Pages: [1]   Go Down

Author Topic: Phase One IQ260 Long Exposure Sample Files  (Read 5245 times)

Doug Peterson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3404
    • http://www.doug-peterson.com
Phase One IQ260 Long Exposure Sample Files
« on: March 12, 2013, 07:48:15 PM »

The First of Many
Over the coming weeks we (Digital Transitions) will be posting a variety of sample images from the Phase One IQ260, Phase One IQ280, and Phase One IQ260 Achromatic on our DT Testing Blog.

Phase One IQ260 Long Exposure Torture Test

The most frequent request we've recevied has been for an example of the IQ260's long exposure capability. Normally you'd expect this to be demonstrated with a softly lit, low contrast scene with daylite balanced lighting; in other words a scene that would show the best-case scenario. But we thought we'd go a step further and do a true long exposure torture test.

So we took the prototype IQ260 and shot in a dim bar illuminated only by the exisiting tungsten lights and shot subject matter with fine detail, deep shadows, high contrast.


VERY IMPORTANT NOTE

This is a JPG generated by a pre-release BETA version of Capture One which provides preliminary support for our prototype IQ260. Neither the hardware nor software is complete, and some minor issues in the file can be observed (notably speckeled black pixels around areas of detail and a few areas of strange color in a highlight). These are the sorts of minor issues we expect to be fully resolved as the hardware and software is finalized for a public launch of the finished product in June.

---------

Download Full Resolution JPGs
See more examples and download full resolution JPGs here:
Phase One IQ260 Long Exposure Example - DT Blog






JoeKitchen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1129
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #1 on: March 12, 2013, 08:45:19 PM »

How did you adjust the light for the second longer shot?  Did you dim the bar lights or close the aperture?  

The 8 minute exposure does not seem as crisp as the 1 minute exposure.  It appears there is a minor increase in noise in the blacks and dark dark charcoals.  Not sure if it is a tiny bit of noise or diffraction.  

Very nice files though.  

I could just be crazy here too.   :D  Where I see what I am looking at is in the first close up, the green and yellow bans around the bottle appear to be not as crisp in the 8 min exposure. 
« Last Edit: March 12, 2013, 08:49:10 PM by JoeKitchen »
Logged
Joe Kitchen
www.josephmkitchen.com

"Photography is one percent inspiration and ninety-nine percent moving furniture."  Arnold Newman
“Don't bother just to be better than your contemporaries or predecessors. Try to be better than yourself.”  William Faulkner

Doug Peterson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3404
    • http://www.doug-peterson.com
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #2 on: March 12, 2013, 09:45:11 PM »

How did you adjust the light for the second longer shot?  Did you dim the bar lights or close the aperture?  

A 3 stop ND from B+W (Schneider's filter brand) took it from 1 min to 8 min. I was a bit concerned at first that the filter might pick up flare, ghosting etc (always a concern for long exposures in an environment with random/chaotic lighting). But as soon as the image was captured I checked it full screen on the iPad and was highly confident no such effects were taking place.

The 8 minute exposure does not seem as crisp as the 1 minute exposure.  

I notice a very minor loss of sharpness. It's hard to know for sure the source of this, as it was not a controlled laboratory experiment. This was done in a bar which was open (e.g. you can see some blur from the bartender spending time behind the bar) and a jazz performance going on. The floors were wooden and some very minor environmental vibration may have contributed. Absolute scientific sharpness was not necessarily my #1 priority in this test. I wanted to see how the sensor handled multi-minute exposures concerning color, noise, and tonality.

It appears there is a minor increase in noise in the blacks and dark dark charcoals.  Not sure if it is a tiny bit of noise or diffraction.  

Agreed there is a minor increase in noise in the deep shadows. Considering the jump from 1 minute to 8 minutes I personally think the increase is surprisingly small. As a point of comparison with the slower-ISO P45+ this would have been a 24 minute exposure. But we post them so you can make your own judgements in that regard.

It's worth noting the luminance noise reduction setting was turned to a very low setting of 8 (out of 100) for these comparisons. A bit more noise reduction did wonders for the deep shadow to quarter-tone transitions and deep shadow noise, but I wanted the comparison to show noise levels before that.

Paul2660

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2980
    • Photos of Arkansas
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #3 on: March 12, 2013, 10:49:54 PM »

Doug,

Have you worked with any iso 50 or 100 files, in normal exposures to see how the noise in the shadows is handled?  1/30, 1/60, to 1 sec?  My big hope is that the new chip will handle the transition better with cleaner shadows.  With the 160 I can usually pull up the shadows at iso50 1 to 1.5 stops and have nominal noise, but it's still there at times.  I am curious if this new chip will handle an even greater push, 2 to 2.5 stops with the same noise or less.

