G1 impact on lens character?

[Jamie Pillers]

I'm not yet reading much here about the images the Panasonic G1 creates with M-mount lenses. I'm very interested in hearing about whether you G1 owners feel the character of your Summicron or Noctilux or Ultron or (your favorite M-mount lens) images look the same using the G1 as they did on film. It would be great if you could describe any differences you see. Thanks much.

[skibeerr]

Go to flickr and do a search on Panasonic G1

I found myself able to pick out foto's made with Leica lenses from the thumbnails.

Cheers,

Wim

[M4cr0s]

Pardon my simple and naive response, but be not the "character" of a lens pretty much identical, regardless of medium? The only exception being the fact that high-resolution digital sensors have a tendency to push lenses up against a wall, resolutionwise (can't imagine that'd be much of a problem with a 2x crop sensor on a Leica-lens though)? Also, due to the AA filter on digital sensors, I presume some lack contrast and sharpness might be evident, hence the need for boosting in PP.

If you search flickr for "panasonic g1 summicron" or similar, I think you'll find a good number of pictures showing what results people have. Over at the Panasonic-forums on dpreview.com there are also a great number of threads with examples/experiences from people trying M-mount lenses on their G1's.

/Mac

[digitalintrigue]

I've used Summicrons, Summarits, Biogons, Planars, vintage Canons, Nikkors, and Carl Zeiss Sonnars, and lots of other lenses on the G1 (M mount and many others, such as Leica R.) The individual lens rendering characteristics are obvious assuming one doesn't shoot at such a small aperture as to masks those characteristics.

[Jamie Pillers]

I have trouble seeing the true character of lenses when looking at small flickr images, except in extreme cases (when photographer tries to show off bokeh, etc.). I'm interested in hearing from those folk that have used the G1 and M-mount lenses to make prints, and whether they still see the lens characteristics they enjoyed when using film.

[ampguy]

From what I've read here (but not seen!) is that the kit lens can look like a Noctilux!! Can't wait for the G1H, but for video use ;)

[Tuolumne]

They all look the same to me, and I have a G1 and a bunch of M glass. They even look the same as the G1 kit lens. There have been a number of threads on the topic here. One poster, can't remember who, said it was because all of the G1's noise processing destroyed the micro-contrast by which lenses could be told apart. YMMV.

/T

[Jamie Pillers]

Thanks. I've read some of those posts and they more often seem to be about the photographer's impression of the camera... not its affect on the lens characteristics that the photographer was enjoying with film.

My hidden interest here is that it seems to me if a small-sensored camera like the G1 can produce images with, for example, the 'bokeh' king Summicron that are no different than the images they were getting using film.... then it seems that film is indeed on its last legs. Why would we bother with the film processing trouble if we can get the same images with digital + M-mount lenses?? Please... I'm not looking to raise the old digital vs. film arguments here; we've all probably had enough of that for now. It just seems that there are a lot of experienced Leica users here that are trying out the G1 and I'm trying to figure out if this digital-plus-high quality lens combo is in fact a big paradigm change... or just another distracting toy.

I was expecting to hear RFF members write about how its fun to use the Leica glass on the G1, but that the digital-ness of the G1's process causes a change to the image quality and that they therefore would need to keep the film bodies in order to get the lens characteristics to really show up in their prints. I haven't heard that yet... one way or the other.

[f16sunshine]

Noise
LotZah Noise
I have reviewed a few sets a friend sent me using FSU M39 lenses (j8,9).
Looks like a P+S. Of course it is his early go at the camera but, I'm not convinced that the true character of any lens will come through without the signature of this sensor/engine masking it.

The concept of the short register EVIL cams is great. I think I will wait t see the offering from Samsung before giving one of the M4/3 cams a try.

[Tuolumne]

Thanks. I've read some of those posts and they more often seem to be about the photographer's impression of the camera... not its affect on the lens characteristics that the photographer was enjoying with film.

