[texandrews]

I did a quick check on the US Consumer Price Index Inflation Calculator for the K-1 Mark II list price in the US in 2018 ($1,999) to see the theoretical price point for the same camera if newly released as of August 2022 after adjusting for inflation. I did this just to get a rough idea of what Ricoh's current list price for the same camera might be if newly released today. The price would $2,345. A K-1 Mark III would have advancements that would presumably cost more to produce, so I don't think a release price of $2,500 to $3,000 a year or three from now would be an unreasonable price.

And I think that everyone should remember that inflation is now higher than it has been in a long time. We've gotten used (in many places) to pretty nominal inflation rates. But another element of this is that inflation this time has to be parsed pretty closely; see Paul Krugman's opinion piece from today's NYT for instance about a major inflation driver in the U.S. right now. Other places will have their own stories. Good luck you guys in the UK! My sympathies to those of you in Turkey, Iran, Argentina....don't want to get political here, and monetary policy is inherently political and the drivers of inflation can sometimes be directly traced back to it, but sometimes it's something else. Just pointing out that the whole world is swimming around in a sea of problematic financial issues.

---

Last edited by texandrews; 10-02-2022 at 09:48 AM.

[repaap]

Money is one thing. As discussed for many pages now. Everyone knows what next k-1III need to be, should they be happy with the price or not.


you want more advanced body with K-3III AF and astrotracer and touchs screen and joystick and 100+ AF points and same RGBiR sensor for AF.

Bigger OVF not needed.
More MP not needed, but seems necessary since tech is what it is.
4K (or even bigger video) not needed.(or this should change to be something really worth while, which would meand faster readout, thinking of heating up and longer record time)

I'd buy it. sadly economy is biting me too so not early adopter this time...(and I'd need to sell my K-3III in order to buy K-1III also, but it is what it is) But I'd definetly buy K-1III. I'v sold k-1 original. Mainly because of my need of better AF. Sure it was beaten up and some scratch here and there. Still images are great, just if AF would work more reliably. One reason to not to buy K-1II, and important one too, for me.

K-3III is very good, because of AF is so much better, and files are very nice too. I can trust it and enjoy photography more because of that. OVF does not make me miss FF so much and files does keep up very well even compared to FF sensors in 24-26 MP range. My DFA*70-200 is crying after FF sensor thou(very nice AF with K-3III BTW) and FA 31 and 77 and...yeah. But it is just me...I can wait, and obviously I will.

[Fogel70]

Of course he would, you can't argue with physics, Fogel, a subject I used to teach for a living.

The amount of shot noise depends on and only on the number of photons that have hit the individual pixel. The longer your exposure, the more photons:

It is the total amount of photons captured by the sensor that is important for images, not the photons each pixel capture. Just like you say one way to collect more photons is using longer exposure. Another way for collecting more photons is using a larger sensor.

Technology can only really change the much lower thermal noise contribution, or in the case of baked RAW files, by doing some sort of averaging out adjacent pixels which also destroys details - why when you push the noise reduction slider the person in the photo now looks like a wax dummy.

One thing BSI improved was fill factor for smaller pixels, so that the effeciency between larger and smaller pixels became much smaller. Although this makes most difference for really small sensors, like those used in smartphones.

DXOMark have almost fraudulent habits like 'normalizing', but if you go to their screen rather than print measurement graphs, that normalizing isn't done, so you can see that the old Pentax K-5 II with its 16Mp of big pixels still actually has better noise performance than the modern Sony A7 IV R, which has 61Mp of smaller ones:

If you would care about images instead of pixels you would see that Sony A7 IVR has lower level of noise in comparison. Normalization is always done if you want to compare stuff. You print them to the same size or look at the full size on same sized screen. If you always pixel peep images at 100% then the screen comparison could be useful for you, but it is not a useful comparison.

[texandrews]

Money is one thing. As discussed for many pages now. Everyone knows what next k-1III need to be, should they be happy with the price or not.


you want more advanced body with K-3III AF and astrotracer and touchs screen and joystick and 100+ AF points and same RGBiR sensor for AF.

Bigger OVF not needed.
More MP not needed, but seems necessary since tech is what it is.
4K (or even bigger video) not needed.(or this should change to be something really worth while, which would meand faster readout, thinking of heating up and longer record time)

I'd buy it. sadly economy is biting me too so not early adopter this time...(and I'd need to sell my K-3III in order to buy K-1III also, but it is what it is) But I'd definetly buy K-1III. I'v sold k-1 original. Mainly because of my need of better AF. Sure it was beaten up and some scratch here and there. Still images are great, just if AF would work more reliably. One reason to not to buy K-1II, and important one too, for me.

K-3III is very good, because of AF is so much better, and files are very nice too. I can trust it and enjoy photography more because of that. OVF does not make me miss FF so much and files does keep up very well even compared to FF sensors in 24-26 MP range. My DFA*70-200 is crying after FF sensor thou(very nice AF with K-3III BTW) and FA 31 and 77 and...yeah. But it is just me...I can wait, and obviously I will.

