.

Its no bad thing if they hang onto it, in my judgement it gives them a get out of jail card for the future, that of increasing sensor size.

 

[jonoslack]

Its no bad thing if they hang onto it, in my judgement it gives them a get out of jail card for the future, that of increasing sensor size.

HI Riley
My feelings exactly.
Vivek - I'll be interested to see how you get on with the G1. I think mine might have to wait a little while!

 

[Riley]

thanks Jono
when i first coined that over at the darkside it caused a hell of a ruckus...

 

[jonoslack]

thanks Jono
when i first coined that over at the darkside it caused a hell of a ruckus...

:ROTFL:

Mind you - I think it inconceivable that they'll increase sensor size - why take away the single competitive advantage (good small lenses) that they seem to have!

I'm not convinced by the 'full frame will win' argument, because, if it's that much better, then I think that those actually wanting it would prefer the M format (you know bigger than full frame).

 

[Riley]

inconceivable or not, ideally the lens sizes wouldnt change, a sensor increase would have to be within the scope that lenses would allow, while being large enough to make it worthwhile.

i dont see FF 'winning' either, the dSLR world is APSC, but Olympus must keep pace with its sector of competition, as it stands now, they are always going to be short of that, and market position suffers accordingly.

 

[jonoslack]

inconceivable or not, ideally the lens sizes wouldnt change, a sensor increase would have to be within the scope that lenses would allow, while being large enough to make it worthwhile.

i dont see FF 'winning' either, the dSLR world is APSC, but Olympus must keep pace with its sector of competition, as it stands now, they are always going to be short of that, and market position suffers accordingly.

Ah - now I really disagree - and yours is a mantra that is parroted over and over again on the Olympus forum.

Let's Look

the Canon 50D sensor is 14.9mm high (pixel density 4.5)
The Olympus E3 sensor is 13.5 mm high (pixel density 4.2)
that is approximately 10% taller (of course they have a different aspect ratio).
The problem right now is that the Panasonic sensors are not as good as the Canon sensors. However, the Olympus budget lenses are in a different league from the Canon ones.

If a decent sensor was used in 4:3, then nobody, but nobody would be able to tell the difference in IQ, but if Olympus put bigger sensors in their bodies then you would immediately lose the real advantage they have in edge definition and vignetting, added to which they would still be using a poor sensor!

What 4:3 needs is good quality sensors - their design is already streets ahead of the kludgy APS-c competition. Changing the design rather then sensor would lose them everything.

Which would be sad (IMHO).

 

[Vivek]

Jono, The pixel densities of 50D and the E3 based solely on the sensor area (and pixel count) are incorrect.

The E3 pixel (light collecting surface) are larger (NMOS) than that of the corresponding 50D's (CMOS).

 

[jonoslack]

Jono, The pixel densities of 50D and the E3 based solely on the sensor area (and pixel count) are incorrect.

The E3 pixel (light collecting surface) are larger (NMOS) than that of the corresponding 50D's (CMOS).

Hi Vivek
fine - I'm sure you're right - but it has no implications on my argument (I just put 'em in because they were there).

 

[Riley]

Ah - now I really disagree - and yours is a mantra that is parroted over and over again on the Olympus forum.

Let's Look

the Canon 50D sensor is 14.9mm high (pixel density 4.5)
The Olympus E3 sensor is 13.5 mm high (pixel density 4.2)
that is approximately 10% taller (of course they have a different aspect ratio).

thats pretty far from the real size story though isnt it, when we go 12Mp and we will, we will be back over 50D density. Its the density thats killing us, pixel size affects noise and DR. It seems the Kodak sensors were larger at 243sqmm, right now we run 225sqmm, the Canon APSC is 100sqmm bigger at 328sqmm.

The problem right now is that the Panasonic sensors are not as good as the Canon sensors. However, the Olympus budget lenses are in a different league from the Canon ones.

if you thought we lost 2/3 stop iso to canon, that would be about right. Im not so sure panny sensors are weak, they just deliver what they can for their size. Noise on the same DoF looks about the same

If a decent sensor was used in 4:3, then nobody, but nobody would be able to tell the difference in IQ, but if Olympus put bigger sensors in their bodies then you would immediately lose the real advantage they have in edge definition and vignetting, added to which they would still be using a poor sensor!

ok, lets say IQ is affected by a slightly larger sensor, you need to resolve how much you shoot wide open v/s the stop of better iso performance you would pick up. Vignetting is controlled from the diagonal, noise and DR are controlled by the area.

Increase diagonal from 21.6mm to 25mm, thats a 15.7% increase
Increase area from 225sqmm to 300 sqmm, thats a 33% increase
As areas are squared, we have much more to gain than we would lose

What 4:3 needs is good quality sensors - their design is already streets ahead of the kludgy APS-c competition. Changing the design rather then sensor would lose them everything.