My hidden interest here is that it seems to me if a small-sensored camera like the G1 can produce images with, for example, the 'bokeh' king Summicron that are no different than the images they were getting using film.... then it seems that film is indeed on its last legs. Why would we bother with the film processing trouble if we can get the same images with digital + M-mount lenses?? Please... I'm not looking to raise the old digital vs. film arguments here; we've all probably had enough of that for now. It just seems that there are a lot of experienced Leica users here that are trying out the G1 and I'm trying to figure out if this digital-plus-high quality lens combo is in fact a big paradigm change... or just another distracting toy.

I was expecting to hear RFF members write about how its fun to use the Leica glass on the G1, but that the digital-ness of the G1's process causes a change to the image quality and that they therefore would need to keep the film bodies in order to get the lens characteristics to really show up in their prints. I haven't heard that yet... one way or the other.

Don't the M8 and Epson R-D1 accomplish the same thing, and film still isn't dead (yet).

/T

[digitalintrigue]

There are several threads/posts about using M lenses, and G1 image processing.

One poster is adamant that the G1 'manhandles' the images, although when asked to show this, he stopped posting about it.

There are other posts about G1 and M lenses, saying that the look is the same regardless of lens. This can't be, because the kit lens is slow. I'll post some photos shortly using a Summicron-C 40 and the kit lens at 40mm.

[Jamie Pillers]

At $4000-ish and $2000-ish, these two bodies are not film killers. But when we start to see things like the $800-ish G1 coming... that's different. Is it the new paradigm? I don't know; I don't own it (yet). :-)

[Jamie Pillers]

Thanks. That's the kind of information I'm looking for. I'll do some more searching around here. And watch future posts of course, to see whether this micro four-thirds thing is a tidal wave or a ripple. :-)

[Jamie Pillers]

Yes... I think this next year is going to provide some interesting thought-provoking equipment from Olympus, Samsung... maybe Nikon. :-)

[Jamie Pillers]

I'd LOVE to see those, because the only M-mount lens I currently own is the Summicron 40! :-)

[Jamie Pillers]

Gotta go to work now... I'll check back tonight to see what's brewing here. Thanks everyone for your responses to my post. Much appreciated. As always, RFF 'rocks'!

[digitalintrigue]

I was just doing some quick comparisons of the Summicron-C 40/2 to an Olympus Pen F Zuiko 38/1.8 when I came across this thread, so here are the shots. This is by no means designed to be a comprehensive comparison of any of these lenses! It's just a single view from my desk.

These were taken on a tripod at the close focus distance of the Summicron (the kit lens and Pen lens will focus MUCH closer -- the last photo shows how close the Pen lens will get.) Focus manual (on the kit lens, too) on a highlight on Gumby's left eye.

ISO 100, shot RAW, converted to DNG, opened in Photoshop CS3, color balance 'as shot', no sharpening or adjustments applied, then 'save as' with quality = 6. The smaller versions were adjusted to 1024 max width then save as with quality = 6.

The kit lens is lacking contrast, I suppose due to the backlighting. Looks like the 40 year old lenses handle that better, with the Summicron having more contrast than the Zuiko.

The Olympus Pen is a 38/1.8 using a Pen F to Micro 4/3 adapter from Taiwan, and the Summicron is using a Novoflex Leica M to Micro 4/3 adapter.

All lenses are at max aperture except the Zuiko which is stopped down from f/1.8 to f/2, to match the Summicron.

Is this a fair comparison? No, as the kit lens is three full stops slower, which means selective focus is very limited (it was also not surprising it was more difficult to focus manually looking through a lens at f/5.6 than at f/2.)

Click on the photos for full size images (4000x3000)

Summicron-C 40/2 @ f/2

http://www.restoration35.com/photo_posts/summ40.jpg
(http://www.restoration35.com/photo_posts/summ40L.jpg)
Olympus Pen Zuiko 38/1.8 @ f/2

http://www.restoration35.com/photo_posts/Pen38.jpg
(http://www.restoration35.com/photo_posts/Pen38L.jpg)
Panasonic 14-45/3.5-5.6 @ 40mm f/5.6
http://www.restoration35.com/photo_posts/kit40.jpg (http://www.restoration35.com/photo_posts/kit40L.jpg)

Close focus distance of the 38/1.8 Pen:

http://www.restoration35.com/photo_posts/Pen38close.jpg

[digitalintrigue]

Just for kicks, a slightly different angle, but all other details the same:

Carl Zeiss Opton Sonnar 50/1.5 @ f/2:

http://www.restoration35.com/photo_posts/sonnar.jpg
(http://www.restoration35.com/photo_posts/sonnarL.jpg)

http://www.restoration35.com/photo_posts/opton.jpg

[Amedeo Contax RF to Leica M adapter + Novoflex Leica M to Micro 4/3]

[ampguy]

Is it just me, or is the kit lens flarey?

