New Pixii 26MP

Okay, here's a teaser for you avid Pixiiologists...

One of the knocks I remember from the handful of reviews that got written about the original 12-megapixel PIXII was that high-ISO performance was dismal. Naturally, I was curious to see if the new 26-megapixel megaPIXII does any better, since I take a lot of photos in dimly-lit places. So I managed to blunder through setting up a simple comparison test. As a reference, I used my Fujifilm X-T4, since it has the same size sensor (26mp, 25x17mm). I used the same lens for both, a 35mm f/2 Canon LTM lens with the appropriate adapters. (I like this lens because it's ridiculously tiny and very sharp.) I couldn't get the images to be quite identical because I'm still trying to figure out how to use the profiles/LUTs on the PIXII (there are seven of them, and they are completely undocumented, so I have no idea what the differences are between "Standard", "Standard S0", and "Default," for example.) But they're at least sort of similar.

Comparing the example photos turned out to be a bit tricky because Fujifilm, like the rest of the world except for France, apparently, increments their ISO ratings in 1/3-step intervals. PIXII does its ISOs in one-half stop increments. They do this because it is the PIXII Way, and one does not question the PIXII Way. So, the closest I was able to get to a head-to-head number was ISO 10,000 on the X-T4 vs. ISO 10,200 on the PIXII. Here are links to a couple of Lightroom side-by-side screenshots that you can scroll around in your web browser:

Full-image reference views: http://ranger9.net/pixiifun/xt4vspixii-fullscreen.jpg

300% crops: http://ranger9.net/pixiifun/xt4vspixii-300pct.jpg

So which is which? Let's pretend I forgot and you tell me...

For ISO 10,000, both are more than acceptable IMO. If the Pixii is the right hand image, yeah. I'll buy one .. eventually, at least. :)

G
 
"These feel like they are generated by an algorithm and offer no extra information. They just recite the specs. It is very common in YouTube in order to generate traffic."

Even if no algorithm was involved.... It is a bit similar to a "Free Energy" con video on YouTube...where some South East Asian person with a coil of copper wire and a capacitor and a sewing needle claims to produce so-called "free energy".
 
Welk as it is supposed to be rubbish at high Isos, it has to be the left hand one, right!?

That was the original PIXII, the one with the 12mp industrial sensor. As for the new 26mp model, well, no one really knows. Except me! And now you... because yes, the PIXII pic was the better-looking one on the right.

Here's another example with people, from yesterday, when I was hanging out in the wings at a ballet rehearsal. There's no noise reduction on either one, although saving as JPEG probably smooshed the noise a bit. The PIXII pic still looks considerably better to me in terms of noise and tonality than the X-T4 pic:

http://ranger9.net/pixiifun/noise_co...i_on_right.jpg

Now before everybody feels like they need to jump on the bandwagon and run out and buy a PIXII, let me just point out a couple of things:
  1. I'm comparing to a Fujifilm X-T 4, which I picked because I have one and because it has a sensor with the same size and megapixel count. Undoubtedly there are lots of other cameras out there (e.g. ones using the 36x24mm Simplex format) that would do better than either of them at ISO 10K.
  2. As I've been discovering, there are various other things about the PIXII that ARE problematic. Battery life is one: I got through only 120 shots yesterday before I hit the low-battery warning! I'm still tinkering with settings to see if I can improve on that, but so far this seems to be a knock on the original PIXII that has NOT been improved on the new model. The battery is an NP-FW50, the same as used on scads of Sony A6000-series cameras, so it's not going to be a problem picking up a spare or two, but if I'm going to be carrying around a lot of consumables I might as well use a film camera, right? :)
I will share some of these other downers over the next few days as I get them documented. Meanwhile, I've still got 12 days left on my 15-day trial, and the jury is still out...
 
That was the original PIXII, the one with the 12mp industrial sensor. As for the new 26mp model, well, no one really knows. Except me! And now you... because yes, the PIXII pic was the better-looking one on the right.

This looks great!

