LX3 RAW/JPEG vs. LX1 vs. F20 comparison

Started Aug 15, 2008 | Discussions
carlos roncatti
carlos roncatti Senior Member • Posts: 2,668
Re: LX3 RAW/JPEG vs. LX1 vs. F20 comparison

A lot better.Theres noise as you say, but it has less digital look.Even the ACR has this kind of effect.I use NX for my Nikon and it produces this noise also, but yet its better than the digital look that ACR( as this Silkpix, but not so evident):
http://forums.dpreview.com/forums/readflat.asp?forum=1039&thread=28716690

The only worry now is to know if this lens problem is going to affect more than a few cameras.

oluv wrote:

i made a new test now and found the RL deconvolution sharpening
algorithm that is built in raw therapee to give the best results with
silkypix-files so far. have a look at this following example and tell
me what you think.

http://www.pbase.com/image/101708062/original

i think RL deconvolution gives the image a more natural look than
normal sharpening. the image gets a slight grain though, but it looks
more natural than the chaotic noise-mess that you get if you sharpen
the silkypix images in a usual way.

-- hide signature --

Carlos Roncatti Bomfim

Ehrik Veteran Member • Posts: 8,014
Re: Thanks...

KIWI Rik wrote:

The F20 sure does do a good job...

That's what I mean about people drawing the wrong conclusions.

It seems to do a good job because:

  • it's zoomed in closer to the detail

  • the F20 image is less enlarged

  • there are more photons falling on each F20 pixel (but there are fewer of them)

Maybe the F20 is doing a good job, but it's not easy to tell from the
comparison because of the unequal settings.

Just my two oere
Erik from Sweden

Kiwi Rik Veteran Member • Posts: 3,351
Re: Thanks...

Ehrik wrote:

KIWI Rik wrote:

The F20 sure does do a good job...

That's what I mean about people drawing the wrong conclusions.

It seems to do a good job because:

in this case

  • it's zoomed in closer to the detail

  • the F20 image is less enlarged

  • there are more photons falling on each F20 pixel (but there are

fewer of them)

Hi Erik,

Couldnt agree more....
--
Cheers

Rik

OP oluv Senior Member • Posts: 1,973
Re: LX3 RAW/JPEG vs. LX1 vs. F20 comparison

carlos roncatti wrote:

A lot better.Theres noise as you say, but it has less digital
look.Even the ACR has this kind of effect.I use NX for my Nikon and
it produces this noise also, but yet its better than the digital look
that ACR( as this Silkpix, but not so evident):

i looked at your images and this is exactly the same problem i have with ACR.

you know that adobe introduced this ugly noise-reduction effect not so long ago, i thnk with version 4.0

before ACR was rendering quite detailed images, tough a bit noisy. but then they decided to build in an internal noise reduction algorithm that you cannot turn off anymore. my LX1 pictures became really ugly, therefore i stopped using ACR.

i don't mind grain if it is looking natural. i hate noise-reduction artifacts or strange noise-patterns.

i am not sure what is wrong with my lens but i discover this problem now on more of my images, especially at tele there is some serious corner smearing. i hope this is only an exception, but my old LX1 has a similar problem, therefore i am afraid that these great "leica" lenses are not that great at all

 oluv's gear list:oluv's gear list
Panasonic Lumix DMC-GM1 Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G 20mm F1.7 ASPH Panasonic Leica Summilux DG 25mm F1.4 Olympus M.Zuiko Digital 45mm F1.8 +1 more
WoodWorks Senior Member • Posts: 1,422
ACR question

oluv wrote:

i looked at your images and this is exactly the same problem i have
with ACR.
you know that adobe introduced this ugly noise-reduction effect not
so long ago, i thnk with version 4.0
before ACR was rendering quite detailed images, tough a bit noisy.
but then they decided to build in an internal noise reduction
algorithm that you cannot turn off anymore. my LX1 pictures became
really ugly, therefore i stopped using ACR.

This may be a stupid question, but on my ACR (v.4.2) there are sliders on the Detail tab that allow me to turn both color and luminance noise reduction off, i.e., move both sliders to zero. Are you saying that ACR still introduces NR with both sliders set at zero?

David

Barry Fitzgerald Forum Pro • Posts: 29,888
Re: ACR question

WoodWorks wrote:

This may be a stupid question, but on my ACR (v.4.2) there are
sliders on the Detail tab that allow me to turn both color and
luminance noise reduction off, i.e., move both sliders to zero. Are
you saying that ACR still introduces NR with both sliders set at zero?

ACR does some baseline NR by default, its been toned down a lot, but still present.

This is on raw files I have used. LR 1.1 was a disaster for it..

Sliders to 0 some still goes on

-- hide signature --

I am not the 'Ghost Hunter', nor am I the Irish actor in the 'Quiet Man'

WoodWorks Senior Member • Posts: 1,422
Re: ACR question

Barry Fitzgerald wrote:

ACR does some baseline NR by default, its been toned down a lot, but
still present.

This is on raw files I have used. LR 1.1 was a disaster for it..

