Panasonic Lumix S1/S1R studio scenes with high-res mode published

Correct. We tried the Lumix S Pro 50mm 1.4 as well, with little noticeable difference.

May 22, 2019permalink

Morpho Hunter

I wonder ..... would you get the same excellent hi res image quality results obtained by the SR1 with the pixel shift technology of Pentax and Olympus, combined with Topaz Gigapixel AI enlargement software? I suspect that there would be very little difference - would anyone out there like to test this (using the free trial software download)?

Apr 29, 2019permalink

With all the Pentax K-1 pixel-shift discussion, i can't help but wonder if the successor to the 645Z will offer pixel-shift. Imagine an 80MP-ish (assuming they up the MP a bit) 645Z2 that pixel-shifts...

Apr 27, 2019permalink

Andrey K

That would be like H6D-400c MS (100MP with pixel shift)

May 11, 2019permalink

Haha, but i like the sub-$10,000 body price of Pentaxes vs. the $48,000 Blad body price. 😉

May 11, 2019permalink

The S1R's pixel-shift results are absolutely amazing! Take a look at how good ISO 3200 looks with pixel-shift, particularly vs. other current FF cameras that can't pixel-shift. Stunning! And of course, at lower/base ISO the results are even more astonishing!

Granted, this is really excessive resolution for almost all situations. However, as this is clearly intended as a landscape camera (not saying it's not fully capable of doing other things very well), these pixel-shifted images could make gigantic prints (measured in feet, not inches) with very fine detail that have very little grain/noise and no moire. Wow!

Now if I could only put my Nikkors on it with full functionality... :)

Apr 26, 2019*permalink

aionios

If a Canon to Panasonic mount exists, id be curious to see how the Zeiss Otus 55mm goes with the pixel-shift modes.

Apr 26, 2019permalink

Zerixos

I think that the Leica glass will be just as good, so you'll end up with great, extremely sharp images.

Apr 29, 2019permalink

Paul S Churcher

How good are the results from all these cameras, and to think how far the industry has come in only 20 years of Digital Photography! Photographers are the true winners....just by the choice of fear available!

Apr 25, 2019permalink

Paul S Churcher

That should have read.....gear available 😊

Apr 25, 2019permalink

Ralf B

well, 'fear' fits many DPR posters.

Apr 26, 2019permalink

Wow, VERY nice sensor! Maybe now Nikon can buy sensors from Panasonic instead, lol.

Fantastically-clean noise profile at ISO 6400 (doesn't pull magenta). Skin tone at ISO 6400 pulls a little yellow-green whereas Nikon and Sony pull a little red.

I was afraid to see this test. But now i have to say that I'm a fan!

Apr 25, 2019permalink

bokehmon22

I'm surprised I'm not seeing Sony is better than this or that crowd in here and Panasonic getting love. Am I the wrong forum? lol

Apr 25, 2019permalink

Naveed Akhtar

I think Sony is producing very different cameras, both Pany and Sony are doing great. There is more of Sony vs Nikon FF debate!

Apr 25, 2019permalink

jrp

The S1R increases resolution by pixel shifting. The Sony is supposed to improve colour fidelity / reduce artefacts. In theory the Sony approach should be better. (Who really needs 187 mpx ? ) Not sure that I’m seeing it here though.

Apr 24, 2019permalink

onlyfreeman

Pentax and Sony have the same approach, but Panasonic and Olympus take it further. In theory I believe the Panasonic / Olympus method is actually better.

Apr 25, 2019permalink

Jorginho

The S1R does 3 shifts per colour. 4 time shifts for resolution. So the colour fidelity is as good as it gets, no moire possible either and the res is 187 MP.

Apr 25, 2019*permalink

https://www.dpreview.com/forums/post/62592160

The S1R's pixel-shift results are absolutely amazing! Take a look at how good ISO 3200 looks with pixel-shift, particularly vs. other current FF cameras that can't pixel-shift. Stunning!

Apr 26, 2019permalink

knickerhawk

See link below for a close comparison of the S1R HiRes vs. the Phase One 100mp with both converted in ACR. It's important to note that (for whatever unspecified reason), the DPR studio comparison is using a Phase One rendering that was done in Capture One, which makes it somewhat of an apples-vs-oranges comparison when it comes to looking at micro detail. Better to use the same converter as I did here:

Apr 24, 2019permalink

https://www.dpreview.com/forums/post/62596240

Similar comparison, including also the Sony A7Riii... It seems to me that the Phase One XF 100MP is better, but the Panasonic S1R's results can be considered impressive, especially when compared to the A7Riii. No sharpening applied at all:

Apr 25, 2019permalink

@knickerhawk:
Awesome work/comparison. Thank you very much! 👍

Apr 25, 2019permalink

The only down side I can see is the default color is pretty blah, but this early in the game it's hard to tell if that is the camera or Adobe's interpretation of the sensor. Playing with a raw file, the profiles closest to realty seem to be either Adobe Landscape or Camera Landscape (pretty much the same gamut but slightly different gamma curves).