Here's another example with people, from yesterday, when I was hanging out in the wings at a ballet rehearsal. There's no noise reduction on either one, although saving as JPEG probably smooshed the noise a bit. The PIXII pic still looks considerably better to me in terms of noise and tonality than the X-T4 pic:

They probably have a built it noise reduction that you can't remove? Cause the lack of details is quite obvious. I wish this can be turned down somehow. One of the things you could test is if RAW is a bit better (Although noise reduction is still applied there as we have seen in previous Fuji cameras).

Now before everybody feels like they need to jump on the bandwagon and run out and buy a PIXII, let me just point out a couple of things:
  1. I'm comparing to a Fujifilm X-T 4, which I picked because I have one and because it has a sensor with the same size and megapixel count. Undoubtedly there are lots of other cameras out there (e.g. ones using the 36x24mm Simplex format) that would do better than either of them at ISO 10K.
  2. As I've been discovering, there are various other things about the PIXII that ARE problematic. Battery life is one: I got through only 120 shots yesterday before I hit the low-battery warning! I'm still tinkering with settings to see if I can improve on that, but so far this seems to be a knock on the original PIXII that has NOT been improved on the new model. The battery is an NP-FW50, the same as used on scads of Sony A6000-series cameras, so it's not going to be a problem picking up a spare or two, but if I'm going to be carrying around a lot of consumables I might as well use a film camera, right? :)
I will share some of these other downers over the next few days as I get them documented. Meanwhile, I've still got 12 days left on my 15-day trial, and the jury is still out...

I don't mind the battery that much. Neither the fixed storage in the camera, as you can easily transfer the images in a usb flash if you really need to.

I would really like to hear your opinion on 3 more things:

1. electronic shutter artifacts
2. Custom LUTs ( I believe they provide very little information on those)
3. OVF quality.

Thanks!
 
That was the original PIXII, the one with the 12mp industrial sensor. As for the new 26mp model, well, no one really knows. Except me! And now you... because yes, the PIXII pic was the better-looking one on the right.

That's very encouraging. I like. :)

Here's another example with people, from yesterday, when I was hanging out in the wings at a ballet rehearsal. There's no noise reduction on either one, although saving as JPEG probably smooshed the noise a bit. The PIXII pic still looks considerably better to me in terms of noise and tonality than the X-T4 pic:

http://ranger9.net/pixiifun/noise_co...i_on_right.jpg

Looks fine to me! I shoot at such elevated ISO settings so infrequently, and capture raw exclusively, so it's just not an issue if the high ISO JPEGs get a little soft. I'd be interested to see what its performance looks like at more normal ISO settings ... ISO 100 to, say, 1600 or 3200. Most of my picture taking happens in and around ISO 200 to ISO 400.

Now before everybody feels like they need to jump on the bandwagon and run out and buy a PIXII, let me just point out a couple of things:
  1. I'm comparing to a Fujifilm X-T 4, which I picked because I have one and because it has a sensor with the same size and megapixel count. Undoubtedly there are lots of other cameras out there (e.g. ones using the 36x24mm Simplex format) that would do better than either of them at ISO 10K.
  2. As I've been discovering, there are various other things about the PIXII that ARE problematic. Battery life is one: I got through only 120 shots yesterday before I hit the low-battery warning! I'm still tinkering with settings to see if I can improve on that, but so far this seems to be a knock on the original PIXII that has NOT been improved on the new model. The battery is an NP-FW50, the same as used on scads of Sony A6000-series cameras, so it's not going to be a problem picking up a spare or two, but if I'm going to be carrying around a lot of consumables I might as well use a film camera, right? :)
I will share some of these other downers over the next few days as I get them documented. Meanwhile, I've still got 12 days left on my 15-day trial, and the jury is still out...

I would compare it, image performance wise, to my Leica CL, that would be interesting. The only negative about the CL, for me, is that there are times when I'd dearly prefer an optical viewfinder/rangefinder over the EVF; it has a similarly short battery life. I never go out for a 'serious' shoot without at least one or two spare batteries, and I only rarely make more than 30-50 exposures in a session anyway. An easily available and inexpensive battery type is a big plus.

