X-T5 at 12,800 ISO

Either is an objectionable artefact and people are kidding themselves that they like it.

JNR wrote:

Erik Baumgartner wrote:

baobob wrote:

and if reducing the noise is important expose to the right ETTR with +2/3 to 1 EV then in PP adjust highlights.

Actually, as Fuji sensors are quite ISO invariant in that range, so long as you are maximizing the sensor exposure, it doesn’t much matter where you set the ISO beyond 1600 or so. You can shoot at -3EV (ISO 1600 instead of 12800) and just push the brightness where it needs to go in post - no significant noise penalty, and you get significantly more highlight headroom to work with. ETTR is really only relevant at base ISO.

Agreed about everything here other than the last sentence. This week I got an opportunity to shoot at a club where the lighting has been pretty much the same as it had been for the past decade (very dark) and so got to do some good comparisons going back not only to the X-T2/20 shots, but earlier Pentax shots (on the 24mp sensor).

To the extent that I had "maximized sensor exposure" correctly, the ultimate processing in the 2000 to 6400 range from the X-T5 showed less noise and equal or slightly better sharpness. Of course, that involved a bit less sharpening and more NR than with the older sensor. A few times the sensor got tricked and lowered the ISO too far (yes, I use auto ISO in these variable light situations) - and the lack of maximized exposure resulted in slightly greater noise issues than with the old sensor. So, proper exposure is more critical with the 40mp sensor - as there seems to be more of a distinct shadow cliff.

I don't think I mentioned anything about image quality of the old vs. new sensors. From my experience playing with the 40MP RAW files, the IQ seems to be quite excellent when processed optimally (meaning not the same as the 24/26MP files).

It is likely that Erik and I have somewhat different understandings of ETTR... but to me it really does mean "maximizing the sensor exposure" and that should include low-light situations when you aren't filling the well.

I don't think there are multiple ways to understand it. ETTR - exposing to the right, means exactly that, exposing to the right.

It is always advantageous to maximize sensor exposure (SS/Aperture), but at anything beyond base ISO, while you still want to expose the sensor as much as possible, this isn't really ETTR anymore as you are now ISOing to the right (AKA "histogramming" to the right) which fundamentally changes the equation.

Exposing to the right is all about maximizing the signal to noise ratio and dynamic range. Recording the brightest highlights just below clipping at base ISO (ETTR) does exactly that. ISOing to the right, on the other hand, does not. While highlights to the right is generally a good idea, and it might may very well be appropriate to do this in most situations (and certainly is for jpeg shooters at DR100), by doing that in high ISO/high DR situations you are significantly limiting the dynamic range that can be captured. Exposing to the right always guarantees the maximum possible DR, but ISOing to the right does not, at higher ISOs, you can easily be sacrificing highlight headroom unnecessarily (and undesirably).

baobob wrote:

Agree

Maximising exposure sensor = what exactly ??

Maximizing exposure - the lowest SS that won't cause motion blur issues along with the maximum aperture that won't cause DOF/sharpness issues.

What I do with the XH2 in dim light is having EV +2/3 to 1. Is it ETTR or maximization of sensor exposure ??

In low light, by shooting at EV +2/3 to 1 stop, you are most likely increasing the ISO, not the exposure (not ETTR). In doing so, you are limiting the highlight headroom and not improving the noise significantly (noise is primarily an exposure thing or rather, the lack thereof). In the ISO invariant range of the sensor (usually around ISO 1600+) there is no advantage to setting at high ISO in-camera vs. pushing the brightness in post. See response to JNR below.

OTH I am not sure as well that the 40 mpx sensor is as ISO invariant as the 26 mpx is ??

I've seen no evidence to suggest this.

a_c_skinner wrote:

Either is an objectionable artefact and people are kidding themselves that they like it.

I think "objectionable artefact" is in the mind of the beholder, and for the rest I'm pretty sure that you wouldn't want someone else's preferred aesthetic enforced onto you.

a_c_skinner wrote:

Either is an objectionable artefact and people are kidding themselves that they like it.

Furthermore, you need look no further than the top photojournalism sports photo from 2022 currently the featured posted image on DPR for the various awards listed... taken in my beloved Williams Arena:

https://www.dpreview.com/sample-galleries/5589856725/best-of-photojournalism-awards-2022/1257728822

What were those judges thinking - who cares about the image when so much noise is apparent? (Sarcasm for those who are wondering).