Apr 24, 2019permalink

MrBrightSide

Like every other camera in comparison to the IQ180, it has real trouble with greens. And also like every other camera it has trouble with the goldy-yellow on the target with the concentric circles under the hank of hair.
Both flesh-tone patches on the Color Checker look good though.

Apr 24, 2019permalink

Andrey K

The fact that it gets compared to Phase One speaks for itself. Remember, this is 35mm mirrorless, not medium format.

May 11, 2019permalink

MrBrightSide

Agreed. Panasonic has definitely upstaged Sony and Nikon and encroached into MF territory. . But soon the medium format companies will introduce some new tech that produces even more unbelievably beautiful and accurate images. That will force the full frame makers to dig deep to come up with a response—it is a grand time to be a photographer.

May 12, 2019permalink

JMD-70

Finally! Thank you DpR!

Apr 24, 2019permalink

Bassman2003

Both Panasonic's show a lot of detail. Very nice job Panasonic. The noise on the S1R is not any better than a 5DMKIII or a D810, but it is a larger pixel count. So still a good job. The S1 is better than both of those camera while keeping/beating detail. So very good job Panasonic.

Detail on both cameras is very high.

The colors are different from Canon and more similar to Sony in many respects.

To the DPReview team, can you tell us which camera is closest to reality when it comes to the studio scene color charts? I know it depends upon the RAW converter for RAW files but just from the published results can you make an assesment for us?

Apr 24, 2019permalink

(unknown member)

Sony and Panasonic have always used colors that are more "accurate". In some surveys that use color charts Sony and Panasonic comes out on top because colors are "more accurate".
But decades ago Kodak scientists learned in photos people do not want accurate. And centuries before that people learned what they saw looked more pleasing during the "golden hour" when the sun was shining a reddish tint. So Kodak, and later Fuji and Canon worked on color science to produce, not the most accurate, but the most pleasing colors.

Apr 24, 2019*permalink

SteveAnderson

Canon has had the most accurate colors for 10yrs or so. Also the strongest CFA for tonality.

Canon around 3.5 delta error, Sony closer to 5.0 and Panasonic 4.5 delta error.

There is a site that tests this on all cameras for the last 20yrs. Easy to look up.

5Dm2 was 3.94 in 2008.
GH1 was 6.59 in 2009.

They push reds equally, with Canon more accurate everywhere else.

I dont care which is better, just want to point that out.

Apr 25, 2019*permalink

gbinoz

I guess that US$5000 APO-Summicron-SL 90mm f/2 lens lens helps a bit.

Apr 24, 2019permalink

pannumon

US$2300 Panasonic 50mm f/1.4 seems to be decent as well:
https://www.dpreview.com/articles/3528119869/panasonic-lumix-s1-s1r-studio-scenes-with-high-res-mode-published?comment=2697094987

Apr 24, 2019permalink

I didn't shoot the whole scene with the Panasonic 50mm F1.4, but I did a daylight base ISO high-res shot - I will be sure to put a side-by-side in the final review.

Apr 24, 2019permalink

Powerdoc

The high resolution mode is impressive, but it's not better than the phase one (quite normal if we consider the cost of this later ....)
The texture of the white part of the card, is visible with the Phase one, not with the Pana.
BTW for A2 print, there will be absolutely no difference between normal and high resolution file

Apr 24, 2019*permalink

s1oth1ovechunk

It is clearly better than the phase one in the Panasonic's lack of color moire.

Apr 24, 2019permalink

evan ts

I downloaded the S1R Hi-Res RAW and processed by Silkypix. It looks more detailed than the ACR.
http://i.imgur.com/g1CQiEt.jpeg

There is same situation in the m43 Hi-Res case. G9 OOC JPEG looks better than ACR.
http://bit.ly/2ZvElUj

Apr 24, 2019permalink

It looks like Silkypix just has a bit more macro contrast added and some light sharpening applied in the demosaicing process, so I suspect the difference in detail is merely perceptual. I agree that it looks better for the most part.

Regardless, it reinforces the idea that multi-shot Hi-Res RAW files really are raw, and need to be edited before output to achieve good results. They should be seen as a starting place, not an end-point.

Apr 24, 2019permalink

GinoSVK

High resolution mode looks exclusive, must be beast for product photography.

Apr 24, 2019permalink

zakaria

I wonder how 1500$ camera beats all of these expensive mirrorless cameras in terms of resolution and details. I am referring to pentax k1.

Apr 24, 2019permalink

s1oth1ovechunk

I think the g9 has the best hi res mode for a cheaper camera

Apr 24, 2019permalink

Pentax started producing their first FF DSLR when they should have presented a FF mirrorless camera. I don't think that Canon or Nikon will present any new FF DSLR.

Apr 24, 2019permalink

walgarch

I'm really impressed by how well the K1 is holding up in this test. It also looks like the S1R is doing really well in the normal shots too, better than the high res mode. Which is definitely a plus if were getting great detail even in the normal mode!

