Reading mode: Light Dark

G1 and "muddiness"

Go to last post
L

liquidsquid

Senior Member
Messages
1,992
Reaction score
264
Location
E Bloomfield, NY, US
So I have my replacement G1 for the other one that managed to somehow get a strange fault which even currently has Panasonic scratching their heads... That is another story.

A thread that hit 150 was talking about image muddiness, and how the person saw it in the foliage and other fine details in the background of landscape images. Basically they are seeing the G1's noise reduction at work, and if it is bothersome, turn down the setting in the film mode they are using if you want to keep the detail in the camera JPGs as high as possible. I always run the cam with NR turned down in all film modes. If noise is a problem in an image, I will target it with a raw processor anyhow.

Note: Not a perfect test, but a description of what muddiness may mean, and how to not get smudged foliage in some situations.

For example, ISO800, G1 with default +0 settings in film mode. You can see the chroma smudging used for NR at work here where the stick's color exits the sticks outlines, this is a common NR technique:



This is the same image using SilkyPix with fairly aggressive NR but favoring the detail:



With ISO100, things are different, the in-camera NR is not as aggressive, and I don't notice much difference between the in-camera JPG and the SilkyPix image until I start applying some sharpening. IMHO the SilkyPix sharpening is much more natural to my eye, but I don't see muddiness in either image.

ISO 100, all default film mode settings, same scene, different crop as above. Moderate sharpening applied in post.



Silkypix of the same with "Pure Detail" sharpening moderately applied.



--
My pictures...
http://12.168.34.68/photos/default.asp
http://www.markwyman.com/photos/default.asp
 
It's hard to judge those pictures because I don't know exactly how they should look, but none of them looks particularly clean to me. But thanks for posting them.

I give up. I'm going to start looking for another camera.
 
If it´s really critical to you, I think you need to use NR -2 or shoot raw, and use your own noise reduction software.

It depends strongly on the pattern in the picture, under some circumstances the incamera NR can produce that muddiness you don´t like. I think especially reds and browns are affected.

A comparison I posted some time ago may interest you: http://forums.dpreview.com/forums/read.asp?forum=1033&message=30532026

G.L
 
I guess if you are looking for 100% crop perfection at a camera's maximum resolution, the G1, or any camera in its class is not for you. Look at the Sony top-end A900 dSLR, take it's output JPEG and reduce it to 12Mp, and then you have what you want.

A lot has to do with the Bayer interpolation from the sensor, and you really can't pull detail from an interpolated signal to make a 100% crop have "perfect" resolution. That is why the Sigma looks better to you, the Foveon sensor at 100% does not need to interpolate up. I have long argued the Foveon's Mp rating is far under what the rating should be to match the Bayer. Meaning a Foveon 5Mp = 10Mp Bayer.

As long as you are pixel-peeping your image, you will likely never be happy.

-Mark
--
My pictures...
http://12.168.34.68/photos/default.asp
http://www.markwyman.com/photos/default.asp
 
Note: Not many cams can pull detail in these lighting conditions anyhow, which was overcast diffuse lighting. This experiment was to refute the muddiness claim, and represents a worst-case scenario for any cam with built-in NR.

A better sharpened crop at 100%, NR off, emphatic sharpen. Silkypix



The original JPEG resized. Crop from left of center.



--
My pictures...
http://12.168.34.68/photos/default.asp
http://www.markwyman.com/photos/default.asp
 
CMurdock said:
It's hard to judge those pictures because I don't know exactly how
they should look, but none of them looks particularly clean to me.
But thanks for posting them.

I give up. I'm going to start looking for another camera.
--
Click to expand...
I don't know that anything short of a $3000 camera is going to give you the sharpness that you're after - but even then, I ask WHY?

Correct me if I'm wrong, but in the other thread didn't you say that you don't print your photos and, instead, prefer to view them on screen at full size - not scaled to fit your screen. Then you go on to say that your monitor is set to 600 x 800. Is this correct? If it is, how do you get any enjoyment out of any photograph. If you'll pardon the cliche, You're never seeing the whole picture. The 100% picture on 800 x 600 would be equivilent to a 40" x 60" photo (yes, I made the size up).

You take a photo of the Golden Gate Bridge to capture the moment. You relive the moment by viewing the photo on screen or on paper. But when you view the photo at 100% on a 600 x 800 monitor, you're not reliving the moment - you're not even viewing the bridge. You're looking at one of the supports of the bridge. If you want to see another part of the bridge, you have to navigate within the picture. That's no fun.

If that's what you're doing as a matter of rule, then I would say that you're more interested the per pixal sharpness than you are in the final image itself. Which brings me back to my original statement - you need a 3000 camera.

You have several reviews, this one, Reichman, DCResource, Imaging-resource, etc. that say this one turns out pictures that are on par and in some cases, better, with entry level DSLRs. But you don't want to believe them because the picture isn't as sharp when it's viewed at some outrageous size.