Agree with some of what you say, and if this are just your desires, then no argument. But I do think that as of this date acceptable 4K video is important. Acceptable means at least apsc sized (remember the k3III is cropped down from apsc), decent record time, touch screen focus, better AF; a joystick would be nice. It doesn't have to be the best AF in the industry for video, because Pentax's lenses are probably a touch slower, but it should be improved, and tracking should be sure. I'd be willing to bet more general potential customers would be more interested in video than astro improvements. Pentax arguably leads here anyway.

[chd]

It is the total amount of photons captured by the sensor that is important for images, not the photons each pixel capture. Just like you say one way to collect more photons is using longer exposure. Another way for collecting more photons is using a larger sensor.

Not so...

I recommend taking the time to watch this. It is both interesting and illuminating, if you'll forgive the expression. Yes, it is a bit over six years old, but as far as I know they haven't released an updated version of physics since then (at least not the sections relevant to digital photography), so it's all still valid. Pixel size is specifically addressed around the 8 minute mark, but really the whole thing (and the whole series) is worth watching. Noise and signal to noise ratio, very relevant to dynamic range, at about the 54 minute mark.

lecture8-13apr16

-C

[WillyKano]

I know it is a waste of time to mention but here are my suggestions for the Mark III:
- in the menu for multiple exposures the ability to change focus between each picture, to do focus stacking; and I know that there are some weardo’s that have already implemented this in an app;
- as a bird photographer I would appreciate a quicker start of the K1 from standby, and a better focusing system, but dreaming of a focusing system like Canon has, which is following the eye of the photographer to focus on the subject he is looking at, would not be possible in the pricing of a K1;
- more pictures per second, with accurate focusing on the moving objects;
- possibility to generate the name of the pictures like YYYYMMDD_sequencenr; or at least the name of the map like YYYYMMDD.
But nevertheless a K1 is a great picturemaker!

[Fogel70]

Not so...

I recommend taking the time to watch this. It is both interesting and illuminating, if you'll forgive the expression. Yes, it is a bit over six years old, but as far as I know they haven't released an updated version of physics since then (at least not the sections relevant to digital photography), so it's all still valid. Pixel size is specifically addressed around the 8 minute mark, but really the whole thing (and the whole series) is worth watching. Noise and signal to noise ratio, very relevant to dynamic range, at about the 54 minute mark.

lecture8-13apr16

-C

And the data he used in his charts is even older. Cameras like Canon 5D and 20D are not that useful to use today for a noise performance. The data he used in the presentation seem to have been from 2008 and earlier.

Today there are more variation on pixel size for larger sensors than it has been in the past, and noise performance do not vary much within same sensor size with different sized pixels, than it does between sensor sizes. The variation on pixel size on larger sensors just have very little contribution to the overall noise in the images they produce.

Today with BSI and dual native ISO sensors make the noise performance completely different than on older sensors.
Cameras with more pixels tend to get more dynamic range on the images if comparing within same sensor size.

The main advantage of larger pixels nowadays are only in the extreme high ISO where sensor readout noise start to contribute a lot to the overall noise.

Edit: In this presentation it seems to be "larger pixels = larger sensor", because all the comparisons seems to be between smaller sensors with smaller pixels vs larger sensors with larger pixels. FI comparison between Canon S70 vs Canon 1D mk II, a 7MP 1/1.8" sensor in Canon S70 vs 8MP APS-H sensor in Canon 1D mk II. Another comparison was between Nokia N95 smartphone and Canon 5D mk II. As far as I can see there is no comparison between different pixel size on same sized sensors.

---

Last edited by Fogel70; 10-03-2022 at 04:20 AM.

[chd]

And the data he used in his charts is even older.

As I said, which you missed or are ignoring, the underlying physics has not changed. You're looking for a specific example and missing how it actually works.

[Fogel70]

As I said, which you missed or are ignoring, the underlying physics has not changed. You're looking for a specific example and missing how it actually works.

I think you ignore the case of using more smaller pixels vs less larger pixels. They pretty much cancel out each other on image level on same sized sensors.

Have you actually looked at images from cameras like Sony A7s III with 12MP vs Sony ATR IV with 60 MP at IS= 6400 - 25600 and see how little difference there is in noise?


Edit: You can also check the ClarkVision web page that the data above in the video you posted are from.
https://clarkvision.com/articles/does.pixel.size.matter/

Their conclusion at the end...

When choosing between cameras with the same sized sensor but differing pixel counts, times have changed. A decade ago, I would have chosen the camera with larger pixels (and fewer total pixels) to get better high ISO and low light performance. Today I would choose the higher megapixel (thus smaller pixels). Modern cameras with high megapixel count, low read noise and low electronics noise allow one to trade resolution and noise. If one wants reasonable dynamic range in a high megapixel camera, the pixels must still be large enough to hold enough photoelectrons to give the dynamic range. Currently (circa 2016), that is not much smaller than 4 micron pixel spacing. For example, the 50 megapixel Canon 5DS(r) fits this criteria, and Nikon's D800 and D810 are in the same league.

