Lars Steenhoff

I've just found photoshop on my laptop so used that and yes, the UHD DC crop is a pixel for pixel crop from the centre of the sensor. You can A/B stills and cine and it's identical. So in this mode with no OLPF it would be interesting to see how much moire there is in motion.
I could guess that perhaps the DC HD mode would be pretty good too. The full frame HD mode not so much but there's a even scaling for HD.
cheers
Paul

Second attempt as i had uploaded the wrong images before.
So long short short,  i was testing resolution of 6k vs FF UHD to check the level of aliasing. The only chart i had was an old one i'd printed off myself many moons ago but should be enough to create issues (but i really need to find a better way...)
But i can clearly see 6K still vs scaled UHD but what i found interesting is if you switch to DC mode i am pretty sure you're just getting a 1:1 crop of the sensor in UHD - it looks similar to the 6K. Either way the DC UHD RAW mode is much clearer than the FF.
But i need to do some better tests but try it yourself and see what you think.
cheers
Paul
 
So top is full frame UHD and lower is DC UHD, see how much clearer the lines are

 

Have to add: Fuji should provide an FW upgrade that would record the gyro data in a separate file to be used for stabilization of any video recording mode in post processing, where you can fine-tune it clip by clip. Can any software accept such input data?

I've been playing around with it and can't fault it really. In fact for me, so far, it's been more reliable than Nikon's own FTZ adapter.
All my EF lenses have worked on it even old weird shit like Tokina 28-70 F2.6
It works well in video mode and has so far been far more reliable and bullet proof than Techart E-mount adapter for Nikon Z.



Going to continue to test but it looks like we did finally get that full frame 4K camera from Canikon.

By the way a lot of things have been discussed in the previous tread 
you can find it here
I would prefer it if there was only a single place to talk about the sigma fp, because it would be easier to follow the discussion.

No mechanical shutter, thats correct.

By the way a lot of things have been discussed in the previous tread 
you can find it here
I would prefer it if there was only a single place to talk about the sigma fp, because it would be easier to follow the discussion.

It does seem to be the Preview Exposure in M Mode where the bug is. On the card, and in playback mode, a still does match the video recording at same manual settings and ISO.
The preview in M mode function is too bright in live-view, giving misleading exposure as you set your settings.
I too tried the manual aperture and focal length input in the optics correction menu but no luck with a fix yet.
I am positively certain this is not going to be a long term bug and in a couple of weeks will be 100% fixed.
Very annoying in the here and now though.
Might be related to another bug... With manual lens and mystery aperture, Exposure bracketing doesn't work.
UPDATE: It does work, camera has to be in continuous shooting mode not single shot. Funnily enough the exposure in live-view now matches the +1 ev shot perfectly, so there's a work-around of sorts.

Two of the most unique cameras on the market today are both L-mount with 10bit capabilities in 4K. The Sigma Fp of course shoots Cinema DNG 4K RAW at up to 12bit to SSD via USB C and 8bit RAW internally. The Leica SL2 shoots very high quality 10bit 4K LOG internally with more of a passing resemblance to Arri Alexa colour science and LOG profile.
I thought I'd get to know the two, if that's the case!
Read the full article

There's no noise reduction on/off switch in the menus so LCD is showing a JPEG or proxy still image with NR added. And it's a very clean sensor anyway, even at ISO 12,800 you have to zoom very very far into the image to see the slightest hint of noise.
I have tried every combination of setting to try and get the exposure to match between stills M live-view and final image when taken, and it's no go... Very strange as it's fine in Cine mode or with a lens that communicates aperture to the body. So I assume it is a bug where the lens aperture is unknown, and that will need a firmware update. For now, I do advise Auto ISO and Exposure Comp in M mode, or use A mode, as a workaround.
Update:
Actually since Exposure Comp (effectively ISO) does not work in M mode on any of the dials and can't be assigned (have to dive into menus for it), here is a better workaround when using lens with manual aperture and you want full manual control with matching LV / stills.
- Use S priority mode instead of M
- Set ISO to Auto and control it with Expo Comp.
- Set aperture manually on your adapted lens (aperture ring)
- Set shutter speed manually
That is pretty much it. Now the problem is gone.