Sliders to 0 some still goes on

Would you mind pointing me at some source of info about what NR ACR does by default, if you happen to have it at hand? I'd be very interested in getting more educated about that.

David

GregGory Veteran Member • Posts: 4,252
Sorry, but these samples cause more confusion than good.

Thanks for your effort, there's no logical explanation why same sized different MP count sensors should be compared pixel by pixel. As Erik has already pointed out, most people don't check the exif, and are left with the wrong message.

oluv wrote:

today i tried to shoot some samples with all three cameras LX3, LX1
and Fuji F20 at the same time (not an easy task).
i shot jpeg/raw with both panasonics and jpeg with the F20:

http://www.pbase.com/oluv/lx3test

the conversion was done with silkypix and i tried to find settings
that were able to extract as many details from the raw-files as
possible. unfortunately the demosaicing algorithms in silkypix are
not very sophisticated so i am not entirely happy with the results. i
hope dcraw will support the LX3 soon, so that i can repeat the test.

my observation so far: LX3 shows less noise than the LX1, especially
in the shadows, but the difference is not really huge. also the LX1
seems to have a weaker anti-aliasing filter, because on the one hand
it is able to render more details than the LX3 on the other hand the
LX1 has quite jaggy diagonals that the LX3 doesn't show.
the F20 still seems to extract most details from most of the shots,
fuji's jpeg-quality is really superb. it is even more surprising as
the fuji cost me 150€ wheras the LX3 cost three times as much.

i also have the impression as if my LX3 lens was misaligned a bit. on
many photos i discovered some blurring on the left part of the image.
i will have to dig deeper into this and exchange it if necessary.

some high-iso shots will follow and i wand to make some crops, so
that you can compare the images more easily.

OP oluv Senior Member • Posts: 1,973
Re: Sorry, but these samples cause more confusion than good.

maybe i'll try to make a test with matched focal-lenghts too.

 oluv's gear list:oluv's gear list
Panasonic Lumix DMC-GM1 Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G 20mm F1.7 ASPH Panasonic Leica Summilux DG 25mm F1.4 Olympus M.Zuiko Digital 45mm F1.8 +1 more
MrClick Senior Member • Posts: 1,886
Re: Sorry, but these samples cause more confusion than good.

oluv wrote:

maybe i'll try to make a test with matched focal-lenghts too.

Great! Looking forward to this hungrily!! I bet the results will be yummier than the LX1/2 upon close inspection!!

I also look forward to Ehrik's opinion after you complete this test. He is a very practical guy with a lot of common-sense!!

Sad thing is common-sense is not that common after all... so shouldn't it be called uncommon sense?!!

-- hide signature --

Click

frobertv Contributing Member • Posts: 538
Re: LX3 RAW/JPEG vs. LX1 vs. F20 comparison

Hi, I got the exact same problem with blurred or out of focus areas, left or right with my Ricoh GRDII. I returned the camera but later thought it had to do with the DOF and the choosen plane of focus. Never got to figure out why it happened exactly but it is common practise with compacts that have a wide aperture lens like the lx3 and the GRDII. Maybe someone who knows more about the technical aspects can shed some light ?. I wonder for instance if the plane of focus is really a plane or a curve. Anyhow, its very easy to make two photos within a few seconds that have slightly different focal characteristics unless on a tripod.

Thanks for posting your very interesting work on the LX3, I will surely get a LX3 since I can see its is a wonderful compact in terms of IQ,
Francois

kernow Forum Pro • Posts: 12,441
LX1

KIWI Rik wrote:

Hi Oluv,

Thanks for doing these.....To me, they are the best set of examples
I've seen with regards to evaluation...I wish reviewers would do
something similar.....You have captured every element that is of
interest, in comparing issues that cause concern...

The LX3 output is much improved over the LX1, which has a similar
ugly output as my FZ50, everything is far to oversharpened and
grainy, especially foliage and people....The LX3 has a far more
natural look, albeit a little softer....But I prefer it to the older
VIII engine...

Regardless of sensor size I think Panny is doing a good job with this
new batch of cameras, compared to last year...

The F20 sure does do a good job...but it has other issues....and far
less features..
--
Cheers

Rik

Yes, but you can process the LX1 image to be softer with less detail if that is what you prefer. You cannot put the detail back that is missing.
--
Oll an gwella,
Jim

http://www.flickr.com/photos/50073525@N00/

[FX07]
[LX1]

[FZ30] * IS/L B-300 * IS/L B-Macro * Minolta No. 0, No. 1 & No. 2 * Sunpak 383 * Benbo Trekker

ratei New Member • Posts: 23
Re: ACR question

WoodWorks wrote:

Barry Fitzgerald wrote:

ACR does some baseline NR by default, its been toned down a lot, but
still present.

This is on raw files I have used. LR 1.1 was a disaster for it..

Sliders to 0 some still goes on

Would you mind pointing me at some source of info about what NR ACR
does by default, if you happen to have it at hand? I'd be very
interested in getting more educated about that.

