Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: Doug Peterson on June 25, 2019, 07:00:23 am
-
The first major update to the Phase One IQ4 is expected to be released to users next week. It brings:
- Automatic Frame Averaging
- Improved storage options
- The ability to load custom C1 Styles into the back
It also fixes the vertical bug we reported on (thanks to our savvy user) and is expected to add a B/T mode to ES when used on a tech camera. It does not include zero latency or Ad Hoc wireless.
We (DT) will have more information, sample files, and analysis up on our website later today.
-
Will automatic frame averaging work on a tech camera with ES? Or only XF.
I am assuming it’s only with XF
Paul C
-
Will automatic frame averaging work on a tech camera with ES? Or only XF.
I am assuming it’s only with XF
It works on any camera the IQ4 works on, including tech cameras and view cameras (e.g. Arca R (https://www.dtcommercialphoto.com/arca-swiss-technical-cameras/)).
Indeed, given how game-changing I expect this feature to be to landscape and architecture shooters, I'd expect a lot of use of this feature on tech cameras.
-
Great, looking forward to seeing your report.
Paul C
-
Our initial write up is here, including raws from Denmark. Just as soon as we can we'll have more raw files shot here in NYC and/or LA to share.
https://www.dtcommercialphoto.com/the-phase-one-iq4-now-has-automatic-frame-averaging/
-
I am not sure that I understand what frame averaging means but the mention of in-camera multiple exposures interests me.
Do you know whether, when doing multiple exposures, the back shows you, when composing the next shot, the previous image or some kind of mask of the previous shot?
Roger
-
Interesting development indeed.
Clear article well to the point. Except for the surprising mention of the Canon 5Dr has some sort of high end 35mm camera... when it is based on the most outdated sensor available on the market today. :D
Cheers,
Bernard
-
Interesting development indeed.
Clear article well to the point. Except for the surprising mention of the Canon 5Dr has some sort of high end 35mm camera... when it is based on the most outdated sensor available on the market today. :D
Cheers,
Bernard
I commend Phase One for adding this feature.
It seems to me that it is essentially Nikon D8xx multiple exposure with a different/better UI.
I do not know if Nikon would have gapless exposures if using ES.
A useful application of that feature would be to eliminate the need for long exposure NR.
I do not understand why many manufacturers have either skipped on multiple exposure feature or generate only JPG as output.
-
I would think that multiple exposure is simply sequential addition while averaging needs to manipulate all the images as a set in memory?
The other cameras may not have enough memory to manage this?
Besides the development for those has been more focused on using the IBIS to do space multi-shot rather than time based one?
Cheers,
Bernard
-
Moreover, the IQ4 can do it entirely free of temporal gaps
I would like to know how can you achieve that, considering the scan time required for each exposure.
-
I would like to know how can you achieve that, considering the scan time required for each exposure.
Each exposure takes the time it takes, but the time gap btw the end of an exposure and the start of the next one can be very short.
Cheers,
Bernard
-
Each exposure takes the time it takes, but the time gap btw the end of an exposure and the start of the next one can be very short.
Cheers,
Bernard
Yes, but you don't capture the entire image at once as if you were using a global shutter. The most you can get is to start scanning the first line of pixels once you finish the last line of pixels of the previous frame, which means you have gaps per line
-
Yes, but you don't capture the entire image at once as if you were using a global shutter. The most you can get is to start scanning the first line of pixels once you finish the last line of pixels of the previous frame, which means you have gaps per line
Yes, indeed.
https://www.dtcommercialphoto.com/phase-ones-electronic-shutter-in-detail/
Scan time appears to be o.3 to o.6s.
Our favorite dealers may be able to confirm if our understanding is correct or not.
Cheers,
Bernard
-
I would think that multiple exposure is simply sequential addition while averaging needs to manipulate all the images as a set in memory?
The other cameras may not have enough memory to manage this?
Besides the development for those has been more focused on using the IBIS to do space multi-shot rather than time based one?
Cheers,
Bernard
In multiple exposure mode Nikon can use averaging overlay mode which should give the same result as Phase One (other options are Add, Lighten and Darken). However, Nikon's multiple exposure allows for a maximum of 9 or 10 exposures. I wonder what is the max number of exposures in Phase One's implementation.
The multi-shot technology requires different technology while multiple exposure is more about software and, as you said, available memory.
-
In multiple exposure mode Nikon can use averaging overlay mode which should give the same result as Phase One (other options are Add, Lighten and Darken). However, Nikon's multiple exposure allows for a maximum of 9 or 10 exposures. I wonder what is the max number of exposures in Phase One's implementation.
The multi-shot technology requires different technology while multiple exposure is more about software and, as you said, available memory.