[kbg32]

I own a G1 and love it. If you are looking to use your M mount lenses, their character is not enhanced at all by the chip in this camera. I have tried Leitz, Zeiss, Canon, and CV lenses. Nothing is better then the kit lenses that are made for the G1.

[digitalintrigue]

It's not just you, the kit lens is definitely lacking contrast. These were taken without hoods, except for the Summicron-C which has the rubber hood.

[digitalintrigue]

The kit lens is great, if you are happy with slow. If you like to shoot wide open, f/5.6 isn't very. :) I'm not sure how a sensor could enhance adapted lenses, as it has no idea any lens is even attached.

[NickTrop]

The primes are better - obvs. The kit does look "flare-y" but what do you expect... how many groups/elements does that thing have? Meh - give me a fast 6/4 f1.8 50mm. I'll zoom with my feet. The kit lens is also slow - little bokeh and flatter. Meh (again). The image is very "digital" - technically acceptable but nothing interesting going on. Don't forget - the lens isn't the only thing with "character" - film has it too. And - "character" is usually seen at wider aps (I'd say up to F4), after that they all look good (50mm primes - that is, not zooms) from a lowly banger $10 Chinon, to a brand new $1000 'Cron - all sharp and "contrasty"... If you want to add character, I think Fred Miranda has a "Character" Photoshop Action he sells for $20 on his site. You control the amount of character with a slider...

[Jamie Pillers]

I gotta get that Character Slider!! And I can't wait until the new Oly comes out with the Ansel Adams scene mode (shutter won't fire if there's anything man-made in the frame).

Nick, regarding your comment about film adding "character": I could rephrase my original question and ask "Does the digital sensor somehow oversharpen your Sonnar lens images? Does it mess with the OOF qualities of your Noctilux images? Does it add too much punch to your Summar images?" I guess I still would like to hear from someone that's made, say, an 8x10 print from a negative out of their Leica M3/50 Summicron kit and another 8x10 print of the same scene from a digital file out of their G1/50 Summicron kit. What difference(s) do they see? Anyone? :-)

[kbg32]

Jamie, you are really comparing apples.

[djonesii]

I have done a fair bit of work with my G1, while I have not done extensive comparisons w/ Kit lens versus my other ones, it goes something like this ..

The need for speed is not met by kit lens, nor in the 25mm length, in my price range, there is no M mount lens either, therefore, I use C mount lenses.
http://www.jonesii.net/2009%2003%2019%2012mm%2025mm%20cine/content/bin/images/large/P1020359.jpg

This is from a 25mm Wollensak 1.5

http://www.jonesii.net/2009%2003%2006%20Flowers/content/bin/images/large/P1020056.jpgThis is from a Kodak 25mm 1.4

There is no way the kit lens can do this.

In the 40mm range, I can afford the 40mm 1.4 CV, and it does a super job IMHO

http://www.jonesii.net/2009%2002%2014%20CV%20shots%20G1/content/bin/images/large/P1010473.jpg
This is no flash, hand held in a dark room. Again, no way that the kit lens can do this.

The 12mm/4.5 and the 25/4, and the 50mm Sumicron are not near as nice to me. and in fact, as soon as I get around to it, they will be on the for sale part of this site.

Much the same is true of the 12mm Cine lenses, they do have some cool looks, but they loose too much in the image circle. The C mount cheapos from surveillance cameras are very sharp, but have little character to me.

Hope this shows you some things you were looking for.

Dave

[digitalintrigue]

I could rephrase my original question and ask "Does the digital sensor somehow oversharpen your Sonnar lens images? Does it mess with the OOF qualities of your Noctilux images? Does it add too much punch to your Summar images?"