JNR wrote:

a_c_skinner wrote:

Either is an objectionable artefact and people are kidding themselves that they like it.

I think "objectionable artefact" is in the mind of the beholder, and for the rest I'm pretty sure that you wouldn't want someone else's preferred aesthetic enforced onto you.

Exactly. The post you replied to sounds a bit like - this is “objectively” better so there is no alternate way of looking at it. Photography being an art form, I vehemently disagree.

As for analog vs digital noise, for me analog is more organic, whereas digital is more like fine sand being sprinkled over the photo. I'm talking about sensor noise, not grain added in post (in-camera or on a pc). I like both and they both have their place. And there is also a place for grainless photos. For character photography, I really like analog grain. But digital sensor noise I really like for adding a touch of dreaminess to landscape.

There are so many photographical techniques. It would be a shame to reduce yourself to in-focus grainless shots.

PeterLeyssens wrote:

JNR wrote:

a_c_skinner wrote:

Either is an objectionable artefact and people are kidding themselves that they like it.

I think "objectionable artefact" is in the mind of the beholder, and for the rest I'm pretty sure that you wouldn't want someone else's preferred aesthetic enforced onto you.

Exactly. The post you replied to sounds a bit like - this is “objectively” better so there is no alternate way of looking at it. Photography being an art form, I vehemently disagree.

As for analog vs digital noise, for me analog is more organic, whereas digital is more like fine sand being sprinkled over the photo. I'm talking about sensor noise, not grain added in post (in-camera or on a pc). I like both and they both have their place. And there is also a place for grainless photos. For character photography, I really like analog grain. But digital sensor noise I really like for adding a touch of dreaminess to landscape.

There are so many photographical techniques. It would be a shame to reduce yourself to in-focus grainless shots.

Admittedly, I am slightly changing the topic because my purpose in adding grain at the time of conversion is to mitigate the unpleasant aspects of digital grain and allow the viewer to concentrate on the image rather than the artefacts. I'm not an advocate for grain to add an aesthetic, but still can appreciate on it the rare occasion when it seems to work well. I'm especially not a fan of smeared details for the sake of attempted grainlessness. That never works for me.

I'm thinking of this because I've had to use added grain much more often on the processed high-ISO X-T5 images because the noise pattern is inconsistent. This does appear to be a slight failing of C1 processing (as opposed to the sensor itself) because the OOC JPEGs (which are comparatively noisy and not quite as sharp) have a far-more consistent pattern. Not that it is a huge difference, but it does trigger the need to add a bit of processed grain for overall consistency.

Objective and objectionable are a long way apart.  Art does all sort of things and isn't amenable to objective, but we can all say what we find objectionable.  It is in the nature of objectionable that it is mostly personal.

JNR wrote:

PeterLeyssens wrote:

JNR wrote:

a_c_skinner wrote:

Either is an objectionable artefact and people are kidding themselves that they like it.

I think "objectionable artefact" is in the mind of the beholder, and for the rest I'm pretty sure that you wouldn't want someone else's preferred aesthetic enforced onto you.

Exactly. The post you replied to sounds a bit like - this is “objectively” better so there is no alternate way of looking at it. Photography being an art form, I vehemently disagree.

As for analog vs digital noise, for me analog is more organic, whereas digital is more like fine sand being sprinkled over the photo. I'm talking about sensor noise, not grain added in post (in-camera or on a pc). I like both and they both have their place. And there is also a place for grainless photos. For character photography, I really like analog grain. But digital sensor noise I really like for adding a touch of dreaminess to landscape.

There are so many photographical techniques. It would be a shame to reduce yourself to in-focus grainless shots.

Admittedly, I am slightly changing the topic because my purpose in adding grain at the time of conversion is to mitigate the unpleasant aspects of digital grain and allow the viewer to concentrate on the image rather than the artefacts.

Have you non-photography friends really noticed artifacts in your photos?

JNR wrote:

I'm not an advocate for grain to add an aesthetic, but still can appreciate on it the rare occasion when it seems to work well. I'm especially not a fan of smeared details for the sake of attempted grainlessness. That never works for me.

No, right. I don't mind grain if it's part of how the photo was made. If somebody adds it for effect in post, it feels like Instagram filters.