Paul Caldwell
Logged
Paul Caldwell
Little Rock, Arkansas U.S.
Photography > http://photosofarkansas.com
Blog> http://paulcaldwellphotography.com

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2902
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #5 on: March 13, 2013, 02:45:04 AM »

Looks very promising indeed, thanks for sharing.

It looks like there is some noise reduction applied (the pastel look), is it possible to disable that to see the "raw" performance too? Some of us prefer to keep noise (except the hot pixels!) rather than applying noise reduction which tends to kill color and detail.
Logged

Doug Peterson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3404
    • http://www.doug-peterson.com
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #6 on: March 13, 2013, 06:30:33 AM »

It looks like there is some noise reduction applied (the pastel look), is it possible to disable that to see the "raw" performance too? Some of us prefer to keep noise (except the hot pixels!) rather than applying noise reduction which tends to kill color and detail.

My usual strategy to accomodate this would be to provide a raw alongside the JPG.

This will have to wait until there is a public version of Capture One which fully supports the IQ260.

So for today you only have my JPG with some minimal noise reduction applied. But soon you will be able to play all you want!

MrSmith

  • Guest
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #7 on: March 13, 2013, 07:08:59 AM »

So did you choose to shoot in a bar because there was lots of hanging around doing the dark frames?
Smart move ;D
Logged

Doug Peterson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3404
    • http://www.doug-peterson.com
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #8 on: March 13, 2013, 07:40:21 AM »

So did you choose to shoot in a bar because there was lots of hanging around doing the dark frames?
Smart move ;D

That didn't hurt! With or without dark frames the process of capturing a half dozen long exposures is, well, LONG!  ;D

Though honestly the reason was that it turns out there aren't that many places around my apartment with so little light to allow an 8 minute exposure at ISO140 even with a 3 stop ND filter. The streets of NYC are too bright!

I'm considering another shot of the skyline or a building with a 6 stop ND filter, but I think I'll pursue other tests long before that.

FredBGG

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1633
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #9 on: March 13, 2013, 01:45:44 PM »



EDIT: IT LOOKS LIKE THIS WAS A SCREEN GRAB ERROR WHEN DOUG made the image to upload.
Most likely a graphics card redraw error. Defect not present if full frame jpeg. He is replacing the bad screen grab.

Viewed at 100% there is quite a bit of speckling noticeable around the text despite the moderate contrast.
See around the letters CHIE and the number 12. This is most likely due to interpolation to replace hot pixels.

There is also something else going on. On the 8 minute exposure the large letter R has an artifact that is larger and looks like
more than a hot pixel replacement. It looks like a vertical line has been replaced. There is also a similar horizontal one
just to the right of the letters SSI on the Classical sign.

While the speckling is more natural looking the vertical and horizontal "line" artifacts are rather nasty.
They are both absent in the 1 minute exposure and present in the 8 minute exposure. If these increase
even more for exposures over 8 minutes it could become quite problematic.

It would be interesting to see the dark frames as jpegs.

However with a 60MP file these artifacts will only be visible on big prints.

It also seems from the other crops that this problem does not show on more natural textures where interpolated
pixels to replace hot pixels will be less evident where there is more texture in the subject.
However it could be a problem with the typical high contrasts in city landscapes.

I have seen hot pixels with Canon and Nikon (D800) I have not seen these line artifacts.

Could this "line" artifact be column calibration that looks fine on a shorter exposure, but rears it's head on longer exposures?
« Last Edit: March 13, 2013, 04:25:39 PM by FredBGG »
Logged

gerald.d

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 436
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #10 on: March 13, 2013, 02:10:14 PM »

Damn you have good eyes Fred.

I'm going to offer up another possibility, because I have seen something similar to this before.

Doug - were LCC's shot and applied?

Is it possible that such artifacts could be introduced if the sensor or lens was shifted ever so slightly between the two exposures, but the same LCC applied to both?

Regards,

Gerald.
Logged

FredBGG

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1633
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #11 on: March 13, 2013, 02:17:39 PM »

Damn you have good eyes Fred.

I'm going to offer up another possibility, because I have seen something similar to this before.

Doug - were LCC's shot and applied?

Is it possible that such artifacts could be introduced if the sensor or lens was shifted ever so slightly between the two exposures, but the same LCC applied to both?

Regards,

Gerald.

I think the artifact would be repeated on the O next to the R if it was a movement.
Logged

TMARK

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1841
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #12 on: March 13, 2013, 02:40:59 PM »

I believe that this is a prototype, of sorts.  I had the line on my P30+.  CI sent a new back, but I believe that the fix involved tuning the readouts.  I doubt this would show up in a retail production model.
Logged

FredBGG

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1633
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #13 on: March 13, 2013, 03:29:35 PM »

I believe that this is a prototype, of sorts.  I had the line on my P30+.  CI sent a new back, but I believe that the fix involved tuning the readouts.  I doubt this would show up in a retail production model.

