The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

GRD2 concerns...

  • Thread starter David Paul Carr
  • Start date
D

David Paul Carr

Guest
Does anyone else have problems with
1) the corner sharpness of the GRD2 and 2) the extremely mediocre accuracy of the GV-2 viewfinder?

I've been testing the GRD2 against the G9 and, frankly, it doesn't doesn't seem as good as the Canon: more noise, less sharp, less able to resolve fine detail and with very soft corners. Poorer high ISO performance too.

It also takes far more effort to get the most out of the raw files from the Ricoh.

This is not to say it's bad. I've made a couple of (80 x 100+ cm @ 200 dpi) big colour prints from GRD2 files that look spectacular but surely a camera of that price with "pro" pretensions should have a better performance with regards to the two issues mentioned at the top?

Oh, and wouldn't it be nice not to have a sandpaper finish on the rear of the viewfinder? As it is it will destroy my glasses in days...
 
S

simon_t

Guest
Does anyone else have problems with
1) the corner sharpness of the GRD2 and 2) the extremely mediocre accuracy of the GV-2 viewfinder?
I do not have either camera, but I am currently choosing between the two (an maybe the LX2; see my other thread).

On the viewfinder, I can only say that I wasn't very comfortable or impressed with it in the store, but I have only handled it for a couple of minutes and have never used a supposedly-good non-DSLR finder before. I may yet start to like it if I get the GRDII+GV-2.

As far as the corner sharpness is concerned, I have downloaded all the RAW files I could get my hands on from the GRDII, GX100 and G9 and I have read Sean's three-way comparison. The results are a little confusing to me, though.

The three-way comparison and GRDII/G9 reviews seem to indicate (1) that the GRDII has better corner sharpness (wide open) than the GX100, (2) that the G9 has better corner sharpness (at a comparable focal length) than the GX100. Comparing the corner shots from the GRDII and G9 (at different equiv focal lengths), the GRDII's corner sharpness seems to be better than both others. On the other hand, if I look at the following test http://ricohforum.com/phpbb/viewtopic.php?f=39&t=553&sid=d3a68d2955090917c9fe8426970bfd26,
the GRDII's edge sharpness (tree branches) seems a lot worse than that of the GX100. What to believe?

The only direct comparison I have seen between the G9 and GRDII is that by Sean. It also seems to show that the GRDII's high ISO noise performance is slightly ahead of the G9.

Apparently, you have tested both and disagree on both counts (corner sharpness and noise). Can you elaborate a bit on your findings? Or would it be possible for you to upload some RAW files via yousendit or some other service?

Thanks,
Simon
 
L

Lewis

Guest
Lol @ the sandpaper finish. I can relate to that...I have the GW1 lens attachment and it seems Ricoh use that rough textured plastic on all their 'accessories'. I put black tape round mine as it was getting white scratches on from my finger nails where it was 'shaving' them evertime they caught the lens edge by the camera grip. Maybe you could do something similar with the rear of your view finder dude?
 
D

David Paul Carr

Guest
I don't really enjoy camera testing but, because I really needed to know which camera I should prefer for a particular job, I have spent part of today doing some comparison shots which I would like to share here.

I have tried to be fairly methodical:

I have taken the same three pictures with the GRD2, GX100 and G9.

Each time, the camera was placed on a tripod with a leveled Manfrotto panoramic head. The aperture was f4 (or as near as possible) and the focal length 28mm for the Ricohs and 35mm for the Canon. Focus was centre point autofocus in all cases and exposure was automatic. All taken at 100 ISO. The only real possible inconsistency is that it is impossible to keep the camera parallel to the subject plane or at least equally unparallel for each shot.
That said, I think these pictures prove my point about the GRD2's edge softness but I would be interested to know what others think.
The Canon files begin with "IMG", the GX100's with "R" and the GRD2's with "_00".
I am sufficiently disappointed with the GRD2's corner performance that I am thinking about returning it. This doesn't just look soft to me, it looks poor quality and certainly not what I expected. I mean, the GRD2 has a prime lens and the others (both cheaper) have zooms...