Apr 24, 2019permalink

https://www.dropbox.com/s/hc8843x5gunklju/Lumix%20High%20Resolution%20test.zip?dl=0

There has been a lot flack in the DPR comments lately about camera manufacturers being behind the phone guys in computational photography. The high res image is a sample of what can be done with a good imager, and excellent lens, and serious computation. Pretty impressive, and maybe some of them aren't so far behind after all.

Apr 24, 2019permalink

WolfpackN64

Computational photography is more about compensation for the tiny sensors in smartphones. Of course computation will be more and more important in DSLR's and Mirrorless cameras, but I'm guessing it won't be in the same way as smartphones do it (as these can decrease image quality depending on the implementation).

Apr 24, 2019permalink

brycesteiner

This is the first time I have truly seen the texture detail in the playing cards!

Apr 24, 2019permalink

Deanaaargh

Well, they all do a much better job resolving the peacock feathers than the phase one.
so borders should take note

(this was a joke)

Apr 24, 2019permalink

Deanaaargh

*birders*

but seriously look at that nonexistent rendering.

Apr 24, 2019*permalink

Hartmonstr

I own both the Hasselblad H6D-100c and the S1R, and made a test using the 120mm Hasselblad and the 24-105mm Lumix, both at f/8.

The resolution is impressive on both cameras. Of course, the 120mm at f/8 is amazingly sharp right out to the edges, while the Lumix kit lens is slightly soft at the edges, but nothing except pixel peeping would discover.

Can't wait to get the Sigma Art for the real test.

Download and view the full size test files here:

Apr 24, 2019permalink

aionios

Any chance you can get the Panasonic 50mm 1.4 to test with?

Apr 24, 2019permalink

Sure the Panasonic 50mm f/1.4 must be a damned good lens, better than the very good Sigma...

Apr 24, 2019permalink

aionios

Turns out they DID use the Panasonic 50mm f1.4 for this test! One of the staff commented in the comments advising this.

Apr 24, 2019permalink

@Hartmonstr Thanks for your post, but wouldn't the 120mm Hasselblad be at at f/10 for the comparison to be meaningful. I agree re: the pixel peeping issue.

Apr 24, 2019permalink

Hartmonstr

David610 agree about f/10 if you're comparing depth of field and bokeh, but it doesn't apply here to a completely flat subject. I kept the aperture f/8 for both lenses because maximum sharpness is generally achieved two stops down from wide open.

Apr 24, 2019permalink

https://www.dpreview.com/galleries/0119553924/photos/3880957/compare

Here is a comparison with standardized resolution between the Panasonic DC-S1R, the Phase One XF 100MP and the Sony A7Riii. I did this INCREASING the resolution, not reducing it and wiping out details... No sharpening was applied...

The S1R in High-res mode gives excellent results, much better than the A7Riii... The Phase One XF 100MP is a little bit sharper...

It seems to me that the way @DPReview does the studio scene shots is favouring the cameras with four thirds sensors, as the photographed area is smaller, so more detail is generated. To avoid this, the photographed area (in square centimeters or square inches) should always be the same.

Apr 24, 2019*permalink

zodiacfml

thanks. some people simply declared it as soft as they're looking at 100% enlargement of the Panny. I mentioned that the DPR comparison tool will not be able to reveal the S1R's capability compared to the MF Phase one camera.

Apr 24, 2019permalink

MamiyaMax

The resolution is far, far better than the Sony. When I downloaded the raw and sharpened it up, the detail was awesome; slightly better than the 100mp sensor.

Apr 24, 2019permalink

samfan

And I used to think the megapixel race was over at 24 MPx and nobody really needs more than 16... Apart from the crazy 50 MPx MF people...

Now I feel stupid.

Just for fun, add Nikon 1 S1 (which, I believe, is the lowest-res camera available to select) to the comparison.

Apr 23, 2019permalink

Merida

Don't believe the hype.

Apr 24, 2019permalink

I still really want to shoot the new scene with my D700. It'll happen. Some day.

Apr 24, 2019permalink

Photoman

D700 Rules! Also handy when you go camping and you forget your hammer for tent pegs ;)

Apr 24, 2019permalink

I still say that few commercial images need more than 18mp. D700 could still be used today. More pixels are for cropping.

Apr 24, 2019permalink

(unknown member)

Dear D700 Fan Boys...Nikon has made better. Please stop Spinning lies.

Apr 24, 2019permalink

samfan

Nikon D700 is the best ever, and also last good DSLR.

Made without compromises (pretty much 95% of D3), uncumbered with video, no bugs that come either unaddressed or at best get fixed 2 years after the fact. Best sensor and IQ available far and wide, all in a package that was a pinnacle of evolution of Nikon's engineering and user experience.

On the other hand it was so good that everyone wanted it, which somehow convinced Nikon that only FX makes sense for them. And it went downhill from there.

