Air 3S photo quality is... disappointing

[Droning on and on...]

I mean seriously, don't you have anything better to do? Sorry...

That was rude and unnecessary in what has to this point been a civil and friendly conversation.

And no, tacking "sorry" on at the end doesn't constitute a legitimate apology. Sorry...

 

[Droning on and on...]

I think that by now it has been well established or at least there seems to be a general concensus that the 50MP photos have issues.

Nope. Not "the general consensus".

Opinions vary across the spectrum. I haven't seen any view (good, bad) seem to predominate.

 

[MegaStoke]

Nope. Not "the general consensus".

Opinions vary across the spectrum. I haven't seen any view (good, bad) seem to predominate.

I'm not sure how any serious photographer can look at Air 3s 50mp photos and think they're acceptable.

The stitched 3x panos on the other hand, are sharp and packed full of detail, and are the only reason I haven't already returned my 3S.

 

[Evan03]

Ive been researching this to. Im kinda in this for the still photos and do have a air3s in route.

Ive researched enough to see that everyone loves to hate something

 

[wco81]

It might be that for still photos, Air 2S or Mavic 2 Pro at 20 Mp on 1-inch sensor is better than the 12 Mp Air 3S photos.

I mostly shoot panos so it's still possible that shooting in 12 Mp mode and shooting at least 2 x 6 panos may yield better results than panos from the Air2S or the Mavic 2 Pro.

Otherwise, it should be interesting what the Mavic 4 Pro does. If they do a QB 100 Mp mode, does that imply a 25 Mp Micro 4/3 shot mode?

That would be the highest base resolution yet and if you make panos with those, the results may yet top everything they've previously released.

 

[Filmarik]

That was rude and unnecessary in what has to this point been a civil and friendly conversation.

And no, tacking "sorry" on at the end doesn't constitute a legitimate apology. Sorry...

Quoting just last sentence out of my post and ignoring the rest is typical of someone who wants to manipulate or mislead. That certainly speaks volume about you. I leave others to form their own opinion, I have formed mine already. Appreciate your Sorry at the end though.

 

[Filmarik]

Nope. Not "the general consensus".

Opinions vary across the spectrum. I haven't seen any view (good, bad) seem to predominate.

Again, your typical agressive meaningless "I know better" kind of comment.
Show me one post or thread anywhere where a pro photographer or at least someone who knows what he or she is talking about has said that the 50MP mode photos from A3S do not have serious issues such as color bleeding through sharp edges of subjects, color fringing, lack of sharpness and detail one would expect to see in 50MP file.

 

[GroovyGeek]

Kind of proves my point. You don't really see problems with Air 3S photos unless you pixel peep. Anyone can nitpick any drone to death if you really want.

Reminds me of the Nikon Z6 III and the Internet nitpicking flaws that have no bearing on real world use.

You don't need to go to 400% at least in the images I have seen. It is clearly visible at 50%. And if it indeed demosaicing then the tonality of the whole image is screwed up and not just the edges. Could explain why PS AI denoisendoes a very poor job on 3S images. The patterned color noise may be throwing off the algos. On top of that the 12MPx images are quite mushy, well visible under normal viewing. But if it does not bother you that is the only thing that matters.

 

[romik]

Easy everyone, no need to get excited . I use 400% as a tool - to understand what is going on. And voila - with a little bit of creative thinking we have - also thanks to 400% pixel peeping - a solution (IMHO) to 50Mpx raw problem.

It turns out that 12Mpix raw file has the same amount of information - INCLUDING SHARPNESS! - as the 50Mpix one - minus demossaic issues.

So if you'd use Lightroom Classic and upscale 12Mpix raw file (use Denoise, then deselect Denoise and select Super resolution) you would get 50Mpix DNG image with the same sharpness as the original 50Mpix - MINUS the demosaicing issues! Mind boggling isn't it?

In fact I dare to say the upscaled version of 12Mpix file has more resolution/sharpness than original 50Mpix...