How could the sensor do this exactly? Why would it oversharpen a Sonnar image as opposed to any other image? Why and how would it change the OOF qualities of a Noct? The light is focused or it's not, how would the sensor change this?

Prime lenses are going to perform differently than the kit, just like a prime lens will perform differently than a kit lens on a D40, just like a prime will perform differently than a general purpose zoom lens does on an F5.

[JoeV]

Interesting comparisons, and discussion. We need to remember, when discussing the kit lens, that when it's being used you are looking at the composite results of the camera's RAW image file being manipulated in software, if SilkyPix is being used to process the RAW files, to correct for off-axis abberations. Panasonic designed the camera, kit lens and software to function as an integrated system (unless you're processing the RAW files using something other than SilkyPix). In this regard the in-camera JPEGs can be revealing, since they, too, operate on the image to correct off-axis abberations when using the kit lens.

When you adapt a 3rd party lens to the G1 it's no longer the same camera/lens/software integrated system.

These sorts of lens comparisons may therefore be more accurate by using in-camera JPEGs. The degree of post-processing the RAW files undergo is dependant on the type of RAW developer software being used.

I also find it funny that the Gumby tests used lens shades on the 3rd party glass, and not on the kit lens. Not exactly a fair comparison?

I'd also like to see the full-frame images, and discuss off-axis abberations between the various lenses, as that's the area where you'd expect the biggest difference between the kit lens and 3rd party glass, for reasons stated above.

~Joe

[digitalintrigue]

As stated, the tests are not comprehensive, and it was even stated that it wasn't fair. ;) This wasn't meant to be a test for flare. I'd be happy to do that, however. :) The contrast differences just happened, and yes, only one lens had a hood, so two lenses were handicapped in that regard, including one of the adapted lenses. It is interesting that a 40+ year old unshaded adapted lens did handle flare better, however.

Silkypix wasn't used, the workflow was outlined.

But that wasn't the reason for the comparison; the biggest difference that adapted lenses offer is speed and the resulting character and rendering differences, which are a direct result of the additional speed. It's pretty obvious (as in the djonesii post) that the kit lens is bokeh-challenged due to its slowness.

Full frame images are linked as stated in the post.

[Avotius]

Luminous Landscape (http://www.luminous-landscape.com/reviews/accessories/m-43.shtml) had a thing on it not too long ago

[Jamie Pillers]

Sensors manipulate images with software. Camera designers have to build in all sorts of "manipulations" into the in-camera software that takes the electronic data from the sensor. The software can correct all sorts of things that they've not been able to solve 'mechanically' yet... things like aberation, vignetting, noise, etc.. Sharpening is a software manipulation that could "revise" the image seen through a Summar for instance.

[digitalintrigue]

Of course, it can do that with the kit lens (although noise isn't lens-related.) But how does it know a Summar is mounted on the bayonet? ;) Or as I said earlier, any lens at all?

[To use an adapted lens, the camera must be set to 'shoot without lens.']

[Jamie Pillers]

No, I'll leave the computer comparisons to someone else. ["apples"... :-) ] I'm trying to understand the difference between sensor-based images vs. film-based images using the same lens. :-)

[Jamie Pillers]

Dave... nice images. Now... go try to make the same images with those same lenses, but mounted on a film camera. Then come back and tell if you see differences. :-)

[Jamie Pillers]

And I'm not interested in comparisons of the kit lenses with M-mount lenses. I'm asking about an apples-to-apples comparison... How do M-mount lens/G1 images compare with images made with the same lens on a Leica M body?

[digitalintrigue]

Well the sensor can't make those adjustments you talk about on M lenses. So I guess I'm a bit confused on that part.

But the images will be very similar. Other than the film look vs. digital look, the other difference will be the angle of view. The bokeh and rendering and individual lens characteristics will still show.

It's no different from when a 50 Summicron is on an M3 or an M8 -- there will still be a Summicron signature, unless the lens is stopped down to the point where it becomes a 'lowly banger $10 Chinon' -- to quote Nick. :)

[JoeV]

Thanks, Jamie, for clarifying the purpose of your inquiry. The only other issue I can think of that might effect the image quality of an M-series lens on the G1 vs a film camera is off-axis falloff, due to the digital sensor requiring a tele-centric lens design to capture more off-axis light. With a non-telecentric lens on the G1 you may have more vignetting and off-axis light falloff compared to a film camera using the same lens.