Apr 24, 2019*permalink

Jon Schick

Very impressed by the lens - the corners are superbly sharp. The resolution from the camera is very impressive too - Panasonic have done well. It looks noisier than some of the competition at ISOs higher than I ever use, but at lower ones it's outstanding. I hope the camera does well with the professional market, because I think it's too heavy for most amateurs to bother with (I was amazed at how bulky it is when I played with one recently but have to admit the build quality was superb) but maybe these capabilities will trickle down to a lighter body at some point in the future.

Apr 23, 2019permalink

alucard1357

Nice.really great balance between noise and details.

Apr 23, 2019permalink

TheRealEagleclaw

The people in the photos of the studio test scene were shot with a ring light. You can see it in their eyes. Damn!!!!!!

Like how long does the high res mode take because this is seriously impressive.

Apr 23, 2019permalink

Not long. Assembling the images takes less than ten seconds. Each individual image takes, well, your chosen shutter speed, as you might expect. So for fast shutter speeds, the whole process is pretty fast. The motion correction is pretty impressive too (there are samples in our S1R sample gallery).

Apr 23, 2019permalink

Max Iso

Looks like we have a new FF ML IQ leader. Good job Panasonic.

Apr 23, 2019permalink

Terrible Photographer

Wow, the Panasonic looks flat out amazing. When downsized to the Phase One file, I'd say it definitely looks better with better resolved fine details and no moire issues on anything. Well done.

What's really even more surprising, is take the multishot file from the Pen-F and compare it to the Multishot file from the Sony and downres to the Sony image size.

When you do that, the Micro 4/3rds camera looks better than the 45MP FF chip! Better resolving fine details (especially in the color circles in the bottom right) and no moire.
The multishot functions in the Oly's and Panasonics I think are doing it right by interpolating up instead of maintaining existing chip resolution. I know my M1X in the high res mode can make a more detailed file than my D810 anyway of the week.

Of course Multishot doesn't work for moving subjects, where the single shot Hi Res chips will still be better.

Apr 23, 2019permalink

Robin Ducker

I have seen images shot with Hi-Res mode 2 including things like moving water and leaves and believe me, the camera is a game changer.

Apr 23, 2019permalink

SDKat

At $6K, with base lens and battery, it BETTER make coffee!

Apr 23, 2019permalink

@SDKat Coffee ... You don't have an assistant?

Apr 24, 2019permalink

DougDolde

Not sharp compared to the Phase One, but a couple orders of magnitude cheaper

Apr 23, 2019permalink

GoodKnight

Very impressive technology. It would be cool to have a 24mp FF pixel shift capable camera. 24mp for high iso performance and pixel shift when you want more resolution and detail. All in one camera a birders sport shooters landscape camera :)

Apr 23, 2019permalink

s1oth1ovechunk

You mean like the s1?

Apr 23, 2019permalink

GoodKnight

Yes exactly what I was thinking

Apr 23, 2019permalink

badi

i think it would be cool if we have this and other multi-shot and computational stuff in all cameras from all brands :)

Apr 24, 2019*permalink

Keep the image at 24mp and make the color awesome?

Apr 24, 2019permalink

juncction

The peacock feathers must be a late addition. It's missing in both spots from the Phase One studio shot.

Apr 23, 2019permalink

UllerellU

What I see is that it is noticeably louder with high ISO than the other three, especially with very high isos where Sony and especially Phase One give a good review, in isos contained the Pentax seems to take the prize. In terms of sharpness is terrible

Apr 23, 2019permalink

(unknown member)

The hi-res mode only goes up to ISO3200 which looks very clean.

Apr 23, 2019permalink

UllerellU

No, is noisier than Sony & Pentax, look at the black, un the night Scene, there is more noise, un both modes, normal & hi-res.

Apr 23, 2019permalink

Hautedawg

I'm impressed by the Panasonic IQ. I don't think we're comparing apples to apples considering the difference in focal lengths and lens quality, but there is a lot of detail and contrast.

Apr 23, 2019permalink

liquidsquid

Wow, that resolution is incredible. The blue, white, and black fabric in the lower right. In other examples you cannot really tell it is fabric. On this cam? It is very clear it is!

Impressed by not only the resolution, but the sticker price :-(.

Apr 23, 2019permalink

caver3d

You get what you pay for. What did you expect? The S1 and S1R are great cameras.

Apr 23, 2019permalink

liquidsquid

I'm only griping because I would get what I paid for if I had the $$$ to pay for it. Being this is nothing more than a spare-time hobby for me and I am not raking it in, it wont happen for me.
Not to mention I would NEVER take advantage of a system like this, I suck.

Apr 23, 2019permalink

bokehmon22

I got the Panasonic S1 for $2250 with free battery and grip pre-order special.

That's just cheaper than EOS R with best EVF, IBIS, new sensor, dual card slot, 4K60, high resolution mode, focus bracketing.

It's the same price after selling my grip for better EVF (5.76 million dot EVF vs 2.4 million), better ergonomic, better built, 2.1million fully touch screen LCD vs 900,000 partial touch screen and no 4K60.

