Phase One IQ4 - Feature Update 1

[Audii-Dudii]

Cool. wondering if you could add 5 stops of ISO (so 1600), cut the exposure to only 1 minute if the results would be nearly identical. should be enough frames for averaging to eliminate the noise as well.

As noted in Post No. 162 above, Bill Claff's testing at Photons to Photos shows the IQ4 loses more than three stops of photographic dynamic range when used at ISO 1600 instead of ISO 50 -- http://photonstophotos.net/Charts/PDR.htm#Phase%20One%20IQ4%20150MP -- which I suspect will have a noticeably negative effect on this photo regardless of how many frames are averaged to reduce noise.

 

[stngoldberg]

Since Phase one no longer supports my Hasselblad H5X and my five HC lenses, I can’t upgrade to IQ4; however the IQ3100 can be frame averaged in photoshop and along with shadow recovery presents a respectable image
Stanley

 

[dougpeterson]

As noted in Post No. 162 above, Bill Claff's testing at Photons to Photos shows the IQ4 loses more than three stops of photographic dynamic range when used at ISO 1600 instead of ISO 50 -- http://photonstophotos.net/Charts/PDR.htm#Phase%20One%20IQ4%20150MP -- which I suspect will have a noticeably negative effect on this photo regardless of how many frames are averaged to reduce noise.

The underlined part of your post is not correct.

As seen in our frame averaging of 60 frames at ISO 25,600, frame averaging has a profound impact on the noise/dynamic range of any given ISO.
https://www.dtcommercialphoto.com/w...akes-it-the-most-average-camera-in-the-world/

Cool. wondering if you could add 5 stops of ISO (so 1600), cut the exposure to only 1 minute if the results would be nearly identical. should be enough frames for averaging to eliminate the noise as well.

I don't think there would be a benefit of increasing the ISO. It seems to me to still thinking be thinking in terms of the photographic triangle. Frame averaging adds an entirely new dimension to that triangle. If one wished to cut the total exposure time (e.g. so one is not standing in the dark for five minutes; or so that you can go make other images), one can just cut the total exposure time; no need to change shutter speed, aperture, or ISO.

He originally opted for five minutes of averaging, which translates to around 150 frames that are averaged into one raw. Using one minute of averaging you'd still get 30 frames of averaging, which, given the really low noise of a single IQ4 150mp raw file, would still have very high quality shadows.

Now whether 30 frames at ISO50 / 2 sec [one minute total] would be better than 60 frames at ISO100 / 1 sec [one minute total] or 120 frames at ISO200 / 0.5 sec [one minute total] would be best is not something I've yet tested; that could be interesting. But my gut says it's still best to use the lowest ISO practical/possible for the base exposure.

 

[Audii-Dudii]

The underlined part of your post is not correct.

As seen in our frame averaging of 60 frames at ISO 25,600, frame averaging has a profound impact on the noise/dynamic range of any given ISO.
https://www.dtcommercialphoto.com/w...akes-it-the-most-average-camera-in-the-world/

For the record, I don't have a horse in this race, so I don't care about the outcome either way. Also, I have been blending multiple files as a noise-reduction technique for nearly a decade (starting with my P30+, in fact) and very much endorse its use when the circumstances allow.

That said, though, none of the test results discussed at the link you cited are apples-to-apples relevant to the proposed test, which is to compare frame-averaged photos taken at ISO 50 and 1600 with the exposure length adjusted as appropriate, not to compare a single-frame photo taken at ISO 50 to a frame-averaged photo taken at ISO 1600.

Because whatever image quality improvements are achieved via frame-averaging will surely apply to both photos and there is no way I can think of that a photo taken at ISO 1600 can close the IQ gap to a similar photo taken at ISO 50, hence my comment (and your gut response noted in your subsequent comments.)

Again, I'm not trying to poop in front of this parade, because I think this is a wonderful tool to have available, only to point out that it's not the magic bullet / universal solution for every photographic situation that some photographers want it to be...

 

[etrump]

The underlined part of your post is not correct.

As seen in our frame averaging of 60 frames at ISO 25,600, frame averaging has a profound impact on the noise/dynamic range of any given ISO.
https://www.dtcommercialphoto.com/w...akes-it-the-most-average-camera-in-the-world/

This is dramatically noticeable flipping through the frames on the back with exposure clip warnings on. Even at ISO 200 there are both highs and lows recovered in the frame averaged exposure. (Street lights and corner shadows)

 

[dougpeterson]

For the record, I don't have a horse in this race, so I don't care about the outcome either way. Also, I have been blending multiple files as a noise-reduction technique for nearly a decade (starting with my P30+, in fact) and very much endorse its use when the circumstances allow.

