Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
05-11-2018, 11:21 AM   #496
Pentaxian
MJKoski's Avatar

Join Date: Aug 2016
Posts: 1,273
Original Poster
That is just something which WORKED very very good with mk1. I did not expect something like that.

BTW, that electronic shutter "sony" bug. It also happens with Nikon D810. Sony sensor...like in K-1.

05-11-2018, 11:24 AM   #497
Senior Moderator
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: Durham, England
Posts: 11,083
QuoteOriginally posted by MJKoski Quote
That is just something which WORKED very very good with mk1. I did not expect something like that.
Sure. But then, the image processing path has changed. Very positive effect in some ways, but it has a negative impact in others. Depending on your priorities, the net result could be better or worse overall.

Let's hope they can add an NR toggle through firmware, that way both camps will be happy
05-11-2018, 11:40 AM   #498
Pentaxian




Join Date: May 2015
Posts: 318
QuoteOriginally posted by madbrain Quote
OK, so that makes two of us ! Found this video, but it has zero explanation .

Having never processed a file in RawTherapee, much less a PS file, it would seem that is not the best way to go for my first try.

OK, going to try this now.
Open the file with Rawtherapee and click the dropdown box top right. Pick one of the "PS iso" profiles. Then queue the file for export by clicking the gear icon or save the file using the save button at your bottom left.

Done. You can of course do much more but that should do it.
05-11-2018, 11:50 AM   #499
Loyal Site Supporter
Loyal Site Supporter




Join Date: Apr 2015
Photos: Albums
Posts: 706
QuoteOriginally posted by MJKoski Quote
It is not voodoo. There was long discussion about this on FM-forums Sony Zeiss 50/1.4 FE thread when initial buyers got the lens. First note was that it has harsh bokeh but no. It was only after disabling fully electronic shutter that users got smooth results.

---------- Post added 05-11-18 at 10:31 AM ----------

Here is a crop of PEF PS file. This is absolutely horrible result:



Now look at it...sharp details are sharp but everything else has melted down to smooth uniform areas. This is much worse than a single exposure. ISO800, f/5.6, Samyang 35/1.4

Here is the PEF (modified EXIF so it opens with current lightroom / any other software able to open mk-1 PS files)
_IMG0149.PEF - Google Drive
So...where's the issue in the bokeh in the images I posted?

05-11-2018, 12:28 PM - 1 Like   #500
Pentaxian
MJKoski's Avatar

Join Date: Aug 2016
Posts: 1,273
Original Poster
Nowhere, you have misunderstood the nature of this phenomenon.
05-11-2018, 12:33 PM - 1 Like   #501
Otis Memorial Pentaxian
Loyal Site Supporter
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 31,436
QuoteOriginally posted by MJKoski Quote
Here is a crop of PEF PS file. This is absolutely horrible result:



Now look at it...sharp details are sharp but everything else has melted down to smooth uniform areas. This is much worse than a single exposure. ISO800, f/5.6, Samyang 35/1.4
Is this to show electronic shutter problems or an issue with PSR. If the former, please show the standard shutter image. I am assuming the latter.

I extracted and processed separately, the four embedded exposures using dcraw. All four are at least as bad as the PSR result. Garbage in...garbage out, as the old data processing saying goes. I then took a close look at the garbage that went in and it was obvious that the DOF is only about 9mm. Without knowing the subject distance, it is hard to say how much blur to expect, but I am hesitant to blame this result on the camera. Miss focus on a ball of fur and one gets an out of focus ball of fur. In all fairness though, the in-focus portions including the transition zones have impressive detail*. The underexposed (we are talking LV 3 or thereabouts**) OOF (far) shadows are nasty as are the OOF (near) highlights to the right.

This is a challenging subject, to be sure, but I would like to see an example (same format, focal length, light, and sensor resolution) of a successful capture of this setup.


Steve

* I remain puzzled as to what detail or tonality of a this subject (hand duster?) looks like.