You get what you pay for and more with S1

Apr 25, 2019permalink

bokehmon22

I got the Panasonic S1 for $2250 with free battery and grip pre-order special.

That's just cheaper than EOS R with best EVF, IBIS, new sensor, dual card slot, 4K60, high resolution mode, focus bracketing.

Compare to Sony A7III, it's the same price after selling my grip for better EVF (5.76 million dot EVF vs 2.4 million), better ergonomic, better built, 2.1million fully touch screen LCD vs 900,000 partial touch screen and no 4K60.

You get what you pay for and more with S1. Best camera I've own and best value I've owned. I had Canon 5DIV and EOS R

Apr 25, 2019permalink

evetsf

@Androole: Citation please - where did you get your information that the S1 is using the same sensor as the others?

It actually doesn't seem that way. This was copied from the Dpreview "What You Need to Know" article published earlier this year:
>>It's worth noting that these sensors are not BSI-CMOS designs. This means that they are not variants of the manufactured sensors we've seen in contemporary Sony and Nikon DSLRs and mirrorless cameras.<<

Apr 23, 2019permalink

Richard Butler

I'd be inclined to go back and edit that article.

Panasonic doesn't say they're BSI designs but the performance characteristics are extremely similar to those of the Z6.

Apr 23, 2019permalink

@Richard Butler - agreed. Panasonic specifically talks about the aspherical waveguide lenses on their S1R, but there is no explicit mention with relation to the S1.

That in itself wouldn't be indicative of anything, but it's valuable to remember that Panasonic also specifically does NOT advertise the BSI aspect of their GH5S sensor (which we know is a BSI Sony IMX299), nor on any of their Sony-sourced 1" BSI sensors. They prefer "High Sensitivity MOS sensor."

Given the near identical sensor characteristics of this 24MP S1, I'm inclined to put 2 and 2 together on this one and suggest that Panasonic merely doesn't want to advertise BSI because it might cast their high-end FSI 20MP G9 and GH5 and 47MP S1R in a negative light given the lack of nuance that customers tend to apply to technology architectures.

Apr 23, 2019permalink

Richard Butler

I was thinking very much the same thing.

Apr 23, 2019permalink

caver3d

I agree. They just avoid saying BSI, and try to distance themselves from Sony. Understandable. But I have the S1, and LOVE IT!

Apr 23, 2019permalink

evetsf

Yes, I'm well aware that Panasonic is very close-mouthed about the sources of their sensors. And it may well be that the S1 uses that same sensor.

However, until someone takes these cameras apart and looks at the maker's stamp on the sensors, saying that the S1 "definitely" uses that same Sony sensor is inaccurate. At this time all we can do is speculate.

Apr 23, 2019permalink

Most of us try to distance themselves from Sony?

Apr 24, 2019permalink

Frank_BR

In high resolution mode, the Panasonic SR-1 almost catches up with the Phase One XF 100MP, which is really a feat.

On the other hand, both JPEG and RAW images from Panasonic use about twice as many bytes compared to Phase One. Panasonic definitely needs to improve the efficiency of data storage.

Apr 23, 2019permalink

liquidsquid

Note that JPEG compression in general results in image file sizes that reflect the amount of detail stored. If all the settings are the same between two images, the one with more detail will be larger. Raw size i snot as easy, and may be due to multiple factors.

Apr 23, 2019permalink

Frank_BR

Picture noise also increases the size of a JPEG file. Panasonic's effective resolution is lower than Phase One, but Panasonic's files are larger. Draw your conclusion.

Apr 23, 2019permalink

The JPEG is twice as many bytes because there are almost twice as many pixels. The Raw file is actually pretty efficient when you realize that the Phase One is storing a Bayer RAW (i.e. 100 MP), whereas the total lack of color aliasing in the Hi Res S1R image suggests a full RGB storage (i.e. 3 times 187 MP). When you convert the file to an uncompressed TIFF, it is 1.04 GB, so that suggests about 3:1 compression in the RAW file.

Apr 24, 2019permalink

Sergey74

Guys, it would not hurt to clean your test scene of dust. The photo shows it... Directly above the letters "Dpreview"

Apr 23, 2019*permalink

I tried hitting that spot with a powerful blast of canned air, but that stuff is pretty stuck. Maybe we'll make it a part of our standard studio assessment.

Apr 25, 2019permalink

evan ts

No Hi-Res OOC JPEG?

Apr 23, 2019permalink

DDWD10

For some reason it was omitted on the S1/S1R. The G9 offers RAW/JPEG.

Apr 23, 2019permalink

turvyT

There has to be some problem with the K1 image. It shows better iq in normal mode rather than in pixel shift mode. I don't remember this being the case some time ago.

Apr 23, 2019permalink

Richmondthefish

What part of the image are you looking at? Pixel Shift looks far more detailed when looking at the BW illustration in the middle left.

Apr 23, 2019permalink

turvyT

Balsamic vinager label, green and red color tubes, low right, old guy pencil sketch, ruler... When contrst is very high (B&W), maybe I could accept a difference, not even clear in that case.