One possible way to allow for unlimited exposures without a ridiculous amount of memory is to average each frame with the previous (average), so you just have to add and divide by 2 (or shift 1 bit in binary numbers). This is not "rigorous" averaging by mathematical definition, but if the brightness of the scene does not change, it would work fine.
-
Actually there is a way to calculate the rigorous average every time a new frame is exposed:
Suppose 2 frames A1 & A2, then average = (A1+A2)/2
If you have 3 frames A1, A2 & A3, you can either do average = (A1 + A2 + A3)/3 or 2*((A1+A2)/2))/3 + A3/3.
Note that the part of the equation highlighted in bold is the same as the average of two images, or "Previous average", so we can write:
Average of n images = (n-1)*(previous average)/n + (new exposure)/n ; for the nth exposure.
-
In multiple exposure mode Nikon can use averaging overlay mode which should give the same result as Phase One (other options are Add, Lighten and Darken). However, Nikon's multiple exposure allows for a maximum of 9 or 10 exposures. I wonder what is the max number of exposures in Phase One's implementation.
The multi-shot technology requires different technology while multiple exposure is more about software and, as you said, available memory.
Current maximum number of frames is 1000. But I don’t get the impression that limit is set in stone, and personally I’m advocating for a higher limit; power users hate limits, even when they are very high.
Back when I used to be a programmer my boss told me “think of the highest value a user could ever possibly want, in even the most absurd use case, then if doesn’t cost anything, add a zero or two” in regards to input fields and testing, which is often just a matter of allowing an extra digit and making some poor sap be the one sitting at a screen testing that a higher number doesn’t somehow cause problems.
-
Current maximum number of frames is 1000. But I don’t get the impression that limit is set in stone, and personally I’m advocating for a higher limit; power users hate limits, even when they are very high.
Back when I used to be a programmer my boss told me “think of the highest value a user could ever possibly want, in even the most absurd use case, then if doesn’t cost anything, add a zero or two” in regards to input fields and testing, which is often just a matter of allowing an extra digit and making some poor sap be the one sitting at a screen testing that a higher number doesn’t somehow cause problems.
Very nice. The camera is probably doing averaging every time a new exposure is performed, as has been suggested here by others.
-
We've updated our article (https://www.dtcommercialphoto.com/the-phase-one-iq4-now-has-automatic-frame-averaging) with new ISO50 comparisons at a high zoom level so you can easily examine the detail/noise change online. We will also have the raws available for direct download in about an hour so you can dive deep into them within Capture One.
-
The IQ4 Frame Averaging Raw Files (https://www.dtcommercialphoto.com/the-phase-one-iq4-now-has-automatic-frame-averaging/) are now available for download via our article (scroll to bottom).
-
Current maximum number of frames is 1000.
you’re article claimed “several minutes in bright sunlight.”
1000 limit equates to @ 8 seconds at 1/125th, 16 seconds at 1/60th, or 32 seconds if you could get to 1/30th (pretty hard to do in bright sunlight).
A long way from minutes. To get to “minutes in bright sunlight” would require a limit much higher than that, more like 5-10000.
Not that I’m particularly worried since I imagine I’ll be using exposures more like 1/8th.
Any clue when they’re gonna let this thing out into the wild?
-
you’re article claimed “several minutes in bright sunlight.”
1000 limit equates to @ 8 seconds at 1/125th, 16 seconds at 1/60th, or 32 seconds if you could get to 1/30th (pretty hard to do in bright sunlight).
A long way from minutes. To get to “minutes in bright sunlight” would require a limit much higher than that, more like 5-10000.
Not that I’m particularly worried since I imagine I’ll be using exposures more like 1/8th.
Any clue when they’re gonna let this thing out into the wild?
Great question! Since 1/30th is not a shutter speed that I expect to allow for gapless capture, the gaps between frames would space the 1000 frames out such that it still allows around 5-6 minutes max capture.
If minimizing the gaps between captures is important (e.g. subject matter is moving in a non-random way), or if going to very long total exposure times (e.g. 12 minutes in broad daylight) is important, then a single normal ND filter (e.g. 2 or 3 stops rather than 10 or 16 or 20) or a single circularizer polarizer filter should get you down to a shutter speed that provides a ton of flexibility.
But to be clear, since Feature Update 1 isn't out until next week, and because this is a fundamentally new way of working with exposure and capture, I don't think anyone (including R+D) has a full answer on how it will be best used or where the envelopes of performance and best practices will be. I'm excited to see how photographers like you use it, where you find it useful, where you find it useless, and how you get the most out of it. We'll do our best as a dealer to do our own testing and learning, but the thing that makes us smartest is that we listen to our really smart clients! So I hope you'll update this thread when you've had a chance to come to some of your own conclusions.