Nothing's perfect, but if the PIXII has a good rangefinder and a decent sensor, it will eventually end up on my shelf. I already have all the lenses I'll ever need... ;)

G
 
I would really like to hear your opinion on 3 more things:

1. electronic shutter artifacts
2. Custom LUTs ( I believe they provide very little information on those)
3. OVF quality.

Thanks!

1) Still working on that. With conventional fluorescent lights, I definitely get banding at shutter speeds higher than 1/125. Newer flicker-free LEDs seem to be no problem; older pulsed LEDs may produce some banding, but I haven't been able to rig up a situation to pin down how much. I haven't been able to text for motion distortion yet. Overall it seems to behave mostly like my Fujifilm X-T 4 when set to electronic-shutter mode; the difference is that on the PIXII that is the ONLY mode!

2) I haven't found any information about the custom LUTs at all. There are seven of them on the menu, but no documentation on what they do. It's possible they apply only to JPEG files, and I haven't shot in JPEG yet. One thing I'd like to know for the future is whether it's possible to create and install user-defined LUTs.

3) The OVF has a magnification of 0.67x, which is a comedown for me because I'm used to the beautiful 1.0x viewfinder on my R-D 1. But one advantage is that the 28mm frameline is easy to see, which is not the case on the R-D 1. The image seen through the finder is very sharp and detailed. The rangefinder patch seems less bright than those on my R-D 1 and Bessa R3M (possibly because it uses mirrors instead of a prism behind the rangefinder window) but focusing is no problem. One very disturbing issue with the finder is that the frame lines (lit by LEDs) seem to flare out at the corners. The flares are outside the frame, so don't interfere with composition, but they are distracting and detract from the perception of quality of the camera.

Another important viewfinder limitation to note is that the rangefinder window is right above the lens mount, so large-diameter lenses can block it and prevent focusing. The critical dimension seems to be about 60mm. The PIXII has a fairly short rangefinder base, so you really shouldn't be using fat wide-aperture lenses on it anyway for reasons of focusing accuracy, but people who have a large variety of M-mount lenses probably will be disappointed to find that some of them are not a good fit on the PIXII.
 
I thought I'd share another Pixii result since this is likely to be a deal-breaker for a lot of people. It's really slow! You can get your first shot off quickly, and a second immediately after if you want... but if you do that too many times, it bogs down because the throughput is terrible. With most digital cameras, you think about throughput in terms of how long it takes to flush the buffer onto the memory card. But with the Pixii, there's also a processing queue that can get backed up, so the queue and the buffer interact in unpredictable ways.

To get a real-world handle on what's going on, I set up a simple test. I set a metronome app to tick once per second, and set up the Pixii to take pictures of a clock face with a second hand. Once per second, I'd push the shutter button, and I'd keep pushing every second whether the Pixii recorded a picture or not. I kept doing this for one minute and then counted how many pictures I got.

(To make sure this was a fair test, I also tried it with my Olympus Pen-F, which is NOT known as a blazing speed demon of a camera. The shutter got a bit laggy after 45 seconds or so, but I WAS able to get all 60 pictures into a minute.)

I started out trying the Pixii using DNG format, which Pixii says gives the best balance of processing time and battery life. I was able to take one picture per second, but only for the first seven seconds. After that, the processing pipeline began to bog down, and I had to wait 23 seconds before I could take another picture. After that, I could continue to take pictures, but only about seven seconds apart. By the time the minute was up, I had captured only 11 pictures.

JPEG was a bit faster, but not much. Again, I was able to shoot one frame per second for the first seven seconds. Then the camera slowed down, usually to about one picture every six or seven seconds. At the end of the minute, I had captured 12 shots.

The Pixii also has a GPR format, which is like DNG but with wavelet compression applied to give a smaller file size. This format was super-slow. After the first picture, I had to wait 17 seconds before I could make another one. After that, the wait would be 15 seconds or so. During the minute, I was able to take only four pictures.

To be fair, the Pixii has "busy" indicators on both the top display and in the viewfinder to warn you when the camera isn't ready. But I think by now most of us expect our digital cameras to be always ready. It's hugely frustrating to see something good happening and your camera won't cooperate.