Apr 23, 2019permalink

Bubu93

Holy sh*t how good is that Leica 90mm APO Summicron in the corners,

I mean, it's a 5000$ lens so it gotta be worth something, but still I'm impressed

Apr 23, 2019permalink

dtibi

Was thinking the same. Easily outperformed the medium format lens.

Apr 23, 2019permalink

dtibi

Or there is some software trickery builtin to the camera/lightroom and removes all CA

Apr 23, 2019*permalink

coasttocoastphoto

I'm a Leica S and SL shooter. I also have a Leica adaptor allowing me to use Hasselblad HC lenses on my Leica S. The Leica lenses are sharper and have more contrast than the Hasselblad lenses. However, I like using both as the Hasselblad lenses produce a very pleasing and sharp enough image. They also make some very nice out-of-focus circles in the background during night scenes, similar to the look of a cine lens. It's like seeing two planes of focus with how the subject is separated from the background. I haven't used a Phase camera or Schneider lenses, but the results with the sharpness test appear to be similar to the Leica and Hasselblad comparison. I suggest trying both. and if you are a Lumix shooter, your best results will be with Leica SL lenses, and in particular, a Summicron.

Apr 24, 2019permalink

It is visibly soft at 187 MP, but that could be the limit of the lens. The biggie is that there is not sign of aliasing and there are times when that could be very useful (like when you want to know what is actually in the image). I can see this camera gaining favor in several analytical fields. The actual resolution is on par with the 100MP Phase One, but without the aliasing. Pretty impressive for FF.

Apr 23, 2019permalink

187MP at f/5.6 is definitely diffraction-limited. But I don't know if it would resolve better at f/4, just based on lens quality.

Apr 23, 2019*permalink

The 187 MP image looks like it would benefit noticeably from sharpening.

Apr 23, 2019permalink

The S1 high-res image looks a little soft too, even at 'only' 96MP. They would both definitely benefit from more sharpening, relative to our standard processing.

Apr 23, 2019permalink

@ Carey Rose. I just downloaded the raw and did some experimenting. It does respond well to sharpening, but interesting that you get to see the result only in PS, because LR won't blow up to 1:1 with that large an image so Adobe has some work to do with LR as cameras go to higher res. LR has always struggled with my 5DSR and with the high res image from the S1R, it is just plain broken. BTW, I got the best results on sharpening the image with Neat Image as it is smart enough to not enhance noise. Importantly, I didn't see any aliasing pop in when I sharpened the image. The bottom of the circles in green/magenta concentric circles is a little fuzzy, but hardly what you would call aliasing.

Apr 23, 2019*permalink

IamJF

It's shooting many lower res pictures and calculates - these are of course NOT diffraction limited!

Apr 23, 2019permalink

@Dragonrider thanks for the info! I'm plugging away on other aspects of this and other reviews, so I haven't had time to dig into the files too much. Interesting (though not surprising) about Lightroom. The 187MP files brought my editing iMac to a crawl, even with 24GB of Ram.

Apr 23, 2019permalink

@IamJF. Diffraction limiting of the lens aperture is not the only challenge here. The pixels are way to big to directly resolve a 187 MP, so they have to using some tricky math to translate edge effects into resolution (something like super-resolution radar). It is a pretty tough math problem and is riddled with opportunity for errors that result in aliasing, so the fact the image shows essentially no aliasing (even after sharpening) is commendable. The place you can see the edge effect is in the concentric circles, where you get more resolution on the 45 degree axis (i.e. the corners of the pixels). Taking multiple controlled shots also gets rid of the effects of the Bayer grid (other than sensitivity loss), so the color transitions are clean. It really is a nice feature and more impressive than the first attempt by Olympus.

Apr 23, 2019*permalink

@ Cary Rose. I just built a new computer with 9900k, 64 GB of RAM, and a GTX 1080, so I don't feel computer limited. PS (and the Neat Image plug-in) handle the file just fine, so LR has a problem. The 16 bit TIFF file is 1.04GB, so it is a pretty fat picture :-).

Apr 23, 2019permalink

zodiacfml

It is not diffraction limited. You take the pixel size of the 47MP FF sensor and not the 187MP output. It is limited by the resolving capability of the system.
Definitely not 187MP quality output but it close or equals the 100MP MF system without moire.

Apr 24, 2019permalink

@zodiacfml. The resolution is way beyond that which can be directly resolved by a 47MP sensor because of the math involved, so, yes it is running into diffraction issues. Just because a different approach to achieving high res is used doesn't alter the fact that the lens aperture is generating diffraction at that resolution.

Apr 24, 2019permalink

@Carey Rose. Just discovered that LR will let you view the 187 MP image at 1:1, but only in the develop window and not in the library window for some strange reason. That sounds more like a bug than a fundamental limitation.

Apr 24, 2019permalink

Fotoni

How long it takes to shoot with high-res pixel shift? There is no delay between photos like Sony has? Pentax K-1 does not have delay. At 100% size high-res PS indeed looks soft, but when downsized to match Phase 100 MP or lower, I can see that high-res is still sharper and more detailed.

