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!

GRDII vs. G9 (vs. LX2) - a couple of questions

....
Use your wrist strap. I've already got 3 dents from a 4' drop onto concrete :p
Good advice; also the wrist strap helps a lot in holding the camera. I´ve put a sliding "bead" on the wrist strap, so I can make tighter over the wrist; now it can dangle safely when using the hand for other things.

But WHAT did they think of when designing that gorgeous leather case? :( There simply isn´t enough room to tuck the wrist strap inside; either you´ll have to remove it, or let it dangle outside and have it snagging on doorknobs and things....
 
S

simon_t

Guest
Thank you all very much for the additional comments. It seems that flash-wise, the GRDII could fulfill my needs. With my current compact (Canon S45), I have found red eyes less of a problem than I had expected, because I often shoot at close range. I am not sure whether the GRDII would be much worse. Of course, an external flash unit will do a better job overall.

Mirroring helgipelgi's comment, and showing that camera selection can be really hard, I am now gravitating towards the camera that I had scrapped off my list before: the GX100. It's only a little bigger and a little noisier than the GRDII, the lens is not much worse (or even better), it zooms, has IS, and it's more than 100 euro cheaper.

As for the flash: I think I may just use my Olympus FL-20 in auto mode in combination with manual aperture/shutter speed settings. When I am not carrying the external flash, I can still rely on the internal flash to give ok results (albeit with less control).

The Dlux3/LX2 looks very good as well, but the smoothing of RAW files would really interfere with my style of photography. The G9 is great specification-wise (except for the narrow lens), but it simply doesn't really appeal to me...

Cheers,
Simon
 
...
The Dlux3/LX2 looks very good as well, but the smoothing of RAW files would really interfere with my style of photography....
As far as I´ve seen so far, this raw smoothing only shows up in 800 and 1600 ISO raw´s. Something dramatic seems to happen when going up from 400 to 800; the image "falls apart", with pixelations, blob-like details, and discernible banding. It seems the highest "true" sensitivity is 400, with the higher settings being only manipulations of underexposed 400 captures.

As for 1600, just forget it´s even there...:cussing: And use 800 only when there really is no alternative (sadly, the otherwise well-performing "I-ISO" includes 800, and often goes there). But 400 seems cleaner than on my Digilux 2 (still talking raw, of course).
 
S

simon_t

Guest
As far as I´ve seen so far, this raw smoothing only shows up in 800 and 1600 ISO raw´s.
To be frank, I'm not really expecting to use any small sensor camera above ISO 400+ with good results. If I can, it's a nice bonus. So it would be fine with me if the smoothing started at ISO800.

However, I stumbled on a set of LX2 RAW files that show this effect very clearly at ISO 100 - so they should represent the best the camera can produce. They can be found at http://www.esnips.com/web/viztygerphotos/
The 'offenders' are P1000460.RAW, P1070119.RAW and P1030834_DNG.dng. If you look closely at the shadows, or try to boost the exposure of dark areas, it is very visible (to me, at least). Boosting the exposure is something I expect to do quite often when shooting high contrast scenes - and I greatly prefer noise to this artificial smoothing effect. I guess that whether it is an issue or not greatly depends on your shooting/processing style.
 
...a set of LX2 RAW files that show this effect very clearly at ISO 100 - so they should represent the best the camera can produce. They can be found at http://www.esnips.com/web/viztygerphotos/
The 'offenders' are P1000460.RAW, P1070119.RAW and P1030834_DNG.dng. If you look closely at the shadows, or try to boost the exposure of dark areas, it is very visible (to me, at least)...
Well, indeed they wouldn´t have made Ansel Adams happy....:cool:

I´ve downloaded, and fiddled with, the dng file you mention (it was the smallest one), and the dark areas are indeed a mess. Still, I´m a bit uncertain what´s really going on here.

After all, it´s a scene that would be a challenge even with compensating development of a b/w medium format negative (dark rock in shadow and foaming water in sunlight). So it´s conceivable that the darkest areas simply fall so close to the treshold that they´ve become posterized.

I´m not familiar with exactly how the transform from linear raw space to the rendering we look at is done (no way to look DIRECTLY at a raw; it has to be rendered in some way), and from a practical point of view it doesn´t really matter; the result matters, and here the result isn´t pretty... But as I said, I´m uncertain as to why. I´m not convinced that there really has to be any smoothing involved to give the results we see in this case.