OTOH, some of the newer Leica glass, intended for the M8, have a somewhat telecentric design (the M8 having a similar issue as the G1, but less severe due to the M8's micro-lens design) and may therefore operate a bit better on the G1 in this regard.

I was also thinking today about older lens designs and how their designs often are compromises between center and edge performance. Those designs that have poorer edge performance would benefit better from the crop factor of the G1, while a lens design better optimized for 35mm film may in fact not perform as well on the G1 due to the center of the lens being compromised for better corner performance.

I look forward to more testing in this regard.

~Joe

[Jamie Pillers]

Joe... I'm learning something here!! Thanks. Yes... the crop factor can affect the outcome. One of the things I've been concerned about actually is that I LIKE that light fall-off and less-than-perfect focus out on the edges. This often enhances the central figure, especially in portraits or portrait-like images. So with less-than full frame sensors I'd lose that characteristic! Thanks.

[digitalintrigue]

The best option (for highest performance center-to-corner) for adapted lenses for the G1 are generally not rangefinder lenses, due to the non-telecentric design that Joe mentions. This is why Leica designed the micro lenses on the M8.

There are other reasons why RF lenses aren't always the best choice, one being that RF lenses can't focus very closely. This is exactly why I was doing a test of the Pen F 38/1.8 vs. the 40/2 Summicron-C, the former being an SLR lens and of course the 40/2 an RF lens.

If the goal is central sharpness and less than stellar corner performance, then rangefinder lenses, especially wide angle, are exactly what you want. Many (though not all) of these lenses exhibit so-called 'smearing' in the corners, due to the light hitting the sensor at oblique angles. This isn't an issue for film but it is with sensors.

There are many threads on getdpi.com about this, do a search for 'corner smearing.' Here is one: http://forum.getdpi.com/forum/showthread.php?t=5566

[KEH]

I really like the way the G1 renders my CV40/1.4 SC. The single coating gives a pastel look to the raw image, which is a great point of departure for further processing. An example:

Original photo, minimal sharpening and clarity:
http://farm4.static.flickr.com/3657/3534978461_b61b8c4a47_o.jpg

With Lightroom 'direct positive' preset, then desaturating the blues:
http://farm4.static.flickr.com/3590/3534978557_18b5af76ec_o.jpg

With Michael Gray's Pantomic X preset:
http://farm3.static.flickr.com/2094/3534978649_8ae7fe297c_o.jpg

Finally, a current favorite taken with this lens:
http://farm4.static.flickr.com/3540/3338548093_c69689ce12_o.jpg

I am reasonably sure that I am getting more of the 'character' I like from this lens than I was getting with my R-D1.

Cheers,
Kirk

[250swb]

I have been using the M mount adaptor for a little while now and there are two differing conclusions to the idea that the G1 affects the image quality, or doesn't affect it.

The first thing to remember is that the Four Thirds sensors are designed to be used with near telecentric lenses, where the light path from the lens to sensor is as near a straight line as possible. In addition the part of the image circle used is a small proportion of that which is theoretically possible and is why Olympus and Panasonic lenses are so well corrected and sharp edge to edge.

With these two things in mind the image quality of a wide lens used on the G1 is disappointing looking at the edges and corners of the image. The IQ falls off because of the light spread and angle it is reaching the sensor, and this wouldn't be apparent on film or a larger conventional FF type sensor. But as soon as you get to a 40mm lens and up the overall IQ is at least equal to what you would get on film, if not better because so many parameters are more controllable. The Bokeh looks the same to me from a 50mm Summicron as it does on Velvia, and the sharpness is unaffected, which isn't surprising when recently the G1 was shown to give a higher resolution than the Nikon D3.

Steve

[Jamie Pillers]

Thanks Steve. So... do you think it will be possible for ANYONE to design a fast wide lens for these small sensor cameras?

[ampguy]

I think it could be possible, but only with Panasonic's assistance since their RAW lens corrections are not in the open right now. Price might be another barrier.

The additional proprietary RAW options and metadata support of camera specific maker lens and sw always gives them a slight advantage over 3rd parties lens suppliers.