Hopefully Sigma will read all of this before putting a firmware update out. I'll send them an email with all the annoyances and bugs in it.

Yes, you're right of course i think on a practical level in those cases it's easier to take the hit for not being 100% physically accurate as a scale is usually pre processing before doing linear comp work (or after). It's nice to run energy conserving convolution filters of lens aberrations over CG footage, works really well, better than blurs.
Yes, these are the kinds of thought experiments i'm trying to wrap my head around. In the case of the edge case (ha pun intended!) even with the full RGGB bayer info would that situation be much different - if the missing pixel between the R and the R is a green - the missing red would still be interpolated. You can of course argue that the luma of the G in that position would be the basis for that interpolation and therefore more accurate. If you scaled the channels down first then output that RGR edge - it would be softened - in your example that applies if you scale up, you are adding data. But scaling down you are removing and averaging. So the thought experiment is what is the best scale to do here - is it nearest neighbour or some kind of non sharpening scale as i can see edge issues with sharpening.
I must assume sigma has tested this if it is in fact what they're doing. 
I must also go back to the basic question i wonder if there's anyway just to get 6K out of this thing. Do we know which specific USB version it's running and what the data rate would be? Also isn't the Red patent to do with lossy compression? If you just do run length encoding on the stream does that not work?
I think sigma should open source the OS
So i've seen a case with the stills doing odd things with exposure, when switching between cine and stills and i am trying to match for testing. But after a reset it stopped. So i reported a bug but wasn't able to give steps to reproduce.
I got caught by having the camera in S mode with a manual lens yesterday - what's annoying is the screen is compensating. So i assume that the camera thinks there is a lens where it can adjust aperture thinks it's stopped the lens down for example and compensates on the LCD but the actual shot is exposed totally wrong. So the M mode is essential.
But it makes me wonder *why* the screen isn't showing an accurate preview of the exposure (and i had the setting on)
So there are some bugs. I've not been massively affected by a shutter lag though. Does it help if you turn off screen black out and also turn on the low previews mode?
cheers
Paul

I can also warmly recommend the Tokina 28-70mm/2.8 which is a bought-up and continued Angenieux design:
https://cameragx.com/2018/04/11/the-truth-about-the-angenieux-28-70-af-zoom/

Now the good news... That Contax Zeiss 28-70mm F3.5 is a stunner, a real gem. So small and light. Almost as small as the Zeiss 55mm F1.8 FE
It does not seem parfocal but it has character. Amazing macro mode too and the Zeiss look.
I have quite enjoyed putting it on the Leica M AF adapter on my Sony a9 too... Autofocus it

Even the Sigma 45mm F2.8 is a fatter lens.

Can confirm I have same issue! In stills mode at least, not in video or cine mode.
It only happens with a passive manual lens adapter where the aperture can't be read by the camera (F--)
Sigma or Pana L-mount lens it is fine
I also get a jumpy flickering of brightness in live-view, like the camera is trying to work out the metering
I think it's a LV metering bug.
Picture comes out same exposure as LV in cine mode... But switch back to Stills LV and exposure is brighter.
Very weird.
Can confirm the slight shutter lag as well.
It's a lot longer than average.

Yes it is, I'm still trying to love it, but for stills this makes me crazy.
It all depends in what fixes the next firmware will bring.
and the shutter lag, that also makes me crazy, because when I think I have the shot, it takes it a fraction later.