So I guess I was a little bit trigger happy returning my 3S but whom to blame when this information wasn't available...

So from now on happy flying and enjoy even main sensor images in HiRes

 

[Droning on and on...]

I'm not sure how any serious photographer can look at Air 3s 50mp photos and think they're acceptable.

And anyone that ever invokes the sort of statement, "I'm not sure how any serious [fill in expert]" can oppose the speaker's opinion loses any meaningful credibility with me.

Plenty of "serious photographers" look at A3 50MP photos and judge them to be excellent.

Opinions are not fact. The things you find so objectionable others don't. The problem is not their subjective judgement, but rather your insistence that yours is more valid.

 

[Droning on and on...]

Again, your typical agressive meaningless "I know better" kind of comment.
Show me one post or thread anywhere where a pro photographer or at least someone who knows what he or she is talking about has said that the 50MP mode photos from A3S do not have serious issues such as color bleeding through sharp edges of subjects, color fringing, lack of sharpness and detail one would expect to see in 50MP file.

Of course I won't. I have absolutely no interest in lifting you out of your ignorance. The reviews are easy to find using Google and YT. If you really have any interest in an informed opinion, you'll do the research yourself.

I've seen more than enough to know you're just another irrelevant critic.

 

[Filmarik]

Of course I won't. I have absolutely no interest in lifting you out of your ignorance. The reviews are easy to find using Google and YT. If you really have any interest in an informed opinion, you'll do the research yourself.

I've seen more than enough to know you're just another irrelevant critic.

There you go. You have just demonstrated to everyone here that you have no evidence to support your claim. So you have resorted to name calling and personal attacks. Quite pathetic frankly. But more to the point of this discussion, if you believe the A3S 50MP photo mode is excellent and flawless and photos taken in this mode satisfy your photography needs I am happy for you. But that does not entitle you to call names everyone who's (myself including) opinion based on rigorous testing, is different. Let's just say that we agree to disagree on this one, shake hands and move on

 

[Filmarik]

Easy everyone, no need to get excited . I use 400% as a tool - to understand what is going on. And voila - with a little bit of creative thinking we have - also thanks to 400% pixel peeping - a solution (IMHO) to 50Mpx raw problem.

It turns out that 12Mpix raw file has the same amount of information - INCLUDING SHARPNESS! - as the 50Mpix one - minus demossaic issues.

So if you'd use Lightroom Classic and upscale 12Mpix raw file (use Denoise, then deselect Denoise and select Super resolution) you would get 50Mpix DNG image with the same sharpness as the original 50Mpix - MINUS the demosaicing issues! Mind boggling isn't it?

In fact I dare to say the upscaled version of 12Mpix file has more resolution/sharpness than original 50Mpix...

So I guess I was a little bit trigger happy returning my 3S but whom to blame when this information wasn't available...

So from now on happy flying and enjoy even main sensor images in HiRes

While Lr Denoising and Super Resolution do a remarkable job it is not the same as looking at genuine detail captured by ( not QB) 50MP sensor. The extra pixels are only computer (AI aided perhaps) generated and while in most instances resemble closely the subject represented by the original pixels captured by the sensor, errors and mistakes are not uncommon. I tried the Super Resolution and sometimes in skies with clouds strange blobs appeared. I agree that they look better than those original 50MB photos. The flaws seen in 50MP mode are hidden and not amplified which is helping the new photos look somehow "better"

 

[romik]

While Lr Denoising and Super Resolution do a remarkable job it is not the same as looking at genuine detail captured by ( not QB) 50MP sensor. The extra pixels are only computer (AI aided perhaps) generated and while in most instances resemble closely the subject represented by the original pixels captured by the sensor, errors and mistakes are not uncommon. I tried the Super Resolution and sometimes in skies with clouds strange blobs appeared. I agree that they look better than those original 50MB photos. The flaws seen in 50MP mode are hidden and not amplified which is helping the new photos look somehow "better"

All the more puzzling what information is hidden in these 12Mpix files. I have my fair share of upsampling primarily using Topaz and never did I see such stellar results. I challenge you to find a single artefact or simplification in samples I posted above. Even tree branches on the horizon which are mushy in 50Mpix mode are nice and clean in upsampled 12Mpix. WTF right?

