https://fdn.gsmarena.com/imgroot/reviews/25/google-pixel-10-pro-xl/camera/gsmarena_022.jpg

Currently it even struggles with buildings, something well known for straight shapes. Wouldn't be surprised if they're feeding a jpeg to the algo instead of raw, since the jpeg samples already exhibit worm-like shapes that shouldn't be there.

https://www.feistyduck.com/newsletter/issue_128_google_debuts_device_bound_session_credentials_against_session_hijacking

Google came out with a mostly cool idea that can eventually turn into a nuisance and impossibility of backing up your logged in sessions (in worst case scenario, as web seems much less vulnerable to monopoly than embedded hardware). From the look of it, currently a TPM whitelist isn't a part of the standard and I've seen none of such plans.

Keep in mind: TrustZone, TPMs and hardware authentication devices are a genuinely cool thing that can save your ass. But whitelists of what third party entities consider as accepted are never right and will prevent you from making improvements or tradeoffs you deem acceptable. We shouldn't need vulnerabilities for capabilities.

Watch DBSC with caution.

Being GMS-less and PIA-failing may force you to buy premium to create an account

*create an account or login (without an existing session/bound mail to avoid SMS)

From the good-to-end-user news, yes, Pixel 10 Pro exposed DCG (dual conversion gain) sensor mode, oddly enough as RAW_SENSOR rather than RAW12 (no quality difference, it's linear Bayer 12-bit at 12.5MP, but a difference for handling data on app side). This will make it much easier to spot regressions done in post processing, compared to sensor's own capabilities.
Contrary to general presumption, yes, smartphone photography is mostly software-defined, but in the opposite sense – software is what mostly degrades it, since a single frame raw (especially with DCG) can often look better than a stacked result ruined by post processing. Stacking is great, but you know. Yes, the raw video with zero denoising recorded from main cam will now look dope. No, gcam doesn't support raw video.

In other words, if you shoot (especially in daylight) a raw in Open Camera and import it to Lightroom Mobile, you're likely to be able to get a better result than stacked stock gcam jpeg. Gcam raws will probably be better than single frame, but Google is known from messing with them in recent years.
Tldr: P10P is now pretty good for non-computational photography without mods as long as you shoot raws.

If someone looks for greatest non-computational photography *with mods*, find something with OV50K or LYT-828, but first figure out how to unlock bootloader... (Yes, I'm temporarily main-camera-centric for the sake of these 3 messages, please never ignore telephoto)

There's a dude on Reddit who must've never read this pdf nor anything remotely similar and insists that DCG is staggered, it's such a weird conspiracy hill to die on. Both DCG and staggered slow down the rolling shutter relatively to same clock utilized on normal mode, but DCG in comparison to staggered isn't affected by ghosting and requires no alignment. Big cameras started to catch up and you may find it at a name like DGO. But yes, for a matter of fact, most of the time you won't see DCG in stock camera apps, OEMs prefer staggered, DCG being exposed to 3p apps is an even rarer exception.

https://i.imgur.com/EkqnsEY.jpeg
If somehow the person behind this comparison happens to read this, just slide chroma denoising to zero. That's not how the raw looks. Incorrect noise profile in dng metadata will give you color leaks and hunting for SNR with any kind of denoising applied in an image editor or viewer is besides the point.

There's a certain quote by a certain disgraced historical person about repeating lies. Now that we have the professional lying machine, we get to see it function as a both way to make them and consume them. Isn't it great?!

If this slop happens to reach the wrong people and affect the market, the people of Türkiye know who to blame for the original AI slop post that started it all once their factory ROMs with provisioning capabilities won't be flashable anymore to help them skip the humongous IMEI tax :(

https://www.androidauthority.com/sony-xperia-monitor-and-control-camera-3593061/

Looks like Sony is completely uninterested in one thing their mobile branch really needs, which is reputation.

As you can see, Huawei is back in the West, even a literal (South) American uses their phone now!

This phone lacks Pro suffix. GN8 is capable of 50MP@24fps, 12.5MP@60fps and 4k120. An analog crop for video only reduces the needed bandwidth (and the needed internal clock) as you can see in the public spec. Knowing the usual drop of max frame rate coming from DCG, it's completely safe to assume that it's capable of 12-bit DCG with a 16:9 crop and PDAF at 30 fps. Most likely the same also applies to 12.5MP mode. After all, DCG exists primarily for realtime use.

Giving a massive benefit of doubt to the company that held it back from Pro models for few years is unnecessary.

In case one would really want to seek for technical justifications, these could be:
1. SNR dip – at a very high ratio, a certain part of dynamic range may suddenly be noisier – unlikely for the expected 4:1 ratio
2. Slower rolling shutter (and frame rate) than the max possible from the sensor in normal mode, usually around half – not a problem here
3. Somewhat higher energy drain, but negligible compared to the SoC – also not a problem here
4. Phase data is usually either from LCG or HCG read rather than combined, so if an ISP can't take advantage of two phase data outputs, use of contrast AF can be more frequent, as LCG has worse SNR, while HCG clips brighter areas – yes, this could suck, but Google controls their ISP sufficiently
5. Theoretically somewhat worse peak SNR at highest ISO ranges, as more circuitry is involved – this makes most sense on GN8, it's a small format and therefore the sacrifices can be more impactful, but at the same time it's worth to keep in mind that it's a DCG sensor rather than a DAG sensor or a sensor with neither, the sacrifice is already in design

But despite it all, wouldn't a more pragmatic explanation exist?
Just as in the first sentence of this message, non-Pro is non-Pro, it's likely to be an afterthought. If to trust what I've recently heard, the latest Pixel Camera app is ready for both 10-bit and 12-bit data, using the 12-bit solely on the main cam of Pro, as they can't use it on the others. That means they had to keep their compatibility for 10 bits up and wouldn't have to worry about P10 in result. If the plan of DCG use came out late in development or had an insufficient priority, it's plausible that the cheaper model missed on their dedication. Now that they could do it, they probably already think it's too late, so see you on Pixel 11.

PS: Android OEMs lack pride in the app ecosystem, they'd rather use a raw-based video app to mislead users in advertising (shot on x) than truly care about what an advanced third party app could access. It's Instagram, WeChat and stock camera, make no mistake.