Images from Panasonic L-Mount cameras

[Chaemono]

Advertisement (gone after registration)

I took thighslapper's Hi Res image and upsized it to 294 MPx. 😁 The PSD file that ON1 creates to roundtrip to LR is 4.7 GB. I thought my Mac Book was going to die. A JPEG export results in 96 MB. 😁 Then I cropped. 😁

 

Norm Res - crop from 47 MPx

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

Hi Res - crop from 187 MPx 😁

Hi Res - crop from upsized 294 MPx 😂

Edited March 17, 2019 by Chaemono

[Chaemono]

And this is the original size from the Hi Res one, BTW (I hope thighslapper doesn't mind).

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[helged]

15 minutes ago, Chaemono said:

I took thighslapper's Hi Res image and upsized it to 294 MPx. 😁 The PSD file that ON1 creates to roundtrip to LR is 4.7 GB. I thought my Mac Book was going to die. A JPEG export results in 96 MB. 😁 Then I cropped. 😁

...

Hi Res - crop from upsized 294 MPx 😂

 

Thanks. Based on the jpgs - the up-sized image appears sightly more detailed compared to the Hi-Res image, whereas the latter is a little more pleasant to look at for my eyes (in the meaning a little less digital than the up-sized image).

An open question, for curiosity: Hi-Res takes 8 images, with the sensor (slightly) shifted, generating a biggie file with improved detail. What about taking 8 identical images, loading these into PS as smart objects, and then taking (e.g.) the mean of these; Layer -> Smart Objects -> Stack Mode -> Mean? This would certainly improve the signal-to-noise ratio that, in isolation, should lead to improved rendering of details. Would this method also give an end result comparable with the Hi-Res and up-sized images shown above?

[verwackelt]

 I don´t think so. As i understand, the HiRes Mode is a little bit like making a pano with a technical camera where you can move the Sensor or like a Normal camera with a shiftlens and the lens is mounted on the tripod but the camera is shifted. 
If you shoot 8 identical images you will not win resolution because all sensorpixel are exposed the same at each shot.
May be you can reduce noise by stacking 8 identical shots with additional software i do not know if PS is able to do so...

[hillavoider]

So SL lenses work fine, auto focus is good? 

[scott kirkpatrick]

I've messed about a bit with Olympus multi-shot hi-res.  In the vision and image science literature, this is called super-resolution.  Imagine that you can displace the sensor by one pixel spacing in the x (horizontal) and then y (vertical) directions, then go back a step in x and finally back a step in y, taking a picture each time.  Then in each pixel position you now have two G, one R and one B pixels' worth of data.  It's as if you were just shifting the Bayer filter, leaving the pixels where they started. That's how Hasselblad has done this for some time.  Now there is no loss of resolution due to the de-mosaicing of the Bayer filters.  For the next four shots of eight total, first displace the sensor by half a pixel spacing in both the x and y directions, and repeat the four step process.  This gives four more pixels' worth of data overlapping the previous four, as if you had pixels with a spacing 1/(sqrt(2)) of the original spacing, but they are "smeared" pixels that overlap each other.  Finally some signal processing tricks (usually something that would be patented, since there are lots of ways to do this) assigns all the information to the larger pseudo-array of smaller virtual pixels  that make up the hi-res JPEG output.  It sounds like a lot of work, but fast algorithms exist for making JPEGs out of Bayer data, and these tricks just use them in a different way.

[thighslapper]

Advertisement (gone after registration)

6 hours ago, hillavoider said:

So SL lenses work fine, auto focus is good? 

Yes ..... but there are major issues with metering etc. using high ND filters ...... and even 3/6 stops at lower light levels that may make this camera unusable for landscape use.

I will discuss this with Panasonic this afternoon. 

If my concerns are an inherent feature of the camera I doubt I will be keeping it. 

[verwackelt]

But does other cameras meter with High ND filters( 6x or 10x) correct?
I did not use them much, but i did metering without them, mount them after metering and do time compensating depending on ND-filter factor…

[mdemeyer]

Very curious to see what you learn about metering.  My brain can’t fathom why something that measures light with the sensor should know the difference between a less brightly lit scene, a reduction in light from stopping down, or a neutral density filter.  All have exactly the same effect on the light intensity hitting the sensor. 

[verwackelt]

May be a 6x 10x ND filter is reducing the light so much, that the reduced light is out of the sensitivity range of the build in Lightmeter...