I haven´t come across any scene even remotely approaching this contrast range since I got my Dlux 3 (Stocholm winter right now is a grey mess..), and I´ve not seen anything like these artefacts in any of my low ISO shots. Also, the smoothing in 800 and 1600 shots isn´t confined to just the darkest parts of the picture; they go all the way up to at least Zone V.
 

Maggie O

Active member
I'm not seeing any "smearing" in P1000460.RAW at all. I'll attach a jpeg of my developed version. To me it looks sharp, and I'm kind of amazed at how well the lens resolves all those individual bricks. You can't really see them at 900xX, but at full size, no problemo.

Also, the Leicasonic doesn't produce .DNG files, so either the file has the wrong EXIF data or there was a SNAFU in converting it to .DNG.
 

Maggie O

Active member
Per, if you're seeing "smoothing" in your RAW files I can't help but wonder if:

1)You've got a faulty camera

2)Maybe you're actually looking at the JPEGs that are produced at the same time

I only say this because, I've never seen any evidence of smearing in my RAW files, but have seen plenty of it in the JPEGS.
 
S

simon_t

Guest
Per, the conversion from RAW sensor data to the output should not lead to posterization (and it would look slightly different). What I would expect, given the small sensor size, is noise in the shadows, even at ISO100. The fact that you don't get noise, but slightly blotched, patterned colors, seems to indicate that there is some smoothing going on.

Maggie, when I converted that same image using Lightroom with a similar shadow boost, the smoothing was very evident on the grey bricks on the lower levels of the building on the left (not visible on your web sized image). Which converter did you use?

Also, a conversion to DNG does not usually affect the RAW data (unless you force it to render a linear file), and it certainly should not introduce such smoothing.

The effect certainly exists in the RAW file of this particular camera, and it sure looks like deliberate processing. But, as I said, you may not notice it, depending on your processing preferences.

Simon
 
Last edited:

Maggie O

Active member
I used iPhoto, which seems to do much better than ACR with Leica RAW files.

Maybe it's the converter and not the camera that is doing the smoothing?
 

nostatic

New member
the lore is that there is a difference between the LX2 and D-Lux3 wrt "processing." Perhaps this is one example of the difference?
 

Maggie O

Active member
Well, I went looking through my RAW files to see if I could find anything like the smearing described and I found a few files that had something like what was described and they all had the same things in common: ISO800, a shutter speed lower than 1/25 of a second, an animate subject (usually my cat) and deep shadows.

My hypothesis: what we're seeing is a fatal cocktail of noise, image stabilization, camera movement, subject movement and more noise. Add those together with any NR from the PP (and just about all of them add some) and you get a blurry, smeared mess in the shadows. So, I don't think it's in-camera NR, but the OIS and other factors.
 

nostatic

New member
so what you're telling us is there's no free lunch? Don't sugar-coat it...we can take it.

mmm, sugar...
 

Maggie O

Active member
You want crazy high-ISO, low noise files that you can extract CIA-level shadow detail from?

Get a C1 MF back or a Canon DSLR.
 

nostatic

New member
no, I want that software program that takes a sat photo and has the "magic enhance" button so I can read the license plate of the bad guy's car.

And I'd like a nice tequila. No lime. :)
 

Maggie O

Active member
no, I want that software program that takes a sat photo and has the "magic enhance" button so I can read the license plate of the bad guy's car.

And I'd like a nice tequila. No lime. :)
I'm not currently at liberty to discuss that, sir. Now I'll have to ask you to take off you shoes...
 
S

simon_t

Guest
Maggie, I'm quite certain that ACR doesn't cause the smoothing, because it results in slightly colored 'blobs' that are 10+ pixels in size. As far as I know, ACR/Lightroom does nothing on these length scales (except for fill light and clarity). See the following crop for what I'm seeing in that file after boosting the shadows quite a bit [it is also visible with less pushing, but I'm trying to make this work on every monitor out there].