Personally, I'd also hope to see the limit raised beyond 1000 frames in future releases; if that's important to you as well, make sure to share that with your dealer!
-
I would like to know how can you achieve that, considering the scan time required for each exposure.
Each exposure takes the time it takes, but the time gap btw the end of an exposure and the start of the next one can be very short.
Yes, but you don't capture the entire image at once as if you were using a global shutter. The most you can get is to start scanning the first line of pixels once you finish the last line of pixels of the previous frame, which means you have gaps per line
I can totally understand why you would think this.
But when I say gapless, I mean just that. My example of a car headlights moving through a desert road is a good one example: the line from the headlights will be continuous not dashed.
The sensor is being controlled in a new way for this feature, so from the time a line of pixels rolls on (in the first exposure) until that line of pixel rolls off (in the last exposure), every line of pixels (and every pixel in that line), is continuously exposing. Subject matter movement is smooth and uninterrupted.
The only caveat is that all of the above applies to exposure longer than a specific shutter speed. Right now it looks like that will be 1/4th or longer for the IQ4 150mp.
Not every scene will require totally gapless capture. For example, averaging 2000 frames of a waterfall at 1/30th won’t technically be gapless but since the motion is semi repetitive (as opposed to the car headlight in the desert which is directional and non repetitive motion) the result will be complete silk not meaningfully different than a gapless capture (which could be done if the shutter speed was lowered to 1/4th or slower).
The raws we have in our article, for example, we’re shot at 1/125th and look quite good to me. Only turn, and the experience of our clients, will guide best practices for how to beat use this tool.
-
O and the maximum number of frames has risen since I previously posted about it.
Previously I said a maximum of 1000 frames. It now looks like the maximum will be ~3000-4000 frames with a maximum total length of ~10-15 minutes.
Both numbers (frame count and max total length) are being worked on still so may change before the release next week or may be changed in future releases thereafter. Personally I’d hope for at least 20 minutes max total time.
-
This is really interesting me.
The IQ4 has been a revelation for me on the Arca RM3di the last 6 months. The UI has made me so much more mobile and agile than on the old CCD backs.
Yet, now this comes along and I see amazing usage for both interiors and exteriors of Architecture.
One thing that would be interesting is how we can actually capture people in use of the spaces which is a request of many many architectural clients.
With the already great dynamic range of this sensor, I feel like this might make my 'lighting' bag lighter...
I'm really looking forward to this download next week.
-
This is really interesting me.
The IQ4 has been a revelation for me on the Arca RM3di the last 6 months. The UI has made me so much more mobile and agile than on the old CCD backs.
Yet, now this comes along and I see amazing usage for both interiors and exteriors of Architecture.
One thing that would be interesting is how we can actually capture people in use of the spaces which is a request of many many architectural clients.
With the already great dynamic range of this sensor, I feel like this might make my 'lighting' bag lighter...
I'm really looking forward to this download next week.
And I'm really looking forward to seeing what images you make with it, and sharing where it does and does not help you achieve your visual goals!
-
We should have a couple new raws to share today. Our team was out happily shooting away this weekend!
-
Doug, we can see how excited you are at frame averaging but here are my comments:
—I think you should avoid the phrase "exposure pyramid" (or whatever pyramid phrase you use) because, although pyramids have four faces, they usually look like triangles. I think "exposure square" or "exposure rectangle" more accurately conveys the fourth dimension that you refer to and does so in the same two-dimensional space that "exposure triangle" is in. A pyramid is a three-dimensional object and quite different to visualise than two dimensional objects like squares/rectangles and triangles. I am being picky I know!
—I don't think the sample images at phaseoneiq4.com show any clear advantage between normal and frame-averaged images. Looking at the first sample in your article, the one of the grassy promontory going out into the sea, the left side is the non-averaged shot and the right shot is the averaged one. As far as I can see, the single frame shot is more pleasingly brighter and more "alive" on the grass and the cliffs than the averaged one. Do you agree?
—I can see an advantage in the7-minute averaged shot of the two buildings with blue roofs on a river but I think that is because darkening the exposure gives more contrast and apparent sharpness in this case.
—Can you publish the histograms for each of the shots to show the differences between them?
—Hopefully, Phaseone will publish a thorough instruction guide?
Roger
-
My seven-year-old Nikon D800 does this, more or less. I use the feature a lot.
The world does not need more blurry-water pics.
-
Sure it does, much preferred to a stop action freeze unless there is something else going on in the image, kayak, canoe, whitewater rafting, etc.
Much more pleasing to the eye, at least for me. But everyone has their own styles for sure.