Edit:
meaning it is important to know how sensitive your Lightmeter is. That is written in technical specs. 
E.g. https://www.calumetphoto.de/product/Calumet-DFM-3-Belichtungsmesser/CALCF0300
has a sensitivity from EV -2,5 to +18. So with a strong ND Filter in blue hour you may be out of range with your build in meter and need to measure without the filter and ad the compensation factor at exposure.
The Range of the S1R is EV 0-18...

Edited March 18, 2019 by verwackelt

[ron777]

There was no mention of the camera settings, in re the ND filters.  If the camera had been set to auto ISO with an inadequate upper limit, might that not cause similar behavior?

[thighslapper]

1 hour ago, verwackelt said:

May be a 6x 10x ND filter is reducing the light so much, that the reduced light is out of the sensitivity range of the build in Lightmeter...

Edit:
meaning it is important to know how sensitive your Lightmeter is. That is written in technical specs. 
E.g. https://www.calumetphoto.de/product/Calumet-DFM-3-Belichtungsmesser/CALCF0300
has a sensitivity from EV -2,5 to +18. So with a strong ND Filter in blue hour you may be out of range with your build in meter and need to measure without the filter and ad the compensation factor at exposure.
The Range of the S1R is EV 0-18...

I suspect this is the issue ...... but it does raise the question of how the SL manages to produce a perfect screen image and meter correctly with a 10 stop ND at ISO 50 in normal lighting and continues to function accurately in even lower light conditions.  

The S1R in the same light levels at ISO 100 shows a dimmed screen even with a 6 stop ND and underexposes by 2.5 stops. I'm not that fussed about the display but the inaccuracy of exposure metering ..... in conditions where I would expect things to work fine.

I fear that my expectations after extensive use of the SL with long exposures, low light and high value ND's are not going to be matched by Panasonics offering. The only question is whether all the other advantages outweigh the 5% of occasions where it will irritate me because it can't do what the SL does .....

[verwackelt]

OK, when showing a good image with that filter i think you are right. it should be within the range and should meter right values.
I am curious if that is only the behavior of your camera or if others will report similar things.
And what Panasonic will say…

[peterv]

So let me get this straight, your lightmeter at 100 ISO sees let’s say sunny sixteen so you get 1/500 and f8,0 and than you say you hang a 6 stop ND (sorry, I have to ask, not an ND6?) in front of the lens. Now keeping the aperture at f8,0 it should say 1/8.

Are you saying

- it does that, but the display looks to dark

- it keeps saying 1/500 and F8,0 and the diplay gets dark

Sorry for one more dumb question, did you set the display to WYSIWYG?

 

[Jon Warwick]

Many thanks thighslapper for your efforts here. It's much appreciated.

The hi res is pretty remarkable ....if i resample upwards your 336normres to the native size of the 335hires (ie, to approx 56"x37"), i see an enormous difference, not only in fine detail but also due to a complete lack of bayer filter "blur". The latter makes the 335hires image look far more realistic and less digital, to my eyes, at that large image size.

I can only imagine what the hi res mode and SL prime could achieve if pointing at a distant landscape!!

Edited March 18, 2019 by Jon Warwick

[W124]

Moin,

 

S1 with Leica Summilux M 35mm Asph FLE and Leica M-Adapter L

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Glühdraht by hen. ry, auf Flickr

Edited March 18, 2019 by W124

[Stuart Richardson]

Hi W124 (and others), 

Any thoughts on how the M lenses fare on the S1R? The photo above is attractive, but does not say much about the performance across the field at longer distances etc...is there vignetting or blur towards the edges? I am quite curious about that....as well as to hear if the S lenses work.

[W124]

Another one with S1 and APO-Summicron-SL 1:2/75 mm ASPH.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Schuh by hen. ry, auf Flickr

[W124]

vor 10 Minuten schrieb Stuart Richardson:

Hi W124 (and others), 

Any thoughts on how the M lenses fare on the S1R? The photo above is attractive, but does not say much about the performance across the field at longer distances etc...is there vignetting or blur towards the edges? I am quite curious about that....as well as to hear if the S lenses work.

I had it on a S1 without "R", firmware 0.xx and jpg only, so I think for a good comparison we should wait for the final version. 

In general it looked like my A7sII IR converted, but not as good as on a M10!

Edited March 18, 2019 by W124

[Advertisement]

Hallo,

Look here: Leica L-Mount

[Stuart Richardson]

I guess that is better than the stock A7S? I do not know much about the conversion process. Better than full glass but worse than on an M?