CI Take on P1 IQ4 Frame Averaging - Whys and why nots, hows and how nots

[Steve Hendrix]

Brad Kaye has put together a comprehensive article regarding the pre-beta version of the Phase One IQ4 150 Frame Averaging Tool. This covers some of the troubling sample files, as well as new files created by Brad in evaluating the tool. I hope this is helpful.

https://captureintegration.com/phase-one-iq4-long-awaited-frame-averaging-nearly-available/


Steve Hendrix/CI

[Steve Hendrix]

Brad Kaye has put together a comprehensive article regarding the pre-beta version of the Phase One IQ4 150 Frame Averaging Tool. This covers some of the troubling sample files, as well as new files created by Brad in evaluating the tool. I hope this is helpful.

https://captureintegration.com/phase-one-iq4-long-awaited-frame-averaging-nearly-available/


Steve Hendrix/CI

Note that some images may have not been "clickable", this has now been corrected.


Steve Hendrix/CI

[DP]

> Key takeaway: Frame averaging a high ISO exposure will result in a decrease in luminance and chromatic noise.

averaging increase S/N always, ISO (high or low) has nothing to do w/ it...

[Steve Hendrix]

> Key takeaway: Frame averaging a high ISO exposure will result in a decrease in luminance and chromatic noise.

averaging increase S/N always, ISO (high or low) has nothing to do w/ it...

DP, thank you for pointing this out. What you have said is of course technically correct, although the reduction in high ISO noise is much more noticeable than a reduction in low ISO noise, which I suppose was the point behind the odd choice of Phase One to use an ISO 6400 daytime shot as an example for showing smooth water.

But this article has an ensuing narrative aspect. If you read further, you will see the section where we approach the idea of low ISO single frame captures vs low ISO frame averaged captures and reveal the noise reduction considerations there as well.


Steve Hendrix/CI

[Paul2660]

The examples I have been able to view from both 6400 and 12800, averaged, make the images appear very close if not exactly the same as images from ISO 400, to 50.  Averaged 50 images I don't see any if any gain in noise reduction.  However the 6400 images have much better color fidelity and details, unlike a single 6400 shot. 

So ISO may have nothing to do with it, but for the P1 solution it seems to make the camera ISOless, in effect everything looks the same.  Except for the motion issues which unfortunately make the use of this tool outdoor very limited.  Still makes sense to attempt a shot unless the conditions are just too harsh.

It's also interesting to see that P1 did not get the XF vibration sensor incorporated into the process, so that the XF is contributing to slight blur.

I noticed this with the official P1 landscape image, and a simple run of Topaz sharpen AI, using the stabilize setting seemingly pulls a lot of of that lost finer detail back.  Made a big different on the P1 test file for sure.

Paul C

[SrMi]

DP, thank you for pointing this out. What you have said is of course technically correct, although the reduction in high ISO noise is much more noticeable than a reduction in low ISO noise, which I suppose was the point behind the odd choice of Phase One to use an ISO 6400 daytime shot as an example for showing smooth water.

But this article has an ensuing narrative aspect. If you read further, you will see the section where we approach the idea of low ISO single frame captures vs low ISO frame averaged captures and reveal the noise reduction considerations there as well.


Steve Hendrix/CI

I assume that Phase One's averaging technology works only with the camera on the tripod. If that is the case, why would one shoot with anything but the lowest ISO possible (while using averaging technique)? I see using higher ISO when shooting handheld and aligning the frames in PS before averaging them. When would you expect users to shoot with high ISO and use Phase One's in camera averaging technique?

[Steve Hendrix]

The examples I have been able to view from both 6400 and 12800, averaged, make the images appear very close if not exactly the same as images from ISO 400, to 50.  Averaged 50 images I don't see any if any gain in noise reduction.  However the 6400 images have much better color fidelity and details, unlike a single 6400 shot. 

So ISO may have nothing to do with it, but for the P1 solution it seems to make the camera ISOless, in effect everything looks the same.  Except for the motion issues which unfortunately make the use of this tool outdoor very limited.  Still makes sense to attempt a shot unless the conditions are just too harsh.

It's also interesting to see that P1 did not get the XF vibration sensor incorporated into the process, so that the XF is contributing to slight blur.

I noticed this with the official P1 landscape image, and a simple run of Topaz sharpen AI, using the stabilize setting seemingly pulls a lot of of that lost finer detail back.  Made a big different on the P1 test file for sure.