You want crazy high-ISO, low noise files that you can extract CIA-level shadow detail from? Get a C1 MF back or a Canon DSLR.
Of course, you'll need a larger sensor than this. That's why I'm only expressing my surprise at artifacts that pop up at base ISO and the fact that instead of noise (which I don't mind) I'm seeing a processed smear (see image above).

the lore is that there is a difference between the LX2 and D-Lux3 wrt "processing." Perhaps this is one example of the difference?
That may indeed be possible. Although this will be hard to determine if we can't agree on whether it's there in a specific file like the one above :) I'll see if I can take a look using another RAW converter as well.

Cheers,
Simon
 
Per, if you're seeing "smoothing" in your RAW files I can't help but wonder if:

1)You've got a faulty camera

2)Maybe you're actually looking at the JPEGs that are produced at the same time

I only say this because, I've never seen any evidence of smearing in my RAW files, but have seen plenty of it in the JPEGS.
1) Possibly; I´ve only had it for less than a week. Nothing else seems amiss, however.

2) Nope. I know what I´m doing...:angel:

Per, the conversion from RAW sensor data to the output should not lead to posterization (and it would look slightly different). What I would expect, given the small sensor size, is noise in the shadows, even at ISO100. The fact that you don't get noise, but slightly blotched, patterned colors, seems to indicate that there is some smoothing going on....
Simon
Well, as I said, I don´t feel up to a detailed discussion of raw conversion algorithms. Still, "slightly blotched, patterned colors" sounds (looks) suspiciously like a description of noise in a signal that has only a few distinct values (like very close to the black point treshold).

Well, I went looking through my RAW files to see if I could find anything like the smearing described and I found a few files that had something like what was described and they all had the same things in common: ISO800, a shutter speed lower than 1/25 of a second, an animate subject (usually my cat) and deep shadows.
Ha! Now we´re getting somewhere! This is exactly the kind of situations where I´ve seen this effect in 800 and 1600 settings (in my case a shaggy dog and a somewhat less hirsute wife, but I doubt the algorithms can distinguish between species...).

My hypothesis: what we're seeing is a fatal cocktail of noise, image stabilization, camera movement, subject movement and more noise. Add those together with any NR from the PP (and just about all of them add some) and you get a blurry, smeared mess in the shadows. So, I don't think it's in-camera NR, but the OIS and other factors.
Quite possible; I don´t have enough knowledge to challenge that hypothesis. In any case, the images in question look terrible. I´ve realized, and accepted that I´ve bought a good 100 - 400 ISO camera that can do 800 in a pinch, and that has a useless 1600 setting, I´ll never even touch again (after all, one can do things to an underexposed 400 file in the converter, and might do it better than the camera).

Finally, I´m intrigued to hear you get good conversion results from the Apple OS engine; I´ve only seen it at work in Preview, and there no editing is possible. iPhoto isn´t enough for my workflow, but I might have a look at Intaglio.

Anyway, within a few hours I´ll have to disconnect and be off for the mountains for a week. So if I don´t answer again, I´m not sulking..:rolleyes: I look forward to see what this thread has come to after a week; I feel I´ve learnt a lot from it already. Thanx all! :salute:
 
S

simon_t

Guest
Well, as I said, I don´t feel up to a detailed discussion of raw conversion algorithms. Still, "slightly blotched, patterned colors" sounds (looks) suspiciously like a description of noise in a signal that has only a few distinct values (like very close to the black point treshold).
Technically, you may be right. However, that would mean that Panasonic is using a bit depth that limits the dynamic range at base ISO (10 bits?), because the signal down to the posterization level is quite clean. Plus, there is some additional interference/spatial correlations leading to the stripy/blob-like character.

My conclusion is that the dynamic range is limited by this effect (whatever it is) instead of by 'simple' noise. Just to be clear, I don't intend to 'bash' this camera, but there seem to be trade-offs involved in any model, and this appears to be one of them.

Per, enjoy the mountains!

Simon
 

helenhill

Senior Member
My hypothesis: what we're seeing is a fatal cocktail of noise, image stabilization, camera movement, subject movement and more noise. Add those together with any NR from the PP (and just about all of them add some) and you get a blurry, smeared mess in the shadows. So, I don't think it's in-camera NR, but the OIS and other factors.[/QUOTE]

I LOOOVE your FCH : Fatal Cocktail Hypothesis
LOL! Cheers ...:grin:
 
Top