 

[Filmarik]

All the more puzzling what information is hidden in these 12Mpix files. I have my fair share of upsampling primarily using Topaz and never did I see such stellar results. I challenge you to find a single artefact or simplification in samples I posted above. Even tree branches on the horizon which are mushy in 50Mpix mode are nice and clean in upsampled 12Mpix. WTF right?

I hear your excitement and I am equally fond of those new tools. Lr can certainly do a remarkable job and you have demonstrated it with your samples. I am not disputing your findings. Just keep your eye on artifacts in organic areas, mainly those without a concrete or distinguishable detail which you might encounter from time to time. I am glad you found a way how to get good results from A3S when wanting to print large, and that you shared it with others in this forum

 

[wco81]

These doesn’t seem to be much innovation in camera sensors in general and especially at smaller sensor sizes, since phones are taking away camera sales.

In one inch and Micro 4/3 sensors, resolution still tops out at 20 Mp. Where DJI could have spent more on sensors would have been BSI or even more costly stacked 1-inch or Micro 4/3 sensors.

So drone cameras are still a couple generations behind some niche compact cameras like the Sony RX100 VII, which I believe has a stacked 20.1 Mp stacked sensor. It’s a $1300 camera but it has a zoom lens which has better reach (24-200 f2.8 to 4.5) and better optical quality than any drone.

Mind you, this camera was released in 2019 and its price has held up.

So there are a lot of ways DJI could have improved camera performance without resorting to QB megapixel gimmicks.

 

[Filmarik]

These doesn’t seem to be much innovation in camera sensors in general and especially at smaller sensor sizes, since phones are taking away camera sales.

In one inch and Micro 4/3 sensors, resolution still tops out at 20 Mp. Where DJI could have spent more on sensors would have been BSI or even more costly stacked 1-inch or Micro 4/3 sensors.

So drone cameras are still a couple generations behind some niche compact cameras like the Sony RX100 VII, which I believe has a stacked 20.1 Mp stacked sensor. It’s a $1300 camera but it has a zoom lens which has better reach (24-200 f2.8 to 4.5) and better optical quality than any drone.

Mind you, this camera was released in 2019 and its price has held up.

So there are a lot of ways DJI could have improved camera performance without resorting to QB megapixel gimmicks.

Good points you are making! I am not going to pretend that I am fully versed in QB sensor technology but there must be something there which is pushing the R&D in that direction. Maybe cost? Some of the videos I have seen from A3S were pretty impressive so maybe that is where QB sensors shine? Maybe over time we will see drones dedicated to photography and drones for videos with different technology in both.

 

[BingBong]

Good points you are making! I am not going to pretend that I am fully versed in QB sensor technology but there must be something there which is pushing the R&D in that direction. Maybe cost? Some of the videos I have seen from A3S were pretty impressive so maybe that is where QB sensors shine? Maybe over time we will see drones dedicated to photography and drones for videos with different technology in both.

It's mainly marketing that pushes QB I suspect. As in the mobile market, the superficially impressive MP numbers wow the wider market. The drone market is tiny in comparison, so has to choose from sensors developed for mobiles.

 

[Filmarik]

It's mainly marketing that pushes QB I suspect. As in the mobile market, the superficially impressive MP numbers wow the wider market. The drone market is tiny in comparison, so has to choose from sensors developed for mobiles.

I get that but you are talking about marketing of gadgets with QB sensors already in them. I am wondering what is pushing the R&D of the QB sensor technology in the first place?

 

[Evan03]

Can a cheap *** lense even take quality photos?
Look at the cost of good full frame lenses. These drones must be cheap as **** with no quality at all on the glass