** LV based on camera settings was 5.9
05-11-2018, 12:44 PM   #502
Pentaxian
Not a Number's Avatar

Join Date: Mar 2012
Location: Venice, CA
Posts: 5,690
I opened the "fur" Raw in DCU 5.8.1. I don't see the "smoothing" of detail except when I turn off Pixel Shift Resolution.
05-11-2018, 12:47 PM   #503
Pentaxian
MJKoski's Avatar

Join Date: Aug 2016
Posts: 1,273
Original Poster
@stevebrot
Issue with PSR which might be affected by electronic shutter as well. But no matter, the way I see this is like:



Have a look at the topmost two waveforms - (one cycle) sine and square (copyright free image from wikipedia). Without accelerator small changes in sharpness work like sine wave (imagine peak as sharp details and bottom as pure blur) whereas with accelerator this smooth change from sharp to blur happens like 1 or 0 - the square pattern. Then have a look at the crop I just posted - the fur suddenly vanishes to totally melted blur instead of smooth analog curve. Test subject is again sheep fur blanket like earlier. Feel free to examine the RAW file.

05-11-2018, 02:45 PM   #504
Journeyman Cat Wrangler
Loyal Site Supporter
SSGGeezer's Avatar

Join Date: Jun 2017
Location: Maine, U.S.
Photos: Gallery | Albums
Posts: 3,570
QuoteOriginally posted by MJKoski Quote
@stevebrot
Issue with PSR which might be affected by electronic shutter as well. But no matter, the way I see this is like:



Have a look at the topmost two waveforms - (one cycle) sine and square (copyright free image from wikipedia). Without accelerator small changes in sharpness work like sine wave (imagine peak as sharp details and bottom as pure blur) whereas with accelerator this smooth change from sharp to blur happens like 1 or 0 - the square pattern. Then have a look at the crop I just posted - the fur suddenly vanishes to totally melted blur instead of smooth analog curve. Test subject is again sheep fur blanket like earlier. Feel free to examine the RAW file.
No they don't it is either digital or analog.
Sample rate has a great deal of how well a digital signal fares accuracy wise also.
05-11-2018, 05:11 PM   #505
Site Supporter
Site Supporter
madbrain's Avatar

Join Date: May 2008
Location: San Jose, CA
Posts: 1,167
QuoteOriginally posted by house Quote
Open the file with Rawtherapee and click the dropdown box top right. Pick one of the "PS iso" profiles. Then queue the file for export by clicking the gear icon or save the file using the save button at your bottom left.

Done. You can of course do much more but that should do it.
Thanks. All the RawTherapee icons look extremely small on my 4K screen. Unlike just about every other Windows program made, F1 does not bring up any manual or help. I couldn't find any PDF manual for RawTherapee 5.x either on their web site. 4.0 was the latest manual and is marked obsolete.
It took me a while to figure out what the "gear icon" might be. Is that the one next to the save button ? I'm using RawTherapee 5.4 .
Is it really necessary to use a processing queue when processing a single image ?

And is there any way in RawTherapee to see the 4 individual frames contained in the PS file ?
05-11-2018, 07:01 PM - 3 Likes   #506
Site Supporter
Site Supporter
madbrain's Avatar

Join Date: May 2008
Location: San Jose, CA
Posts: 1,167
QuoteOriginally posted by MJKoski Quote
That is just something which WORKED very very good with mk1. I did not expect something like that.

BTW, that electronic shutter "sony" bug. It also happens with Nikon D810. Sony sensor...like in K-1.
OK, I played your little game. Use tripod PixelShift on my K-1 II at ISO 800 . The lens I chose is an FA50/1.4, stopped down to F1.8 .
The subject is part of the inside of my K280 grand piano. There was a bit of daylight coming through the front door and windows, but ceiling LEDs were also turned on to help a bit.

I shot using infrared remote control, with 3 seconds delay. This automatically disables the SR in the camera.

I shot at two different ISOs - 500 and 800, single frame. I rendered these with the "auto-matched curved - ISO medium" profile.
Then, I shot again at those same ISOs, but with PixelShift . I rendered these with the "PixelShift - PS ISO medium" profile.
Output files were all JPEGs at the default settings - 92 percent quality. They were in the 5MB range.