[Racer X 69]

Photons?

You Tube

You Tube

[repaap]

Agree with some of what you say, and if this are just your desires, then no argument. But I do think that as of this date acceptable 4K video is important. Acceptable means at least apsc sized (remember the k3III is cropped down from apsc), decent record time, touch screen focus, better AF; a joystick would be nice. It doesn't have to be the best AF in the industry for video, because Pentax's lenses are probably a touch slower, but it should be improved, and tracking should be sure. I'd be willing to bet more general potential customers would be more interested in video than astro improvements. Pentax arguably leads here anyway.

Yeh, it was more like my own preferences. Acceptable video is needed now a days. I have K-3III and it has okay video. And it should be absolute minimum. They should still get that readout faster. also if you have just 7 minutes of 4K it is a bit lame. Sure you can use it and get okay result, but it is FULL HD 1080 camera. even if it says 4K.


I'm shooting video with different system, I'm going to buy one more professional camera for video this time, that won't be able to take photos, just video...


I'm shooting photos with Pentax K-3III because I do think that it is very good camera. K-1 would be one heck of a camera if it had K-3III tech -even if it would still shoot 5 FPS-. As long as keeper rate would be 9 out of 10, like it is with k-3III. Eye AF and ability to find and track human and actually animals, even if it is not written in specs (animal eye af), is very nice. Could be better still, but it is huge step in right direction..


joystick. it is surprisingly good and touch screen is actually very nice on K-3III. I had some doubts and actually refused to use them at first, since I was accustomed to shoot with my Pentax at sertain way...but slowly it was growing on me. I'd like to have 'moonlander', but if that would mean that there will not be touch screen...it would become complicated.

but that is just me.

[Ian Stuart Forsyth]

Of course he would, you can't argue with physics, Fogel, a subject I used to teach for a living.

The amount of shot noise depends on and only on the number of photons that have hit the individual pixel

This is about as nonsensical as arguing that one should measure the speed of an automobile based on the revolutions of a wheel regardless of the diameter of the wheel, sure we could have 2 cars can travel at the same wheel RPM but the speed at which they travel will be different if these cars had different wheel sizes.

When evaluating a SNR you always have to understand the bandwidth it represents. Looking at a very narrow bandwidth will decrease your SNR. Looking at the SNR across a larger bandwidth will increase your SNR just because you are looking at the SNR with a narrower bandwidth does not mean that you have captured less SNR it just means that you did not have the acuity to see the variations when sampling while only using a wide bandwidth. With an imaging sensor we are able to capture all the data that falls withing the bandwidth

With imaging sensor with greater resolutions, one is able to view the image at narrower bandwidths, but this does not mean we lose SNR based on more noise, if we view that image based with a set bandwidth or resolution that signal represents very different areas within the image. the one viewed at a narrower bandwidth represents only a small portion of an image with less impact while the other when viewed with a wider bandwidth represents much more of the image with a greater impact on the image .

---------- Post added 10-03-2022 at 08:10 PM ----------

Technology can only really change the much lower thermal noise contribution, or in the case of baked RAW files, by doing some sort of averaging out adjacent pixels which also destroys details

How is combining adjacent pixels and losing detail any different than not capturing the detail in the first place? We at not really combining pixels when we normalize, we are placing the recorded data in the area that data represents (a smaller area)

---------- Post added 10-03-2022 at 08:21 PM ----------

The amount of shot noise depends on and only on the number of photons that have hit the individual pixel.

You do realize that the goal of one of the sensor designers I have been following has the goal of only recording were a single area who is hit by a single photon strikes that sensor, have any guesses as to how and what he predicts the sensors performance will be?

A hint he has been working on it for a few years with much more success than what is presently available. One photon per one photo site or Joint as he calls them

[clackers]

It is the total amount of photons captured by the sensor that is important for images, not the photons each pixel capture.

Completely incorrect.

A RAW image is like a spreadsheet.

It is simply a list of the individual values of each pixel.

Couldn't be further wrong!

That is also why you can never 'see' a RAW image, only a JPG or similar preview of it after processing.

All pictures are captured as pixels, and they are rendered on paper or on screen as pixels.

[clackers]

Not to be contrarian, but normalizing is helpful. I look at it two ways. First of all, it makes sense that you would set a target print size that you will compare all sensors at.

You just don't need that extra guesswork, Rondec, because SNR is a ratio. 18% on the DXOMark test means 18% of the pixels are noise, whatever their size, whatever the sensor they're on.

Extra pixels don't get you any more light or lower noise, they just help software noise reduction in postprocessing, by averaging.

[clackers]

With imaging sensor with greater resolutions, one is able to view the image at narrower bandwidths, but this does not mean we lose SNR based on more noise

?

See Post 134.

Increase the resolution, shot noise goes up.

'Bandwidth'?

Why did you call ... this ... any of this ... 'bandwidth', Ian?