That said, though, none of the test results discussed at the link you cited are apples-to-apples relevant to the proposed test, which is to compare frame-averaged photos taken at ISO 50 and 1600 with the exposure length adjusted as appropriate, not to compare a single-frame photo taken at ISO 50 to a frame-averaged photo taken at ISO 1600.

Because whatever image quality improvements are achieved via frame-averaging will surely apply to both photos and there is no way I can think of that a photo taken at ISO 1600 can close the IQ gap to a similar photo taken at ISO 50, hence my comment (and your gut response noted in your subsequent comments.)

Again, I'm not trying to poop in front of this parade, because I think this is a wonderful tool to have available, only to point out that it's not the magic bullet / universal solution for every photographic situation that some photographers want it to be...

I think we agree on everything here.

 

[JimKasson]

Now whether 30 frames at ISO50 / 2 sec [one minute total] would be better than 60 frames at ISO100 / 1 sec [one minute total] or 120 frames at ISO200 / 0.5 sec [one minute total] would be best is not something I've yet tested; that could be interesting. But my gut says it's still best to use the lowest ISO practical/possible for the base exposure.

The amount of photon noise will be the same for all three cases. However, to the degree that the sensor is not ISOless, the frame-to-frame uncorrelated component of the read noise will be less in the higher ISO cases. The IQ4 150 is not ISOless, since it uses the Sony-licensed Aptina DR-Pix conversion gain technology.

http://www.photonstophotos.net/Charts/RN_e.htm#Phase One IQ4 150MP_16

So there will be some advantage to going to ISO 400 or so, to take advantage of the increased conversion gain and consequently lower input-referred read noise. However, this benefit will only be apparent in the deepest shadows.

Jim

 

[Bill Caulfeild-Browne]

Although I updated my XF/150 a week ago, I hadn't had a chance to play with it yet, so here's my first crack at seeing what it can do with noise.

First, the original scene exposing for the highlights at ISO 3200. C1 defaults.

At 100%,the shadowed rear wheel and black tire, pushing the shadow recovery to maximum.

Using the Frame Averaging - I think this was 3 seconds, 11 shots at the same ISO 3200.

There's a full moon tonight so I'm going to see how it can handle really low light! (if you hear howling, it may be me.)

 

[algrove]

Bill
Thanks for this.
Just to be clear you used the same ISO and f stop for before and after. What was your exposure time for the before?
Thanks as I'm trying to get a feel for adjustment time when using frame averaging.

 

[Boinger]

Bill
Thanks for this.
Just to be clear you used the same ISO and f stop for before and after. What was your exposure time for the before?
Thanks as I'm trying to get a feel for adjustment time when using frame averaging.

Exposure time does not change.

You take the shot as you would when taking a normal shot. So if the scene calls for f/8 at 1/15s iso 100.

You shoot it like that but just turn on frame averaging and select the number of frames you want to average.

The length of time for the averaging / blur is entirely your judgement.

 

[dougpeterson]

Exposure time does not change.

You take the shot as you would when taking a normal shot. So if the scene calls for f/8 at 1/15s iso 100.

You shoot it like that but just turn on frame averaging and select the number of frames you want to average.

The length of time for the averaging / blur is entirely your judgement.

Just a quick note that if the light is at all changing, or the scene is moving, it behooves you to aim a bit under for exposure when frame averaging. The IQ4 live RAW histogram is a great tool to do so.

 

[Bill Caulfeild-Browne]

Bill
Thanks for this.
Just to be clear you used the same ISO and f stop for before and after. What was your exposure time for the before?
Thanks as I'm trying to get a feel for adjustment time when using frame averaging.

Both shots were identical - ISO 3200, shutter 1/4000, F14. I'd have gone with a higher ISO but it was too bright.

I note, however, that the IQ4-150 is essentially noiseless even at ISO 3200 IN PRINT. You can see the noise on screen, but it's barely noticeable in print.

 

[algrove]

Both shots were identical - ISO 3200, shutter 1/4000, F14. I'd have gone with a higher ISO but it was too bright.

I note, however, that the IQ4-150 is essentially noiseless even at ISO 3200 IN PRINT. You can see the noise on screen, but it's barely noticeable in print.

Why did you not just use say ISO50? Too long an exposure?

 

[Bill Caulfeild-Browne]

Why did you not just use say ISO50? Too long an exposure?

I did. In fact I tried several different ISOs but my purpose was to see what I could recover from higher ones.
Lower ISO's are even better because the dynamic range is wider.