I will be surprised if Resolves does rescale in anything different than the image native gamma, that is in whatever gamma the values are at the point of the rescale operation. But if anything, some apps convert from sRGB or power gamma to linear for scaling, and then back.
You can do various transforms to shrink the ratio between extremes, and this will generally reduce ringing artifacts. I know people deliberately gamma/log transform linear rendered images for rescale. But it is mathematically and physically incorrect. There are examples and lengthy write-ups on the web with what might go wrong if you scale in non-linear gamma, but perhaps most intuitively you can think about it in an "energy conserving" manner. If you don't do it in linear, you are altering the (locally) average brightness of the scene. You may not see this easily in real life images, because it will often be masked by detail, but do a thought experiment about, say, a greyscale synthetic 2x1 image scaled down to a 1x1 image and see what happens.
I have a strong dislike for ringing artifacts myself, but I believe the correct approach to reduce these would be to pre-blur to band limit the signal and/or use a different filter: for example, Lanczos with less lobes, or Lanczos with pre-weighted samples; or go to splines/cubic; and sometimes bilinear is fine for downscale between 1x and 2x, since it has only positive weights. On the other hand, as we all very well know, theory and practice can diverge, so whatever produces good looking results is fine.
Rescaling Bayer data is certainly more artifact prone, because of the missing samples, and the unknown of the subsequent deBayer algorithm. This is also the main reason SlimRAW only downscales precisely 2x for DNG proxies.
It is actually possible to do Bayer aware interpolation and scale 3 layers instead of 4. This way the green channel will benefit from double the information compared to the others. You can think of this as interpolating "in place", rather than scaling with subsequent Bayer rearrangement.
Similar to how you can scale a full color image in dozens of ways, you can do the same with a Bayer mosaic, and I don't think there is a "proper" way to do this. It is all a matter of managing trade offs, with the added complexity that you have no control over exactly how the image will be then debayered in post. It is in this sense that rescaling Bayer is worse -- you are creating an intermediate image, which will need to endure some serious additional reconstruction. Ideally, you should resize after debayering, because an advanced debayer method will try to use all channels simultaneously (also, see below).
This is possible, and you can definitely extract more information and get better results by using neighboring pixels of different color because channels correlate somewhat. Exploiting this correlation is at the heart of many debayer algorithms, and, in some sense, memorizing many patterns of correlating samples is how recent NN based debayering models work. But if you go this way, you may just as well compress and record the debayered image with enough additional metadata to allow WB tweaks and exposure compensation in post, or simply go the partially debayered route similar to BRAW or Canon Raw Light.
 
In any case, we should also have in mind that the higher the resolution, the less noticeable the artifacts. And 4K is quite a lot of pixels. In real life images I don't think it is very likely that there will be noticeable problems, other than the occasional no-OLPF aliasing issues.

Just to make the matter less abstract, the problem of scaling a 6K Bayer pattern into a 4K Bayer pattern is that this...

...needs to be turned into this:

Which is only possible with weird interpolations.
Or, on a larger scale, if you have 3x3 of the above four pixel groups....

....how to scale them into 2x2:

However you do it, it will result in weirdness, since you're not actually interpolating adjacent color values (as you would do in an undebayered image).
 
 

Binning is also scaling. Hardware binning will normally just group per-channel pixels together without further spatial considerations, but a weighted binning techique is basically bilinear interpolation (when halving resolution).
Mathematically, scaling should be done in linear, assuming samples are in an approximately linear gamut, which may or may not be the case. Digital sensors, in general, have good linearity of light intensity levels (certainly way more consistent than film), but native sensor gamut is not a clean linear tri-color space. If you recall the rules of proper compositing, scaling itself is very similar -- you do it in linear to preserve the way light behaves. You sometimes may get better results with non-linear data, but this is likely related to idiosyncrasies of the specific case and is not the norm.
 
re: Sigma's downscale
I assume, yes, they simply downsample per channel and arrange into a Bayer mosaic.
Bayer reconstruction itself is a process of interpolation, you need to conjure samples out of thin air. No matter how advanced the method, and there are some really involved methods, it is really just that, divination of sample values. So anything that loses information beforehand, including channel downsample, will hinder reconstruction. Depending on the way the downscale is done, you can obstruct reconstruction of some shapes more than others, so you might need to prioritize this or that. A simple example of tradeoffs: binning may have better SNR than some interpolation methods but will result in worse diagonal detail.