David

For me I get a no NR with my RAW files also by setting Noise Reductions sliders (both Luminance/Color). This is with ACR 4.3 and above (and LR versions related to it). I also turn off the sharpening to zero and apply third-party NR first in Photoshop before using USM. I get better looking High ISO that way. Waiting for LX3 to be supported by Adobe before I conclude this camera is for me.

AdamT
AdamT Forum Pro • Posts: 58,597
Thats why I like the FZ30 ----> Comments

VERY Sharp sensor - looks like its widescreen compact Brother, the LX1 has that also !! the LX3 looks soft in comparison in the same converter (heavier AA) ..

the LX1 retains its sharpness at ISO400 but is a lot clearly noisier than the LX3 in RAW - the JPGs are a washout with the LX1 compared to the LX3 (the TZ5 blows it away too)

ISO800 is showing that Venus 4 is miles better than any previous Pan imaging engine by miles (wanna see it stretched to the limit and still working - try a TZ5 with its tiny 10Mp 1:1.3" CCD perform !!) but it's still no match for a 6Mp sensor which was designed with low noise in mind .. the LX3 RAW and cunning work in neat image is struggling to match the F20 JPG - downsize to 6Mp and the F20 will still probably be on top .. the only compact I know of which will outprint an F20/30/31 at ISO800 is the ancient Canon S30 in RAW and that's 3Mp on the same size sensor

-- hide signature --

Please ignore the Typos, I'm the world's worst Typist

 AdamT's gear list:AdamT's gear list
Canon PowerShot G1 X Panasonic Lumix DMC-GX85
IIIV Regular Member • Posts: 438
Re: LX3 RAW/JPEG vs. LX1 vs. F20 comparison

I downloaded this RAW image of yours...

http://www.pbase.com/oluv/image/101706377/large

and applied NR on it using Noiseware Pro and here's the result...

http://www.savefile.com/files/1732779

I say it looks pretty good.

OP oluv Senior Member • Posts: 1,973
Re: LX3 RAW/JPEG vs. LX1 vs. F20 comparison

wow noiseware does a perfect job on this. i tried it also with neatimage, but it struggled in this case.

did you use some particular settings for this file? could you please elaobrate a more on this. i am willing to try out noiseware if ithe results are that good.

would you mind if i put your result on my page too?

 oluv's gear list:oluv's gear list
Panasonic Lumix DMC-GM1 Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G 20mm F1.7 ASPH Panasonic Leica Summilux DG 25mm F1.4 Olympus M.Zuiko Digital 45mm F1.8 +1 more
Barry Fitzgerald Forum Pro • Posts: 29,888
Re: ACR question

WoodWorks wrote:

Would you mind pointing me at some source of info about what NR ACR
does by default, if you happen to have it at hand? I'd be very
interested in getting more educated about that.

The best way to see it is to compare raw from ACR and another 3rd party one, this is a free one here..

http://www.rawtherapee.com/

You can clearly see the difference, though it is not as bad as before.

Most noticable with high ISO raw files...notice the slight smoothing, esp on darker areas, like the jacket. Leave aside differences in colour/wb etc..

LR 2:

RT:

Luminance NR set to 0 on both

I am not the 'Ghost Hunter', nor am I the Irish actor in the 'Quiet Man'

GregGory Veteran Member • Posts: 4,252
Re: Sorry, but these samples cause more confusion than good.

oluv wrote:

maybe i'll try to make a test with matched focal-lenghts too.

That would be highly appreciated! Thanx in advance.

sorinx Contributing Member • Posts: 707
Re: Not strange

If at the pixel level LX3 is almost as good as F20 then it is a great camera. Considering that is has more pixels and faster lens.

It means that in real life is much better than F20.(considering that you will take a picture as good as you can; not take a larger frame and use only center 6mil mixels)

sorinx Contributing Member • Posts: 707
You are missing something important here

Adam-T wrote:

VERY Sharp sensor - looks like its widescreen compact Brother, the
LX1 has that also !! the LX3 looks soft in comparison in the same
converter (heavier AA) ..

the LX1 retains its sharpness at ISO400 but is a lot clearly noisier
than the LX3 in RAW - the JPGs are a washout with the LX1 compared to
the LX3 (the TZ5 blows it away too)

ISO800 is showing that Venus 4 is miles better than any previous Pan
imaging engine by miles (wanna see it stretched to the limit and
still working - try a TZ5 with its tiny 10Mp 1:1.3" CCD perform !!)
but it's still no match for a 6Mp sensor which was designed with low
noise in mind .. the LX3 RAW and cunning work in neat image is
struggling to match the F20 JPG - downsize to 6Mp and the F20 will
still probably be on top .. the only compact I know of which will
outprint an F20/30/31 at ISO800 is the ancient Canon S30 in RAW and
that's 3Mp on the same size sensor

The focal length was different for LX3, so that only the center 6MP were used to compare with 6MP camera. THis is like dooing digital zoom.

I am sure that if we zoom LX3 to the same focal length as F20 it will be at least as good.

Keyboard shortcuts:
FForum MMy threads