I then uploaded them to imgur, which stripped all the EXIF data, including ISO and PixelShift data.

I won't tell you which one is which. I will leave it up to you figure that part out.

First image :



Second image :



Third image :



Fourth image :



You can pixel-peep all these images in your browser at the full 36MP resolution by clicking them, "view image" in Firefox, and then double-clicking the magnifying glass. I await your answer.

Adding the rules ::

We have 2 ISO 800 images, both with accelerator/NR . One with PS, one without PS.
And 2 ISO 500 images, both without the accelerator/NR . One with PS, one without PS.

Please state which 2 shots you think are with ISO 800, ie. with NR, and which two shots are ISO 500, ie. without NR.
And then, for both ISO 800 shot, figure out which is PS and non-PS. And same for both ISO 500 shots.

"best" is certainly a valid answer, but I was hoping MJKoski (and others who are RAW NR purists) could actually figure out when it's happening and when it isn't.

Last edited by madbrain; 05-12-2018 at 01:19 AM.
05-11-2018, 08:42 PM - 2 Likes   #507
Pentaxian




Join Date: May 2009
Location: Somewhere over the rainbow
Photos: Gallery
Posts: 781
QuoteOriginally posted by stevebrot Quote
That is why it is important to expose for the shadows,
I think it is more important to expose for where you want white to fall within the sensor ability to convert light information into data. If you expose for the shadows you are unknowing placing where you record white within the saturation space you are allowed and clip tonal values in zones 10-8.

QuoteOriginally posted by stevebrot Quote
though I am still not convinced that shot noise is as much a factor in blotchy capture at the low end as random misplacement of values by the A/D converter into the few bits available to express zones 1-3
You can easily test this by reduce the read noise that is captured in a image, this can be done by selection an exposure at base iso ( where we see the highest level of read noise) and then raising the iso to the point just before the camera uses NR and or stops using gain)
I am using the D800 as that is the only camera that is charged now

You can see that at iso 100 it produces 3.4 electrons and at iso 1600 it produces 2.3 electrons

To test this lets take 2 images that use the same exposure f/2.8 1/6 while changing the iso levels, one at iso 100 and the other at iso 1600 and look at objects that fall in zone 1-3



The top one is iso 100 and the bottom is iso 1600, for the iso 1600 I had to adjust the output brightness to the same level as the iso 100 shot, and because I increase the iso to 1600 I clipped all the data found above middle grey but as you can see that when I decease the read noise there is little effect on the noise we see between the 2 image.

However if I was to lift the shadows to the point that these tone values enter into the middle and highlights as you would in a HDR image then yes we would see some noise variances. Sony sensors produce so little read noise that they are really noiseless until we dig deep into the shadow well below the 10 zones that we can display as a medium, its the shot noise that decides how much noise we see in most of our work until we start doing HDR and compressing the full DR the sensor can capture into that of what we can display our images, shot noise is still greater contributor in a lot of what we shoot.
05-11-2018, 09:42 PM   #508
Otis Memorial Pentaxian
Loyal Site Supporter
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 31,436
QuoteOriginally posted by Ian Stuart Forsyth Quote
I think it is more important to expose for where you want white to fall within the sensor ability to convert light information into data. If you expose for the shadows you are unknowing placing where you record white within the saturation space you are allowed and clip tonal values in zones 10-8.



You can easily test this by reduce the read noise that is captured in a image, this can be done by selection an exposure at base iso ( where we see the highest level of read noise) and then raising the iso to the point just before the camera uses NR and or stops using gain)
I am using the D800 as that is the only camera that is charged now

You can see that at iso 100 it produces 3.4 electrons and at iso 1600 it produces 2.3 electrons

To test this lets take 2 images that use the same exposure f/2.8 1/6 while changing the iso levels, one at iso 100 and the other at iso 1600 and look at objects that fall in zone 1-3



The top one is iso 100 and the bottom is iso 1600, for the iso 1600 I had to adjust the output brightness to the same level as the iso 100 shot, and because I increase the iso to 1600 I clipped all the data found above middle grey but as you can see that when I decease the read noise there is little effect on the noise we see between the 2 image.