Every tested camera should have similar quality lenses. You could test this and pick proper lenses. Now it is like cheating because Panasonic has premium lens and others don't. It is good know how good results high-res can give, but not for this kind of comparisons.

Apr 23, 2019permalink

otto k

Sony and Pentax don't raise the resolution of the output, they just cycle 4 positions to eliminated issues with Bayer. Panasonic (S1, S1R, G9) up the resolution so also have higher requirements for the lens. Similarly for Olympus (EM5II, EM1II, etc).

Apr 23, 2019permalink

IamJF

Sonys 85/1.8 or 90mm (and Nikons f1.8s) are perfectly sharp in the corners stopped down - they are just cheaper 😉

Apr 23, 2019permalink

No noticeable delay between shots that I can tell. Assembling the image takes under ten seconds.

Apr 23, 2019permalink

https://www.dpreview.com/articles/3528119869/panasonic-lumix-s1r-studio-scene-with-high-res-mode-published

Hmmm. There doesn't seem to be anything that has enough fine detail to discern the difference between 100MP and 187MP. Quite an unusual problem to have!

(For what it's worth, the unsharpened RAW results from the Panasonic certainly do not contain near 187MP of acute detail, but do very well when compared only to the 100MP Phase One at the same scale. Including the K1 and A7r II in the comparison makes the detail too small to discern, since they are far lower MP cameras).

I don't think anyone at Panasonic would be particularly disappointed if you said their camera 'only' resolved 100MP.

Apr 23, 2019*permalink

(unknown member)

Look at the black/blue splotchy rectangle towards bottom right.

Suddenly you can see the texture of the printed cloth. Even that is too large to really judge just how much detail there is.

Apr 23, 2019*permalink

Oh wow, that's pretty cool, actually.

Apr 23, 2019permalink

ZeBebito

Don't see the point of the fake high-res mode and the high iso performance is subpar. But still, Panasonic made the right move upgrading to full frame. More options is always good.

Apr 23, 2019permalink

jdu_sg

"fake" is irrelevant. Either the camera feature produces results or it does not. Seems like Panasonic's version delivers. Why would you argue against increased functionality ?

Apr 23, 2019permalink

(unknown member)

The Panasonic hi-res shoot make the A7Riii pixel shift shot look like it was take with a cheap camera phone. I can see why ZeBebito is so upset.

DPR needs to update their scene because ll the line charts are too big to judge the amazing detail. Look at the texture in that black/blue splotchy rectangle. You can see it is printed cloth now.

Apr 23, 2019*permalink

...and the camera produces better results with the High-Res... It's impressive...

Apr 23, 2019permalink

ZeBebito

"The A7Riii pixel shift shot look like it was take with a cheap camera phone."

@Scotty if you are going to troll, at least learn how to write... "lookS"... "takeN"... you have been here like 2 weeks, please behave.

Apr 23, 2019*permalink

(unknown member)

ZeBebito,
Be civil. We both (along with virtually everyone) agree,,,
the Panny hi-res mode is noticeably better/more detailed than the A7Riii pixel shift mode.
You agree, right?

Apr 23, 2019permalink

MamiyaMax

I don't see any fake resolution at all. When I downloaded the raws and worked on them all, Sony was poor and the Panasonic actually slightly beat the 100mp sensor. I'm seeing a few comments about fake resolution that are completely false.

Apr 24, 2019permalink

Mark Hollister

Fake high-res? If the S1R is out resolving the other cameras and you can see greater detail than with the other cameras, then it's not fake, it's real.

It's like saying Margot Robbie isn't as pretty as she looks.

Apr 24, 2019permalink

Searching

High rez mode is really stunning.

Apr 23, 2019permalink

https://tinyurl.com/y25v32mg
https://tinyurl.com/y58d27mo
https://tinyurl.com/yysmte6d
https://tinyurl.com/y572ogj8

Here are some widget URLs I generated to compare the detail of the S1R high-res vs S1R vs Z7 vs A7rIII pixel shift. These are normalized to a common resolution:

Apr 23, 2019permalink

fpessolano

Interesting, I am happy I decided to stick with my Z7

Apr 23, 2019permalink

BTW I downloaded RAW files and opened in C1. A7riii image looks a bit brighter than S1R by about 0.28 EV.

Apr 23, 2019permalink

@fpessolano By default, Nikon's images indeed have more vibrant color to my eyes.

Apr 23, 2019*permalink

(unknown member)

Why did you normalize them?
You removed a huge amount of pixels from the Panasonic images.
Before you removed the pixels, the Panasonic images showed how much detail is lacking in the A7Riii pixel shift images. It is night and day.

Apr 23, 2019permalink

How were the images normalized to a common resolution? Reducing the resolution? This would just wipe off the advantages of the high-res mode...

Apr 23, 2019permalink

@Scotty, I normalized them because that's what would be done to print them at a common size.

@IdM, They were normalized by selecting the "Comp" button at the top of the widget.