Was your line visible on short exposures or just long exposures?
Here is appears to be only on exposures longer than 1 minute.
Maybe..... and I mean maybe it's the tolerance of the column readout holding up at 1 minute and not at 8 minutes.

It could also be a one off freak thing and attributable to the prototype. Also looks quite retouchable and in practice may not be too much of a big deal.

It would be interesting to see if this gets noticeably worse at longer exposures such as 15 minutes or so.
Logged

Doug Peterson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3404
    • http://www.doug-peterson.com
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #14 on: March 13, 2013, 03:50:40 PM »

Viewed at 100% there is quite a bit of speckling noticeable around the text despite the moderate contrast.
See around the letters CHIE and the number 12.

Definitely. This is the speckling I mentioned in the notes "Neither the hardware nor software is complete, and some minor issues in the file can be observed... notably speckeled black pixels around areas of detail [...] These are the sorts of minor issues we expect to be fully resolved as the hardware and software is finalized for a public launch of the finished product in June."

I don't know to what extent this speckling will be present in the final product. I suspect we'll see a lot of improvement in this area.

There is also something else going on. On the 8 minute exposure the large letter R has an artifact that is larger and looks like
more than a hot pixel replacement. It looks like a vertical line has been replaced. There is also a similar horizontal one
just to the right of the letters SSI on the Classical sign.

To be honest I don't have a clue what happened here. I agree it's pretty ugly. If you look at the full resolution uncropped JPG (available for free download) this abnormality is not present.

The "quick comparisons" (the 100% crops posted in this thread) were based on screen grabs. So my best guess is that the beta version of C1 which provides preliminary support for the prototype IQ260 "hiccuped" while I was doing the screen grabs and was performing correctly when it processed the image.

Since it's not present in the raw file and I can't duplicate the issue in C1 (I tried reimporting it, adjusting the file, panning around, and haven't seen it) I feel comfortable replacing the screen grab here with one that does not.

Edit: the initial screen grab files don't show the anomaly either. So best guess is this happened because of an errant click/drag/mistake when I was copying the screen grabs into the photoshop template with the borders/labels/logo and selecting/nudging them around the screen.

Doug Peterson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3404
    • http://www.doug-peterson.com
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #15 on: March 13, 2013, 04:08:01 PM »

I believe that this is a prototype, of sorts.  I had the line on my P30+.

"calibration errors" (manifested as a subtle line through the image, usually more visible at higher ISOs) are rare but well understood in digital backs.

I would not be shocked to see that sort of thing (or other work-in-progress anomalies) on a prototype (this is not a "prototype of sorts" - it is most definitely a prototype; the shipping version won't be finalized until June) so that was a very good guess. However, in this case, I'm fairly certain now it came from me accidentally nudging a selection in Photoshop. It probably didn't help that I was doing the Photoshop work in the evening after a long day of work.  :D

The graphic will be corrected and replaced shortly...

FredBGG

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1633
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #16 on: March 13, 2013, 04:21:10 PM »


Since it's not present in the raw file and I can't duplicate the issue in C1 (I tried reimporting it, adjusting the file, panning around, and haven't seen it) I feel comfortable replacing the screen grab here with one that does not.

Edit: the initial screen grab files don't show the anomaly either. So best guess is this happened because of an errant click/drag/mistake when I was copying the screen grabs into the photoshop template with the borders/labels/logo and selecting/nudging them around the screen.

Yes the full frame jpeg does not have the "ugly" artifact. Must be a graphics card screen redraw issue when you did the screen shot.

I'll edit my previous post.
« Last Edit: March 13, 2013, 04:37:07 PM by FredBGG »
Logged

TMARK

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1841
Re: Phase One IQ260 Long Exposure Sample Files
« Reply #18 on: March 13, 2013, 04:39:03 PM »

"calibration errors" (manifested as a subtle line through the image, usually more visible at higher ISOs) are rare but well understood in digital backs.

I would not be shocked to see that sort of thing (or other work-in-progress anomalies) on a prototype (this is not a "prototype of sorts" - it is most definitely a prototype; the shipping version won't be finalized until June) so that was a very good guess. However, in this case, I'm fairly certain now it came from me accidentally nudging a selection in Photoshop. It probably didn't help that I was doing the Photoshop work in the evening after a long day of work.  :D

The graphic will be corrected and replaced shortly...

Don't Drink and Photoshop!
Logged
Pages: [1]   Go Up