The Pixii needs a lot of breathing room, but it feels pretty responsive as long as you're shooting at a leisurely pace. If your dream camera is something like a knob-wind Leica IIIg, it may not bother you at all. It works best in situations where you can control the pace -- scenery and building shots, informal portraits, maybe model shoots with a relaxed model. But if you're trying to keep up with your kids playing or your cat acting crazy, the Pixii will drive you nuts and you should stay far, far away from it!

There, now nobody can say this is a fanboy review...
 
I thought I'd share another Pixii result since this is likely to be a deal-breaker for a lot of people. It's really slow! You can get your first shot off quickly, and a second immediately after if you want... but if you do that too many times, it bogs down because the throughput is terrible. ....

There, now nobody can say this is a fanboy review...

LOL! Thank you for the objectivity. :D

I remember having similar pipeline issues with the Sony F707 once upon a time, and never mind an insane amount of shutter release to exposure lag! Despite that, I figured out how to time my exposures and shoot the motorcycles when racing at the Isle of Man "Manx GP" event. It took some effort but made the photos all the better to see...

My Hasselblad 907x has some pipeline issues when used with the eshutter as well, never mind the 300ms scan time per capture. Such things don't bug me much, I just have to figure out how to get what I want out of the camera. It does limit some kinds of photo taking, of course. I guess that's why I have so darn many cameras in the closet... LOL! :angel:

G
 
LOL! Thank you for the objectivity. :D

I remember having similar pipeline issues with the Sony F707 once upon a time, and never mind an insane amount of shutter release to exposure lag! Despite that, I figured out how to time my exposures and shoot the motorcycles when racing at the Isle of Man "Manx GP" event. It took some effort but made the photos all the better to see...

My Hasselblad 907x has some pipeline issues when used with the eshutter as well, never mind the 300ms scan time per capture. Such things don't bug me much, I just have to figure out how to get what I want out of the camera. It does limit some kinds of photo taking, of course. I guess that's why I have so darn many cameras in the closet... LOL! :angel:

G

Funny, I was thinking about the Hasselblad 907x as well. Not that I've ever tried one, but it seems similar in that the user-experience stories I've read seem to divide between those who appreciate its beauty and don't mind its peculiarities, and those who are driven absolutely crazy by its frustrations and imperfections. It's a little less explicable in the case of the Hasselblad, which has DJI's money behind it, but it's a good reminder that boutique cameras are exactly that!
 
Ok, so when is the third version being released? ;)

Seriously, it wouldn't surprise me if a software update came along eventually to address some of this stuff. The latest revisions in the release-software channel were in 2020, so I think they're about due.

One of the interesting conceptual things about the Pixii is that instead of using custom application-specific integrated circuits (ASICs) like the big boys do with Expeed and BIONZ and so on, the whole schmeer runs on an ARM Cortex M0 processor with an integrated FPGA accelerator. FPGA stands for "field-programmable gate array" and is basically a blank ASIC -- the chip designer can design logic structures onto it and can update them at any time. Another interesting thing about an FPGA is that it can use "IP blocks" -- IP standing for intellectual property and essentially being licensable chunks of logic that perform specialized functions. So if Pixii SAS found some IP blocks that provided faster throughput than the ones they're using now, they could pretty much re-program the whole camera via an over-the-air update.

Am not saying they WILL, only that it's a possibility, and they designed it the way they did specifically to make that kind of thing possible.
 
I started out trying the Pixii using DNG format, which Pixii says gives the best balance of processing time and battery life. I was able to take one picture per second, but only for the first seven seconds. After that, the processing pipeline began to bog down, and I had to wait 23 seconds before I could take another picture. After that, I could continue to take pictures, but only about seven seconds apart. By the time the minute was up, I had captured only 11 pictures.


Ouch. Never expected for a new digital camera to compete for slowness of film ever again and with an internal drive no less. The unpredictable nature is the most annoying part.

Seriously, it wouldn't surprise me if a software update came along eventually to address some of this stuff. The latest revisions in the release-software channel were in 2020, so I think they're about due.

