Today i took some shots at night and i confirm. The result is a disaster now in Android 10

[doublem101]

I've been reading all the posts about the bad photo quality on Android 10 but honestly i didn't had the chance to test it in detail before.
However today, i had a spare time and it only took a couple minutes to see what everyone are talking about.
I already had taken some indoor photos and i noticed the high noise in darker areas but i went outside and i couldn't believe on what i was looking at. Slightly blured photos (like a shaken photo) and no detail at all, it looked like a watercolour painting. On some photos, darker areas were pinkish, something that i complained before in Pro mode on long exposures but never noticed it in any other mode.
However, i don't know if it's related or not but the few Gcam versions that works, aren't much better neither. Still a little better than stock app but really really worse than before. I'm not sure if they aren't already optimized or somehow related with the stock app bad quality. I don't know how Gcam works but what if the main issue isn't the stock APP itself but something behind it that makes Gcam works so bad also?
Sorry Asus, but i can't accept this quality from €500 device. One of the things that made me buy this phone was his night photo quality and now it takes worse photos than my old device?
I know you are looking at the problem already but i become frightened when i read that the differences you found aren't considerable. They are. If you didn't tested yet at night, outside, please do it. Theres were i found the major differences, not indoor.
I don't a have photo took in Android 9 under the same circunstances in this same places but please look at this photo. This isn't acceptable at all.

[mastershu]

CH_ASUS

https://zentalk.asus.com/en/discussion/comment/32828#Comment_32828

Can you please help differentiate between HDR Auto, HDR off and HDR++ (HDR+ Enhanced).

These are distinctly different (depending on scene).

---

Are there camera changes in P? Yes, there have been plenty, from day 1 (not necessarily every update).

But what we are discussing here is Q from latest P - correct? Yes, some want to throw in more variables - P-189, this and that, and we'll try to get to those as well, but in order to systematically discuss this topic, throwing things around is not helpful.

We cant fix what we havent changed in the code. What we are trying to do here is to help everyone figure out where the discrepancies are.

We've already said that HDR+ Enhanced has been adjusted , and from our own testing we see generally this being visible in shots where there is a strong light source in the scene (like the sun).

The changes in HDR++ (HDR+ Enhanced) is to give the scene more contrast in "most" situations (no, it wont be in all ) due to feedback that HDR++ generally produced "flattish and bright" images.

For Auto and Night Mode, there are no changes made from P to Q (latest P to Q).

If some dont want to believe this, then no matter what we do or say here won't help you.

---

Some users say they notice camera quality improving if they clear their camera settings (by clear camera app data) and/or restore camera settings to default (inside Camera app's own settings).

It has also been mentioned that there is a suspicion that those units which experience performance issues would have camera image quality results impacted.

These speculations I currently have no judgment on, whether it is possible or not. From what we know, it is unlikely - but I wouldnt rule it out as impossible.

These issues might resolve themselves then by a device reset, or not, or some may have their issues resolved with a future update.

However they still do not change that from what we know, our camera team has not altered any algorithms for Night Mode or Auto, or low light, or noise reduction/etc. going specifically from P to Q.

View post

@CH_ASUS
hdr off is with man farthest away in pic hdr on is when man is closest in pic then hdr --

[isaac_brody]

@CH_ASUS I did a 100 percent crop on these images you posted. 189 is on the left, Q on the right.
Compared side by side 189 is of course flatter, but it also retains a lot more detail in the shadows.
And In 189 you can also see detail in the bark of the tree and on the snow. Overall it looks like it's resolving fine detail and sharpness much better than Q.
The Q version has much higher contrast at the expense of dynamic range and details in the shadows. That lost detail and the noise reduction resolves a softer image, especially on the tree bark and the snow.

[off2fly115]

isaac.brody

@CH_ASUS I did a 100 percent crop on these images you posted. 189 is on the left, Q on the right.

Compared side by side 189 is of course flatter, but it also retains a lot more detail in the shadows.

And In 189 you can also see detail in the bark of the tree and on the snow. Overall it looks like it's resolving fine detail and sharpness much better than Q.

The Q version has much higher contrast at the expense of dynamic range and details in the shadows. That lost detail and the noise reduction resolves a softer image, especially on the tree bark and the snow.

View post

I 2nd this completely

[CH_ASUS]

isaac.brody

https://zentalk.asus.com/en/discussion/comment/33467#Comment_33467

Thanks, @CH_ASUS for doing these tests.

I'm noticing that the noise reduction is indeed less aggressive in 189/191 versus Q. In your last photos with the figures, I can see less noise reduction in the background including what looks like some chroma noise in 189, while Q has noise reduction that kills the chroma noise along with detail.

Additionally, the higher contrast in Q gives a perceptual sharpness bump but also looks like it reduces dynamic range and details. I'm wondering if the additional contrast is throwing out detail that is then smoothed over with the higher noise reduction.

One solution would be to add picture profiles to the settings that allow users to pick their poison. Some folks might like the higher contrast more aggressive noise reduction, while others might prefer a flatter profile with less noise reduction that captures a good base image that is then finished in Snapseed/Lightroom.

Out of curiosity, how is testing done to improve image quality?

View post

HDR++ algorithms have changed as I mentioned - HDR++ should not be used in low light as (even in P) those algorithms are not made for low light.

A side effect of the increased contrast could be that in certain scenes the noise will go up on the dark parts - it could be so yes.

As for Night Mode and Auto, still shouldnt be any sweeping changes except "on device" variations.. like if you take 10 photos of the same thing, not all 10 will turn out the same (even with the same phone/firmware).

[CH_ASUS]

off2fly115

https://zentalk.asus.com/en/discussion/comment/33469#Comment_33469

The 3rd party app camera experience was far superior in Android 9. I know 3rd party camera apps don't go under the same tuning the stock camera does but, I just can't help but notice the worsening quality after the latest Android 10 update. Even though the you and the camera team at Asus say you guys haven't changed any coding for the camera, I am sure something ELSE must've changed the camera quality. As for stock quality camera, the pics there are also overly exposed. I can try to post some example pictures later. I already have posted multiple pictures here on the thread somewhere so maybe you can look back at those?? I know you guys are hard working and I respect you all. But if these many people say the camera experience is this much different then there has to be something that is causing it. Not our hardware or software. Hopefully you guys can release an update soon fixing all of the latest bugs 🙂 - Best regards Branden.

View post

Even if the camera team didnt adjust the "ASUS PART" of the image quality (this refers to Asus Camera) - can there be changes inside Android 10 that would affect how 3rd party apps' cameras behave?

I agree, I would not rule it out as not possible. But I think the discussions like these are what is most constructive.

We are not saying "you cannot see any difference!" or "your eyes are wrong!". We are simply saying what we've done and what we haven't done - but there could ofc be many more factors (and generally is) to these observations.

As for 3rd party apps, we'll try to do some tests as well from our end and see what we can do. Is it any specific app you are seeing the most difference? Facebook Messenger, or Snapchat, or Instagram? Or doesn't matter / similar between them?