However if I was to lift the shadows to the point that these tone values enter into the middle and highlights as you would in a HDR image then yes we would see some noise variances. Sony sensors produce so little read noise that they are really noiseless until we dig deep into the shadow well below the 10 zones that we can display as a medium, its the shot noise that decides how much noise we see in most of our work until we start doing HDR and compressing the full DR the sensor can capture into that of what we can display our images, shot noise is still greater contributor in a lot of what we shoot.
I'm sorry, but this makes little sense. I wrote a whole lot more, but have decided to not include the detailed critique. I am becoming bored, but will at least suggest that the A/D converter clips at or near the point of clean linear response with that point being evaluated as 0 volts (zone 0). There is no deep basement on the low end.


Steve
05-11-2018, 10:37 PM   #509
Pentaxian




Join Date: May 2009
Location: Somewhere over the rainbow
Photos: Gallery
Posts: 781
QuoteOriginally posted by stevebrot Quote
I'm sorry, but this makes little sense.
How so ? read noise plays very little role in the noise we see from zones 1-10 and that shot noise play more of a significate role, its not until we want to process parts of the scene that fall deeper than what is found zone 1 and want to place them within the zones that can be displayed on our medium be it print or monitor.

If you look at the test images when I have reduced the read noise and it done little to change the noise found it the zones displayed. Now if I reduced the shot noise by increasing the signal then we would see cleaner tones in those displayed tones.

QuoteOriginally posted by stevebrot Quote
I am becoming bored, but will at least suggest that the A/D converter clips at or near the point of clean linear response with that point being evaluated as 0 volts (zone 0)
When I am discussing zones I am referring in the zones that we can display 10 being white and 5 being middle grey
05-11-2018, 11:08 PM   #510
Site Supporter
Site Supporter
Mallee Boy's Avatar

Join Date: Sep 2006
Location: South Australia
Photos: Gallery | Albums
Posts: 3,845
QuoteOriginally posted by madbrain Quote
OK, I played your little game. Use tripod PixelShift on my K-1 II at ISO 800 . The lens I chose is an FA50/1.4, stopped down to F1.8 .
The subject is part of the inside of my K280 grand piano. There was a bit of daylight coming through the front door and windows, but ceiling LEDs were also turned on to help a bit.

I shot using infrared remote control, with 3 seconds delay. This automatically disables the SR in the camera.

I shot at two different ISOs - 500 and 800, single frame. I rendered these with the "auto-matched curved - ISO medium" profile.
Then, I shot again at those same ISOs, but with PixelShift . I rendered these with the "PixelShift - PS ISO medium" profile.
Output files were all JPEGs at the default settings - 92 percent quality. They were in the 5MB range.

I then uploaded them to imgur, which stripped all the EXIF data, including ISO and PixelShift data.

I won't tell you which one is which. I will leave it up to you figure that part out.

First image :



Second image :



Third image :



Fourth image :



You can pixel-peep all these images in your browser at the full 36MP resolution by clicking them, "view image" in Firefox, and then double-clicking the magnifying glass. I await your answer.

I'll play....no 2 is shot with pixel shift. @ 500 iso & No 1 @ 800 .... but now I am guessing.

Last edited by Mallee Boy; 05-12-2018 at 03:20 AM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
accelerator, camera, claims, crop, dslr, exposure, full frame, full-frame, image, information, iso800, iv, k-1, k-1 mk2, k1, mess, mk2, op, pentax k-1, sensor, settings, software, sony, unit
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Macro Best moderate priced macro lens for newbie HGMerrill Photographic Technique 18 10-20-2014 06:08 PM
Old Moderate Mitt is Back! boriscleto General Talk 3 10-10-2012 02:12 PM
"Moderate" Mitt jeffkrol General Talk 2 10-08-2012 01:59 PM
Excesive noise in moderate light, please help Al_s14 Pentax K-r 4 08-03-2011 03:28 PM
Pentax lens all-in-one with moderate zoom fevbusch Pentax DSLR Discussion 17 01-27-2007 05:13 PM



All times are GMT -7. The time now is 03:11 AM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top