Thanks Steve. So... do you think it will be possible for ANYONE to design a fast wide lens for these small sensor cameras?

[250swb]

So far neither Olympus or Panasonic have shown any desire with four thirds to build a fast and wide lens. At the moment I think the real problem they have is not the designing of such a lens, but the eventual size of it and its impact on what is supposed to be a miniature system.

The new Panasonic 7-14mm is supposed to be a little jewel of a lens, but only f4. This is mitigated on the DOF front by the inherent depth of focus that a 7-14mm would have anyway in normal shooting, and perhaps limiting even an f1.4 lens to significant DOF effects only with close ups. And while f4 isn't a fast speed yet again this is mitigated a bit by the image stabilisation making slower shutter speeds practical.

So perhaps this reflects the way Olympus and Panasonic are looking at the specification of their lenses, and also in keeping the faster speeds to the longer lenses where it makes more difference to the hand holdable performance of the lens?

Steve

[djonesii]

Dave... nice images. Now... go try to make the same images with those same lenses, but mounted on a film camera. Then come back and tell if you see differences. :-)

For the flowers, as near as I know, there is no way to even attempt this on 35mm as the lenses are 16mm movie lenses.

I have tried indoor color, no flash, and given up on it, b&w works out good, but in 35mm, I just could not get nice clean images. MF with the Bronica RF645 gives me a shot at it, but the Fuji GA645 with it's little built in fill flash gives digital a run for it's money.

Dave

[Benjamin]

Well let's see.. http://cgi.ebay.co.uk/Lens-Summicron-2-35mm-Tilt-Shift-f-Panasonic-G1_W0QQitemZ350208554143QQcmdZViewItemQQptZCamera_ Lenses?hash=item518a0e749f&_trksid=p3286.c0.m14&_trkparms=72%3A2109|66%3A2|65%3A12|39%3A1|240%3A13 18|301%3A0|293%3A2|294%3A50 (http://cgi.ebay.co.uk/Lens-Summicron-2-35mm-Tilt-Shift-f-Panasonic-G1_W0QQitemZ350208554143QQcmdZViewItemQQptZCamera_ Lenses?hash=item518a0e749f&_trksid=p3286.c0.m14&_trkparms=72%3A2109%7C66%3A2%7C65%3A12%7C39%3A1%7C 240%3A1318%7C301%3A0%7C293%3A2%7C294%3A50)

Oh dear :(

[digitalintrigue]

Can the kit lens do this? :D

Carl Zeiss Sonnar 50/1.5 @ f/2.0

http://www.restoration35.com/photo_posts/bamboo.jpg

http://www.restoration35.com/photo_posts/dragonfly.jpg

[KEH]

http://kirkh.smugmug.com/photos/563639670_Pjr7F-O.jpg
CV 35/1.4 MC - a lens I was never thrilled with on the RD1. But it seems quite lovely on the G1. Zone focused at F8.

Cheers,
Kirk

[mawz]

There's a few other things that need be considered.

Remember the crop. If the lens's signature at wide aperture is defined by any aspect of the edge performance (vignetting, softness, non-flat field rendering) you simply won't see it, or at least that portion of it. My Nikon 35mm f1.4 AI-S becomes rather boring at wide apertures because of this.

Remember the pixel density. The G1's sensor is extremely demanding on lenses. It has over twice the pixel density of the M8 (The M8's sensor is almost exactly twice the size) or your average high-MP FF DSLR. Consider what a 24MP M8 or 48MP D3 would demand for lens performance even in the sensor. This means the kit lens is a truly stunning design with regards to resolution. It has to be to deliver adequate results from the G1's sensor. This also means that any lens not capable of delivering very high resolution numbers will not stand out when compared to the kit lens. Conversely lenses with high centre performance and poor edge performance may be better on the G1 than they are on larger formats.

Telecentricity matters. Realistically you're going to want SLR lenses or native lenses for the wide focal lengths due to this. A cheap Tamron 28/2.5 Adaptall outperforms most of the 28mm RF lenses on the G1 for this reason. Its not a serious issue with 35mm+ lenses. This also means that native m43 lenses are going to require processing adjustments for optimum performance.