Paul C
-
Doug, we can see how excited you are at frame averaging but here are my comments:
—I think you should avoid the phrase "exposure pyramid" (or whatever pyramid phrase you use) because, although pyramids have four faces, they usually look like triangles. I think "exposure square" or "exposure rectangle" more accurately conveys the fourth dimension that you refer to and does so in the same two-dimensional space that "exposure triangle" is in. A pyramid is a three-dimensional object and quite different to visualise than two dimensional objects like squares/rectangles and triangles. I am being picky I know!
Point taken! It's definitely an imperfect analogy. Hard to figure out what metaphors and phrasing to use for something that fundamentally diverges from traditional photographic exposure theory.
—I don't think the sample images at phaseoneiq4.com show any clear advantage between normal and frame-averaged images. Looking at the first sample in your article, the one of the grassy promontory going out into the sea, the left side is the non-averaged shot and the right shot is the averaged one. As far as I can see, the single frame shot is more pleasingly brighter and more "alive" on the grass and the cliffs than the averaged one. Do you agree?
I don't agree at all. Look in smooth surfaces at high magnification and the difference in noise is profound.
Since the ISO 50 frame is already very good (the IQ4 150mp at ISO 50 is already the best standard-capture raw file you can get), the difference is only meaningful in the very deep shadows. But the "very deep shadows" come up frequently in landscape and architectural photography.
—Can you publish the histograms for each of the shots to show the differences between them?
This implies you haven't downloaded the raws, which goes a long way to explaining why you missed the difference in the shadow quality, which, IMO, is not subtle.
Go the IQ4 frame averaging raws (https://www.dtcommercialphoto.com/the-phase-one-iq4-now-has-automatic-frame-averaging/) and if you want to see the as-captured histogram just create a New Variant.
—Hopefully, Phaseone will publish a thorough instruction guide?
Couldn't speak to that. But DT will certainly be eager to discuss this feature with any and all of our clients who want to use it, both so we can share what we know and so we can learn what they have to teach us!
-
Feature Update 1 is now live!
https://www.phaseone.com/en/SupportMain/Camera-Firmware/IQ
Our support team is working until 6:30pm ET today in the case any of our clients wish to update and have questions.
A reasonable warning: any firmware update of any device (phone, camera, computer etc) carries a very small risk of “bricking” that device. Of course every IQ4 comes with a five year warranty that includes a loaner during any repair or service, but with the holiday weekend upon us I would discourage anyone from updating if they have an important production/job/shoot this weekend. Or if you update, do it early today so we have time to assist you in the (rare) chance of any issue.
Happy shooting everyone!
-
Our support team has the following detailed write up of the install process:
https://www.dtcommercialphoto.com/updating-your-iq4-with-the-creative-control-firmware/
-
Doug, we can see how excited you are at frame averaging but here are my comments:
—I think you should avoid the phrase "exposure pyramid" (or whatever pyramid phrase you use) because, although pyramids have four faces, they usually look like triangles. I think "exposure square" or "exposure rectangle" ...
It’s the exposure tetrahedron! See https://en.m.wikipedia.org/wiki/File:Tetrahedron.gif
-
It’s the exposure tetrahedron! See https://en.m.wikipedia.org/wiki/File:Tetrahedron.gif
My favorite shape!
-
Update: the max frame and maximum duration were both raised for the production version of the firmware. With an IQ4 150mp the max frame count is 3600 and the maximum total duration is 2 hours. This is a major improvement from the original 1000 count and 12 minutes originally expected.
-
What impact does frame averaging have on the 'count' of the 4150? Is only the resulting raw registered in the count or all frames?
Victor
-
Each Frame Average session, is counted as (1) actuation. No matter the total number of frames used.
Paul C
-
Thanks much, Paul. I only asked as I do not have my 4150 yet...... maybe next week.
Happy 4th......
Victor
-
Hi Victor,
I wondered the same thing, especially when P1 started talking about 3600 frames possible in a single average.
That would run up the count a bit!.
Looking forward to your feedback on the new back.
Paul C
-
Anyone have any results to share, positive or negative, of their own testing of the new firmware?
-
Anyone have any results to share, positive or negative, of their own testing of the new firmware?
Not from me. I was hoping that it might allow in-camera multiple exposures (just two or three shots) but I now realise this is not the point or a capability of the feature.
I now appreciate the feature has two benefits: you can take slow exposures to blur moving objects, such as, water and clouds, without using a neutral density filter or incurring colour casts and you can neutralise any “random captured noise” (quote from P1’s release note), presumably, relevant where you have shadows in your image that you cannot expose for as much as you would like.
These benefits are only available if the objects of your image that you want to be sharp and your camera remain absolutely rigid during the entire image-taking process (unless you want ghosting or ICM).
Roger