Paul C

Paul, I expect some changes to the Frame Averaging Tool that will contribute to lessening the impact of vibration/movement on multiple images. We shall see.

Regarding the ISO 50 images, did you not see the example at the bottom of the article that showed noise creeping in with just a bit of post processing?


Steve Hendrix/CI

[Steve Hendrix]

I assume that Phase One's averaging technology works only with the camera on the tripod. If that is the case, why would one shoot with anything but the lowest ISO possible (while using averaging technique)? I see using higher ISO when shooting handheld and aligning the frames in PS before averaging them. When would you expect users to shoot with high ISO and use Phase One's in camera averaging technique?

This is part of the conundrum for determining use cases for frame averaging. It has a lot of power, but there are obviously many situations it does not work for. It has commonly been used for astrophotography, as one example. But I feel that the focus on high ISO is perhaps misplaced, and in fact that was the point of our article, other than high ISO, what use is this tool?

If anything, the benefits of improved image quality at low ISO were the most compelling aspects of the frame averaging tool in our view. Low ISO still exhibits noise even with modest post processing, but not with frame averaging (or at lest with much less). And the beautiful gradients that Brad Kaye produced for the interior shot were so much nicer than the single shot frames.

These are reasons to use frame averaging, not just for high ISO situations.

But also the point of the Beta labs philosophy is not so Phase One can tell you how to use the tool, but for photographers themselves to create workflows that the tool has merit for. Photographers have always created workflows and uses for tools that manufacturers have never even considered. Here's your chance to do so.


Steve Hendrix/CI

[rogerxnz]

Great article on your website, Steve.

I found it helpful and objective. Frame averaging has pluses and minuses and it will be interesting to see what uses photographers develop for it.
Roger

[Steve Hendrix]

Great article on your website, Steve.

I found it helpful and objective. Frame averaging has pluses and minuses and it will be interesting to see what uses photographers develop for it.
Roger

Thank you Roger.

I hope that the testing that our team produced at the bottom answered the questions you and others have of the quality improvement with frame averaging, regardless of ISO. You need to read past the Phase One samples that we show at the top of the article. Those don't really do a good job of telling the critical components of the frame averaging story.

I realize some people may have seen the Phase One samples at the top and thought they already had covered that, but if you have, then you've missed some of the overlooked the critical advantages of frame averaging that we cover in the lower half of the article. So for anyone who has stopped short, please read on!


Steve Hendrix/CI

[Steve Hendrix]

FYI - Brad Kaye told me this morning that we have discovered this:

While the Vibration Delay/Seismometer of the XF is not observed, the delayed Drive Mode (Timer) is enabled and the XF will acknowledge that setting. We have updated the article to reflect this.


Steve Hendrix/CI

[Wayne Fox]

FYI - Brad Kaye told me this morning that we have discovered this:

While the Vibration Delay/Seismometer of the XF is not observed,

I tested this quite a bit. The vibration detect worked fine for me.  I handheld the camera and the exposure didn’t start until shortly after I sat the camera down and quit moving it.  It appears to “start”, but the actual exposures and the progress indicator don’t seem to start/move until the camera is stationary or 8 seconds passed.

[Steve Hendrix]

I tested this quite a bit. The vibration detect worked fine for me.  I handheld the camera and the exposure didn’t start until shortly after I sat the camera down and quit moving it.  It appears to “start”, but the actual exposures and the progress indicator don’t seem to start/move until the camera is stationary or 8 seconds passed.

Thanks Wayne for demonstrating this. We were told by Phase One that the beta version did not acknowledge the Vibration/Delay mode. However, with your findings, we have confirmed our IQ4 units are also acknowledging, so Brad Kaye has now updated that content on the blog.


Steve Hendrix/CI

[Steve Hendrix]

Frame Averaging recognition in Capture One browser

One of our clients contacted us with a feature request for an identifier for frame averaged shots in the browser, so they could easily see which shots were frame averaged and which were single capture without having to look at the metadata for each shot.

I imagine there may be at some point a "sort by frame average" command from the sort tool (similar to "sort by sequence" for focus stacking), but this still doesn't address the ability to be able to - at a glance - see which shots in the browser are frame averaged vs single frame.

In the meantime, Brad Kaye has developed a workaround using the Cap[ture One naming tokens that can effectively push an immediately visual recognition factor for frame averaged files in the browser.

Instructions below at the link:

https://captureintegration.com/tech-tip-iq4-frame-averaging-file-organization/


Steve Hendrix/CI