Again, if I misunderstood what you said from the other thread - I apologize.
 
--Looking at your photos on your site, I see you like to take outdoor photos. You have some good ones. As for the one you've submitted here, this is a challenging photo to do anything with.....snow in the foreground, bright grey sky above. Not sure what you expect to get out of this??? Not the G1's fault, IMO.

All the best,
W.C.

http://www.flickr.com/photos/29320396@N05/show/
Digilux 2
FZ18
'Are you warm, are you real, Mona Lisa?
Or just a cold and lonely, lovely work of art?'
  • 'Mona Lisa', recorded by Nat King Cole - 1950
 
in order to get out better detail in the shadows you have two choices..

HDR or fill flash..

no PP can do miracles, but if you want you can try out noise ninja or some similar dedicated noise/sharpen softwares..

but anyway i suggest you use a powerful external flash..

hope this sheds some light (pun intended..)

--
tommaso ferrarese

L 1 0 / O L Y 2 5 - 2 . 8 p a n c a k e / L X 1 / O L Y M P U S F L 3 6 / l e i c a S F 2
4 D / n i k o n S B 2 6 / l e i c a C F
http://www.flickr.com/photos/tommasoferrarese
http://homepage.mac.com/tommasoferrarese/stage
http://homepage.mac.com/tommasoferrarese/IR
shadowgraphy
 
I know this particular circumstance is a very hard one for any camera. Bright foreground, bright sky, diffuse lighting, fine non-contrasty detail to try and reproduce. My Sony R1 used to do fairly well with these scenarios, it was an impressive camera, but the G1 is managing just as well, though it requires Raw processing to match up.

Call this a stress test. Thus why it is on my experiments page... I needed some way to reproduce what a gentleman was calling "muddiness" in the G1 output and a way it was obvious to show how it is mitigated when it appears.

Modern NR processing smears chroma since your eye cares less about chroma than it does about luminance. However these techniques have a lot of trouble in these circumstances (which are rare), and it is IMHO better to selectively NR an image in areas where noise is a problem than to blanket-NR an image.

Note: Not an image I would normally take unless there was a Jackolope in the trees I was trying to capture ;-).

--
My pictures...
http://12.168.34.68/photos/default.asp
http://www.markwyman.com/photos/default.asp
 
Yes, I know all that, but thank you for reminding me. My real argument is with the Bayer sensor design. However, some manufacturers -- especially Canon -- seem to be able to make Bayer images look good. Panasonic has failed, in my opinion.

I think I've decided to buy the Sigma DP1 or DP2. Here is a page from their DP1 site with sample images. The perfection of those images is astonishing. No Bayer camera comes close. I do wish the sensor size were a bit bigger than 4.6 MP, but eventually they will increase it.

http://www.sigma-dp1.com/sample-photo/

For all of you who don't know what a sharp image looks like, take a look at those pictures.
 
Oh, I switch back and forth between looking at the full image and the reduced image. In another post, I posted this link to the Sigma DP1 images. This is the kind of sharpness I want to see. If you look at those images at 100%, the beauty of the details is just breathtaking. There is very little beauty in the details of a Bayer image because all Bayer images are a bit blurry.

http://www.sigma-dp1.com/sample-photo/
 
I have downloaded the evaluation version of SilkyPix and fixed up a few of the G1 images myself, and was not satisfied.
 
And the recent review of the G1 on this very site shows it. Direct comparison with Canon and the G1 shows better definition of detail.
 
Man, you just don't get it. Try this with a full-sized G1 image:

1. Sharpen
2. Downsize to 2500 x 1875 (4.6Mp)
3. Weep at the incredible sharpness as your eye hurts.

The point is, the 12Mp Bayer even though the R,G,B pixels are spread out does still capture more detail in the Luminance channel (where it counts) than said Foveon sensor with the R,G,B pixels stacked. The numbering for resolution between the Foveon and the G1 is off, but image for image on the same scene at the same resolution, the G1 will produce the sharpest results.

--
My pictures...
http://12.168.34.68/photos/default.asp
http://www.markwyman.com/photos/default.asp
 
You must log in or register to reply here.

Similar threads

Guy Parsons
Exploring sharpening in Silkypix
Replies
15
Views
3K
Guy Parsons
Guy Parsons
luxborealis
DxO PureRAW 5 Review: Worth the Upgrade?
2
Replies
29
Views
16K
Jackpot101
J
lokatz
Noise reduction: Is it worth it? How much improvement?
2 3
Replies
44
Views
20K
Breakfastographer
B
M
Got curious and did a comparison of DXO and NX Noise Reduction.
Replies
7
Views
3K
monte12345
M
lokatz
Noise Reduction: Comparing DxO and LRC
2
Replies
31
Views
10K
Vince P
Vince P

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top