I think the Rs are good for photography because they tend to show two different renders wide open (glowing) and stopped down, especially the Mandler designed ones (the lux's you quote). I found the 90 summicron very poor with flare, almost unusable. The 24 is not a Leica design and it shows. And the 28mm f2.8 Mark II is excellent. The last version of the 19 pretty good too, much better than the contax 18. 
Yes, i have focus gears for the M lenses. Studio AFS make some aluminium gears that you can twist on, like the Zeiss gears but they do them with scalloped inserts that will hold around the smaller barrels of the Ms. The Zeiss versions don't go small enough. They're really good and easy to twist on and off when needed.
The issue with the Ms are cost. I'm heading for a 90 APO and the latest 28 summicron. But i may have the sell the first born. But actually you get what you pay for (up to a certain point)
I don't find Resolves debayer very good. I am trying to get some fixes to Nuke that will allow these DNGs to go through there. 
I don't see there being any reason why we should see coloured pixels if the algorithm understands what the content is, right? I mean it's all recreated - so why false colours? In this case i believe this artefacts look like the AHD(?) algorithm.
cheers
Paul

Seems the 4k of the fp is not perfect, but I knew that, still its nice to see some test.

compared to the zcam e2
And heres is a nice guide for the contax lenses.
http://www.reduser.net/forum/showthread.php?92044-Contax-Zeiss-Survival-Guide
Most users tend to think the 28-70 is the worst zeiss. but its not true, it is actually the zoom that is not push pull from the range.
 

Seems the 4k of the fp is not perfect, but I knew that, still its nice to see some test.

I had to snap that up. Good find mate
Now I am waiting for my Fp to arrive. Sent the previous one back, and bought it for £1650 inc. 45mm F2.8. Yes, it may be an import. But if it breaks I can fly out, get it repaired in Japan, have a holiday, and still be £100 up vs the ridiculous UK price
Hopefully we will see some juicy firmware updates soon including Sigma-LOG
Or they might just call it V-LOG and nick it from Panasonic

This looks like a sincere question to me. I withdraw my claim you‘re trolling and apologize.
I for myself have a professionally printed A2 version of my test chart. However, for a zone plate chart, even that shows signs of remaining printing artifacts. And Marc doesn‘t have any usable printed version.
However and because we only needed a small fraction of the test pattern, there is an established alternative in the testing scene: monitors! As odd as it may seem at first glance, they can be put to good use if a number of rules are respected. I know about at least one professional lens calibration company which is using monitors to display test charts. The most important rule is that monitor subpixels (their projection when photographed or filmed) must be MUCH smaller than a sensor pixel. Other rules are that the test pattern must be resolved with no aliasing, and that there is no flicker. We obeyed these rules. Which is also why we know to have no extra moiré effects from the pixel grid.
It is also the reason why:
1. the monitor shows a small fraction of the test chart only, such that the remaining part is fully resolved by the 3840px wide monitor. The original test chart is 8400px wide and 9.6MB large. It‘s pixels are displayed at 100% or 1:1. There is no moiré in the screen display to the naked eye. The test chart file was carefully created to avoid aliasing as much as possible, by myself. Which is no easy task for a zone plate chart.
2. the monitor appears so small in the video, as it is far away. This makes the monitor pixels and subpixels disappear completely, there are more than 10 monitor pixels per camera pixel ... Hard to beat with any printed chart! It also ensures that we have spatial frequencies beyond 4k to test for. This is crucial for the test and any attempt to reproduce our results!
3. minor sources of blur (lens, focus, motion) destroy results as we depend on high spatial frequencies being resolved.
Btw, the article DOES contain a link to a test video before cropping.