The raw files can be found here (I hope - this is the first time I try to use yousendit).

http://www.yousendit.com/transfer.php?action=batch_download&batch_id=Mmd0QndEb0JreERIRGc9PQ

Yes, I know this isn't scientific!
High ISO comparisons will have to wait for another time...
 
Last edited:

sizifo

New member
I've had a grd II for a month and just had a quick look through my photos. On most photos I find the corner sharpness very difficult to judge, but for sure I have many examples which display no problem whatsoever. Could there be something off with your lens?

Is f4 an important factor in your assessment?
 
D

David Paul Carr

Guest
F4 simply because I wanted to avoid any issues that might be connected to using the lenses either wide open or fully stopped down. Working, I suppose, on the basis that most lenses give their best performance stopped down a couple of stops from the widest aperture.
Yes, there may be something wrong with the lens. And yes, it is difficult to judge...
 
V

VladimirV

Guest
I just had a look at your first file and the G9 is pretty unsharp at the corners and the GX-100 is sharpest. I think you should try a different GRD II or GRD I and see if there is any difference, maybe your camera has a problem with the lens.
The G9 appears worst to me with the least details and softest image, maybe I need to use different settings for processing it. The GX-100 seems best in regards to details and corner to corner sharpness.

Regarding your 2nd point the OVF is not very accurate but neither is the G9 OVF and at least the Ricoh shows you 100% whereas the G9 shows only 80% and is obstructed by the lens. I have the GV-1 viewfinder but almost never use it on the camera as I prefer framing with the LCD, this is why I do not like dSLRs ;). I use it more as a scouting tool to help compose and visualize the shot before taking the camera out. It's also helpful to decide if I should use the 21mm lens or not :).
 
S

Sean_Reid

Guest
I haven't had time to look at your files yet but a couple thoughts:

1) It's impossible to compare the G9 lens and the stock GR II lens at the same focal length. That being the case, depth of field is going to be a confounding variable when the G9 is at a 35 mm EFOV and the GR2 is at its 28 mm EFOV. When the GT-1 is added to the GRII, the files are soft in the far corner.

2) Any camera tested for resolution must be absolutely square (vertically and horizontally) to the test board.

3) It's essential to do resolution testing with focus bracketing and then comparison of the resulting files to find the distance of peak focus. AF with these cameras is not consistent or accurate enough for critical res tests.

4) Resolution can change at different focus distances.

I'd have to look at my results again but I don't recall either camera being soft in the corners at the 28 mm/35mm (respectively) effective fields of view.

If your GR2 is soft in the corners at F/4, the lens may be defective.

Also, I see the best results from the GR2 using C1 4.0 or SilkyPix.

Photographers who are not used to RF cameras may not realize that no window finder can show fixed frame lines that will be accurate for all distances. As I mentioned in the GR2 review, the GV2 finder is conservative, much like the frame lines in an M8 or R-D1. With any of these, one has to learn how the edges of the picture, at various subject distances, will fall, relative to the framelines.




Cheers,

Sean
 
7

7ian7

Guest
David, from what I've seen of your work, I certainly take your impressions seriously. The idea of a defective camera only comes to mind because your particular frustrations with the GRD2 lens/sensor haven't really been expressed elsewhere. It is irksome; my guess is we'd all like to be done with the endless micro-scrutiny about these tools, but so far none really offers the kind of reassurance we experience when using the bigger guns, or the film forbears of these kind of cameras. It'll get there. Meanwhile, I imagine that wrestling with these limitations across the three best contenders must exacerbate your dilemma, even if it is sort of an enviable headache.
 
7

7ian7

Guest
(All that said, I'm using my GX100 all the time, and I don't actually remember a day when I experienced truly frustration-free results from my Hexar, or Pentax 67, etc etc)
 
D

David Paul Carr

Guest
Sean, I take your point about testing. I wouldn't have posted anything if it hadn't been for Simon t's request. And, yes, I suppose I will have to get used to the conservative framelines although I can't for the life of me see why they need to be so inaccurate. After all, an optical viewfinder is a pretty basic piece of equipment...