I really do hope so but I am not sure the have the capacity:
"SAS Pixii has 3 employees, and employs a few interns. This tight structure gives it real agility, which allows it to evolve the design of its device very quickly."
https://www.gazettebourgogne.fr/art...pareil-photo-fabrique-en-france-depuis-40-ans

The argue for agility, but not even achieving stable 1fps means this is not an agile camera :)
 
Right now it's still trickling upward through the Intertubes, but by morning US time you should be able to watch my initial YouTube review of the Pixii:

https://youtu.be/XPvCgE4wY4Q

I tried to keep it short because I figure viewers can take only so much of my droning voice and unprofessional delivery, but I did manage to cover the basic physical layout, menus, the app, and the problems I think are the most likely deal-breakers for most users.

Things I did NOT include, so may have to cover in a follow-up video, include: the viewfinder (good but not as nice as a Leica, Bessa, or Epson, mostly because the framelines flare out in the corners); consequences of not having a mechanical shutter (fluorescent lights do produce banding at shutter speeds above 1/125; pulsed LED lights produce banding at unpredictable shutter speeds because different LEDs use different pulse-clock rates); and how to use the top-panel meter readout (it has a few tricks that are useful but confusing at first.) I'd also like to do a photo shoot with live models so I can get more samples of how skin tones look under various lighting conditions.

All of that may or may not happen, though, because I am down to two days left on my 15-day return privilege, and I still haven't decided for sure whether I should keep the Pixii or not! Still, now you can watch the video and see why I'm in a quandary.
 
Right now it's still trickling upward through the Intertubes, but by morning US time you should be able to watch my initial YouTube review of the Pixii:

https://youtu.be/XPvCgE4wY4Q

I tried to keep it short because I figure viewers can take only so much of my droning voice and unprofessional delivery, but I did manage to cover the basic physical layout, menus, the app, and the problems I think are the most likely deal-breakers for most users.

Thanks, very nicely done. Interesting and intriguing!
 
Thanks for taking the time to do the excellent review. I’d still rather have the French girlfriend, or the Canon .095 lens you showed, and I still don’t think that the people who have detailed why they don’t want one/would never buy one, “hate” the camera, and I still don’t think the fact it’s made in France has anything to do with the hate they don’t actually have, and my reasons for not buying one in this day and age were not only confirmed, but amplified by your review……and that was before you even got to what you were listing as negatives for the camera.


But, I really enjoyed the review as it was more informative and well put together than many reviews done by people who do youtube reviews “for a living”. Nicely done. Thank you for helping to fill the strange factual information void surrounding this camera. And, yes, I’ve looked at everything else available.
 
I get you about the existence or non-existence of haters, and I certainly didn't mean here… but some of the comments I've been seeing on, for example, DPReview fora struck me as falling into the category of “unreasoning hatred.” I remember the same thing when the Epson R-D1 came out (yes, I bought one of those too, and I still own and use it.) There just seems to be a small but vocal group of people who feel that nobody except Leica should be allowed to make a rangefinder camera, at least not an expensive one, and when anybody tries, their reaction is “How dare they?!?” It seems odd to me, and probably you too, because we came up during the era when Leica M3 users and Nikon SP users and Canon 7 users (and a few Contax IIa diehards) were all very active partisans, so the idea that there's only One True Rangefinder never really resonated with us. Besides, I feel like haters in general deserve to be laughed at a bit!
 
Thanks for the great review. I found it really informative and entertaining. I'm really interested in the Pixii although it is right outside my price range and my children are too old to sell ;). It is exciting to see a company having a crack at producing a new camera with a different approach. If I had been in the market your review would have given me a lot to think about and I'm not sure which way I would have gone. My M lenses all take 39mm filters except for the new CV 50/1.5 (43mm) so lens size is not a problem. I am basically a film shooter and shoot digital in the same way (which is why I like not having a screen and my phone would stay in my pocket) so the processing time and battery capacity would not bother me. Looks like I might end up buying it but for now I'll have to make do with my XE-1 and Novoflex adapter.

Cheers,
Clay
 
Back
Top