Apr 23, 2019permalink

Ok, but this normalization is not well done... Resolution is reduced, detail is wiped off... The resolution should rather be increased, in order to see whether enlarging the image with less resolution gives simular results...
@dpreview change that!!!

Apr 23, 2019permalink

@IdM, I'm not sure what downsampling algorithm Dpreview's widget uses but to my eyes the high-res S1R image still shows noticeably more detail after downsampling vs the equivalent normal-res S1R and the Z7/A7rIII pixel-shift images.

Apr 23, 2019permalink

PWPhotography

I see about the same between downsampled S1R and A7r III pixel-shift. But then you lose the purpose of high-res if you have to reduce to get sharp and contrast details at pixel level.

Apr 23, 2019permalink

evan ts

Try normalize other images up to the S1R Hi-Res size.

Apr 23, 2019permalink

https://www.dpreview.com/galleries/0119553924/photos/3880957/compare

Done:

Apr 24, 2019*permalink

Nice comparison. However pls note that FE 85/1.8 doesn't quite fit into that league.

Apr 24, 2019permalink

Thanks! However, if you check dxomark.com, the FE85/1.8's sharpness is said to be fantastic (sharper than the 85mm/1.4 GM). I don't think that the difference in the results is due to the lens...

Apr 24, 2019*permalink

The Leica used on S1/R is an APO lens. I think FE 85/1.8 is indeed a great lens, but there is a chance that the larger summicron 90 could be a slightly greater one in IQ.

Apr 24, 2019permalink

(unknown member)

Horshack,

You said you normalized them to see what they would look like when printing them at a "common size".
That seems pretty silly. My smartphone takes pictures that look very detailed when printed at a "normal size".
But how will these look when cropped and displayed on an 8K TV or 5K monitor?

Apr 24, 2019permalink

@Scotty, they'll look the same considering I normalized them to 45MP, which is sill larger than the largest display medium you listed (8K screen, which is 34MP).

Apr 24, 2019permalink

(unknown member)

Horshack
Do you know what "cropping" is?
I mentioned it, but you must have missed it. It is very relevant to why your removing millions of pixels before comparing was a bad idea.

Apr 24, 2019*permalink

@Scotty, Yep, I'm familiar with cropping. It's what photographers do either to match a given format aspect ratio (which 45MP is still sufficient for the output mediums you mentioned) or when they're focal-length limited. If you believe the latter scenario is much more common than the former then sure, me removing millions of pixels was a bad idea. But that's not the case, unless you're a wildlife/bird-shooter (which isn't suitable for multi-shot pixel shifting anyway since those aren't static subjects) or you intentionally shoot with a shorter focal length than the situation requires.

Apr 24, 2019*permalink

(unknown member)

I tend to shoot a little wider so that if I change my mind on a scene later I can crop or leave wide. I also don't always carry my larger "longest" lenses, but cropping adds that extra reach with a hi-res image. And as you mention many times we all need to crop to a different aspect ratio.
So just going from 3:2 to 16:9 you lose 17% of your resolution. Then it is not uncommon to crop another 10-15% on top of that, and if we are trying to add "reach", 30-40% is not uncommon.

Apr 24, 2019permalink

@Scotty, that's a pretty specific scenario you've devised. Perfectly valid but quite specific. Do you believe that scenario is more common than the ones I mentioned, again keeping in mind the limited situations where multi-shot pixel shifting can be utilized due to the demands on camera-rig stability and subject matter movement?

Apr 24, 2019permalink

(unknown member)

Horshack,

It was not meant to be one specific scenario, but many or a combination. I frequently use 16:9, so all those times there is a 17% loss of pixels (more if shooting mFT), and it goes up from there depending on the situation.

I'll be curious to see how useful the hi-res mode is and how "quick" it is.

Apr 24, 2019*permalink

@Scotty, Understood, and certainly there are additional scenarios where upsizing to a common resolution is a better way to depict the resolution advantages of the S1R's pixel shift feature. I chose downsizing to a common resolution because it's what I believe represents the most common use case - as new higher MP sensors and pixel-shift modes become available I believe most will utilize the technology to increase the acuity of the print/render sizes they already use. Bayer sensors leave a lot of resolution on the table so back-filing the mosasic gaps with shifted pixels is a great way to realize the full resolution potential of existing pixel pitches.

Apr 24, 2019permalink

Astrotripper

There's one obvious take away from this. Leica SL 90mm f/2 is excellent optics, as is Schneider Kreuznach 80mm f/2.8. Can't say the same for Pentax 77 and Sony 85.

Kinda shows what kind of money you need to pay to get the most of those super resolution bodies.

Apr 23, 2019permalink

Tical

And yet we're not taking into account the print quality / print method factor: thinking such resolution could be useful to look at images on (any) display is out of question and print, whatever the method, is a huge leveller, believe me.

Apr 23, 2019permalink

wy2lam

Too bad Pentax stopped making its 85, it's probably up there if not better.

Apr 23, 2019permalink