7ian7, I suspect we have a similar approach to these little cameras. I have a gut feeling that there is something very interesting to be done with them. At the end of last year, when I looked through my pictures, I was surprised how many of the ones I thought of any value had been done with these (as opposed to my big "pro" Canons).
I suspect it may have something to do with being forced to think (and concentrate) a little harder if you want to get the best out of them. Also, if you think of them as a mini view camera rather than an ersatz Leica - using the lcd as you would a ground glass screen - the whole process becomes very interesting...
 
7

7ian7

Guest
Nice David, I agree.

And to Sean, I don't mean for you to be done with your micro-scrutiny!!! I'm happy to pay for the benefit of your efforts in that regard!
 

scott kirkpatrick

Well-known member
Sean, I take your point about testing. I wouldn't have posted anything if it hadn't been for Simon t's request. And, yes, I suppose I will have to get used to the conservative framelines although I can't for the life of me see why they need to be so inaccurate. After all, an optical viewfinder is a pretty basic piece of equipment...
I think the point Sean was referring to is that even a perfect optical finder shows the angle of view correctly only for one object distance. For the M2-7's that distance was 1 m. The M8 brought this reference frame distance down to 0.7 m, which is not very realistic, perhaps out of a desire to always err on the side of showing less than you will actually get. As you focus further away, the lens to image distance decreases and the angle of view increases. The effect is larger the longer the focal length.

This drives people who believe that what you see should be what you get slightly crazy. The sanity-inducing mantra that results is that you must learn to know your lenses and the viewfinder is only a guide, hopefully with enough outside the indicated frame to see what you are really doing.

scott

edit: Sean and I overlapped.
 
S

Sean_Reid

Guest
And, yes, I suppose I will have to get used to the conservative framelines although I can't for the life of me see why they need to be so inaccurate. After all, an optical viewfinder is a pretty basic piece of equipment...
The fact that they're a basic piece of equipment is exactly why they can't be accurate at all distances. See the "Introduction to Rangefinders" article. What many of us often wish for is that frame lines were optimized for middle distances (say 1 - 2 meters) even at the cost of some accuracy loss close up.

Cheers,

Sean
 
S

Sean_Reid

Guest
I think the point Sean was referring to is that even a perfect optical finder shows the angle of view correctly only for one object distance. For the M2-7's that distance was 1 m. The M8 brought this reference frame distance down to 0.7 m, which is not very realistic, perhaps out of a desire to always err on the side of showing less than you will actually get. As you focus further away, the lens to image distance decreases and the angle of view increases. The effect is larger the longer the focal length.

This drives people who believe that what you see should be what you get slightly crazy. The sanity-inducing mantra that results is that you must learn to know your lenses and the viewfinder is only a guide, hopefully with enough outside the indicated frame to see what you are really doing.

scott
As so often is the case, Scott is on the money and concise here.

Cheers,

Sean
 

Lili

New member
Also, if you think of them as a mini view camera rather than an ersatz Leica - using the lcd as you would a ground glass screen - the whole process becomes very interesting...

Very good point; that is very much how I find myself using mine; like a very fast and tiny view camera. Use the LCD to establish the edges of the frame so to speak.
 
S

Sean_Reid

Guest
There was a broader question posed above about seemingly conflicting test results from different sources. Speaking generally, and only generally, here's my advice on that point.

1. Look at the methods used for testing by each source. For example, any resolution testing (including res. charts, MTF, real world objects, etc.) is suspect if it wasn't conducted with careful focus bracketing. Otherwise, an important confounding variable is introduced.

2. Consider the possible confounding variables in the test and consider the results accordingly.

3. Draw upon multiple sources whose methodology and approach you trust.

4. Remember that there can be sample variation among cameras, lenses, etc.

Again speaking generally, I often see comparison tests published on the web which really don't tell us anything at all about the nominal quality they're supposed to be testing. One can never entirely eliminate confounding variables in these sorts of tests, but the best possible effort should be made to do so.

Lastly, it's important to remember that nothing is ever proven. The best that we can do is to deny the null hypothesis. (The null hypothesis states that there is no relationship between A and B.)

Cheers,

Sean
 
S

Sean_Reid

Guest
Very good point; that is very much how I find myself using mine; like a very fast and tiny view camera. Use the LCD to establish the edges of the frame so to speak.
I've always thought about it that way and discussed the Leica D2 that way when I first tested it.

Cheers,

Sean
 
Top