Which would be sad (IMHO).

it would actually be better still with a bigger sensor. The patent allows that 25mm diagonal is the maximum, thats a 20x15mm sensor with an area of 300sqmm. As Canon APSC is 328sqmm, the nMOS/pMOS configuration would allow for less wiring and hence more area for well sizes, we would actually be infront.

so we could tackle DR and noise to comparable levels with APSC
the OVF would be larger b/se of the larger sensor
and the lens suite as you indicated is among the best
whats not to like for 1/4 stop stopped down for vignetting
(which can be eliminated incamera via software anyway) ?

 

[jonoslack]

I'm sorry Riley
I hear the arguments, and I've seen them all before, and I think they universally add up to much less than a hill of beans.

You can do as many percentages and diagonals and then apply areas to linear results as you like, but face it, the sensors are very nearly the same vertical height, and that is what one is normally taking pictures with:

You are never going to convince me that increasing sensor size by that much is going to improve image quality noticeably when other things are equal.

Anyway - the Canon has really good high iso, with a pixel density of 4.5, and the E3's is decidedly dodgy at a better pixel density of 4.2. It's the panasonic sensor which is the trouble NOT the 4:3 design.

The G10 has a tiny sensor in comparison with a pixel density of 34 instead of 4, and people are being kind about it's high ISO performance.

However - using full frame quite a bit (as I do). I realise how incredibly fast lens performance falls off at the edges, I'm quite certain that the Olympus wide angle lenses would NOT be okay using a larger image circle. (both the 9-18 and the 7-14 are starting to lose definition at the corners anyway.

To be frank, I think that arguments to the effect that Olympus should abandon their frying pan and jump into everyone else's fire are a load of old cobblers

 

[Riley]

the determinant of well sizes aside, and that imposition on performance, ISO, DR etc, there is more to sensor size as a physical limit in play though Jono, there is the perception of sensor size.

Olympus are losing money in the SLR div, sales are holding up only b/se they sell them very cheaply. If they sold for what they were worth volume would fall, and the tiny share they have already would shrink further to the 'inferior' APSC. Whatever they are doing at the moment isnt working, its as simple as that. as they say in the classics, people aren't buying it.

As to vignetting etc, the larger the crop factor the less vignetting is as an issue, there is scant difference between 2x and 1.73x but a measley 1/4 of a stop. Opticly lens resolution would improve, as the MTF of 60/20 we use now would become a more ideal 45/15. The lenses would actually perform better by 33%.

 

[jonoslack]

the determinant of well sizes aside, and that imposition on performance, ISO, DR etc, there is more to sensor size as a physical limit in play though Jono, there is the perception of sensor size.

Olympus are losing money in the SLR div, sales are holding up only b/se they sell them very cheaply. If they sold for what they were worth volume would fall, and the tiny share they have already would shrink further to the 'inferior' APSC. Whatever they are doing at the moment isnt working, its as simple as that. as they say in the classics, people aren't buying it.

Assuming you are right (I haven't a clue) I would say that is because they haven't got their marketing right in the larger world, and that their comparatively poor marketing would make matters even worse if they made their product like everyone else's.

I simply don't believe that the success or failure of a mid market camera hinges on very marginal differences in ISO results possible between aps-c and 4:3. More to the point I simply don't believe that the general public has the faintest idea about these things. Most of the people on dpreview certainly don't!

Lots of my friends have dSLR's these days, of various types - the vast majority of them really don't have any comparative information, nor could they care less. They bought their cameras based on what took their fancy when they were in Jessops - or what their mate told them, or which has the most pixels.

They are also profoundly uncritical of the results they get - if the picture's good they like it (whether it's grainy or whatever). I suspect that they're right as well! (and good usually means that little Jimmy has a smile on his face)

As to vignetting etc, the larger the crop factor the less vignetting is as an issue, there is scant difference between 2x and 1.73x but a measley 1/4 of a stop. Opticly lens resolution would improve, as the MTF of 60/20 we use now would become a more ideal 45/15. The lenses would actually perform better by 33%.

I'm sorry - figures again - but I'm extremely suspicious. What I KNOW is that I have NEVER had a good wide angle lens for either APS/c or full frame which weighed less than 900gm - most of them never get good even if you stop down, and by the time they've stopped having crap corners the middle's no good.

I'm not worried about vignetting that much either - it's resolution at the edges and corners which bugs me.

The rules of 4/3 and the diameter allowable aside, I can see absolutely clearly that if you put the magnificent 7-14 on a larger sensor it would be just as bad as all the rest. . . and so would the nice new 9-18.


If you say to me:

Actually, for people who read dpreview, soft corners and vignetting really don't matter a damn because the most important thing in the reviews is high ISO capabilty - therefore camera companies should sacrifice everything to the holy grail of high ISO

Then I couldn't deny that would be a wise course if that was the 'real market'.

But I'd think it really really sad to throw away their one advantage to become another pentax (without the nice primes). Definitely a case of throwing the baby out with the bathwater!

But if you tell me that the Zuiko lenses would perform just as well if you increased the sensor size by that much, then I'm sorry, you won't get me to agree!

And if you tell me that mid and low range camera sales depend principally on the quality of the image then I can only laugh.