S30, not S21.
What is ToF?? Why wouldn't the writer explain this?
so they acknowledge that ToF sensor in scamsung s20 and note 20 series are complete garbage... why don't you change it back to heart rate sensor, scamsung
Anonymous, 16 Aug 2020Any camera can be used as depth, they choose to use 2 or 5 MP because they are cheap and old. ... moreThink before what you say.
You called iPhone's bokeh as garbage ?
iPhone was the first to introduce secondary telephoto and the portrait shot.
AnonD-909757, 15 Aug 2020The specs sheets on websites like Kimovil for most phone give a 0.5MP.
LIDAR is a ToF sensor ... moreI know LiDAR is one type of ToF.
What stage current ToF is won't cost much, around 10$. I am not denying that.
I was referring to your concept 12MP ToF.
Just think about it.
12 million Laser Light beams are bursting through a small source, How costly that can be ? I'm not even sure if it is possible to make. Even if someone makes it it will be too much power hungry.
And as you don't want help from main sensor and don't want that type of complex algorithm to make 3D image by taking like 20 photos.
You need 12,000,000*30 Laser beams per second to smoothly capture a 3D scene.
That's not just going to drain the battery in a few minutes, that should theoretically also fry the device in few minutes because of the heat that will be generated.
Don't count me stupid and think why do one need 12 million beams ? One can do it with 100 beams.
Actually that's for your understanding. Those 100 beams LiDAR's one beam resolution is bigger than a whole 12MP sensor.
What I'm trying to say is if you want to use big sensors(lIke 12MP) then you also have to increase the Laser Beam's resolution which is impossible for now.
Are you thinking you can just get better result by just putting a bigger sensor ?
This isn't regular photography. This is why I said you should know about the LiDAR and how it works.
Remember the Light capturing theory(equation) you gave ?
Think about that. If you don't get enough light back then using 12MP sensor will give worse result than a 0.5MP sensor because 10% beam won't reach into 12MP sensor because of too many pixel divider. Over 10% of all beams are going to fall on pixel divider and wont reach to sensor.
I guess I was able to clear the confusion you're having. Too use 12MP sensor you need to bump up the resolution of the Laser beam too.
Your more megapixel ultrawide theory is very good.
But you're forgetting something here. Smartphone camera pixel sensor has limitation.
For this there is a suitable pixel count. Different people have different choice. But that choice has to be 1.
You're saying 12MP is the best. Which means 12MP will give best results overall.
Now you're saying 16-20 is best for ultrawide/telephoto.
Doesn't that mean you are going to get worse image quality than the 12MP ?
It's going against your own theory. Think about it.
AnonD-909757, 16 Aug 2020True, I didn't even knew Samsung had 5Mp macro camera, I usually only look at high end de... moreI think, You should like Xiaomi Mi 10 Ultra's camera.
That camera specification is just incredible.
It also fills most of your concept phone.
Note 20s not even yet hit the market and s21 leaks start lol.
Galaxy S21 will come with LiDAR.
AnonD-909757, 16 Aug 2020I never pretended I had any.
And you know what ? Peoples often throw AI at everything as if a... more"AI" for colour balance is as easy as comparing what's in the frame with a bunch of templates. You can't template focus. It's a physical focal length setting. That example your showed also rely on you already knowing the image you're altering.
You can work around the issue with having two unrelated focus sources, but it will leave you with issues every time it's fluctuating in the transition zone.
Dear 'GSM Arena' : thank you for providing 'Samsung Galaxy S21' information.
AnonD-762416, 16 Aug 2020The issue with transitions is because you're switching between two completely different m... moreI never pretended I had any.
And you know what ? Peoples often throw AI at everything as if artificial neural networks were the only thing they could do nowaday, but THIS is actually a perfectly adapted problem for one, as literally they are great at those kind of things, they can do WAY MORE complex tasks than just figuring out the common data between two different methods at their overlapping range, and when I say way more complex, I mean like literally rebuilding incomplete data, like that :
But again, as most of the tools companies have, AI isn't user properly, it often do jobs that should be hand coded, and many features that it would allow are kept untouched.
Anonymous, 16 Aug 2020Any camera can be used as depth, they choose to use 2 or 5 MP because they are cheap and old. ... moreAs depth, yes, but as a precise and good depth like a ToF is ? Hell nah !
A regular image isn't good enough as you need to compute the depth out of the image(s) while a depth sensor like a ToF LIDAR already have timing data which are really easy and fast to convert into distance data.
That's exactly why a true ToF depth sensor is required, to allow realistic looking bokeh, almost impossible to tell appart from a real optical one.
You mean like they waited a while before making 5MP macro that gave a great result as opposed to the 2MP one many insist on putting despite having a crappy result ?
Or like the bazillion other things that smartphones desperately need but they don't do ?
A single 12MP ToF sensor is actually cheap, there are many 20¤ good 12MP camera, and as I said, ToF actually cost less as they are less complex.
And I would take a more expensive phone with a camera setup that make sense over one that is expensive anyway and have a totally illogical and even stupid camera setup, like the Galaxy Note 20 Ultra that is well above 1100 and yet have a stupid 0.3MP depth sensor that could either be replaced with a better one for a slightly higher price or simply ditched as it is useless at this resolution.
AnonD-909757, 15 Aug 2020YOU have no idea what you talk about. A depth sensor IS a regular camera except it does 2 thi... moreThe issue with transitions is because you're switching between two completely different methods for depth sensing. You can pretend it doesn't matter and that you somehow has software that covers every scenario, but then you'd just be a liar.
AnonD-909757, 15 Aug 2020ToF give you depth map that for properly using it without taking ages matching its data with t... moreAny camera can be used as depth, they choose to use 2 or 5 MP because they are cheap and old. Phones that use the "zoom" cam for fake/garbage bokeh, use the main camera as depth.
If ToF needed 12Mp, they would be making already.
Then you all will be crying because decent phones cost more than $1100 and will be even more expensive.
Shiv..11, 15 Aug 2020Yes broo!! We really need to see these damn 2MP macro and 2MP depth batch should be ditched al... moreWell, in budget that's one thing as it is cost restricted, but when high end or even flagship have 0.5Mp or 0.3Mp depth and/or 2Mp macro, or 8Mp ultrawide or even 2Mp portrait...
AnonD-754814, 15 Aug 2020I think Samsung was one of the first who came up with 5MP macro camera instead of useless 2MP.
True, I didn't even knew Samsung had 5Mp macro camera, I usually only look at high end devices specs, and almost never at midranger of Samsung (except special one like the A80), so I missed that.
The Galaxy A90 5G from September 2019 do have a 5MP macro camera while Xiaomi did it on December 2019 with the K30 5G.
Apparently Samsung have the following phones with 5MP macro camera :
A90 5G, A51, A71, S10 Lite, M31, A31, A71 5G, A51 5G and M31s.
So except the A21s which is obviously too budget/low cost and have a 2MP macro, all other have a 5MP, interesting to know !
Xiaomi on the other hand did that with their recent higher hand devices, like the K30, Note 9, Note 10, Mi 9 etc.
Can't say I'm that bothered. Had one on the P30 Pro and it didn't appear to make anything any better. I don't miss it on my current device anyway.
AnonD-754814, 15 Aug 2020I think most ToFs are 0.3MP(including Samsung's)(which is made by Sony).
LiDAR or ToF ?
... moreThe specs sheets on websites like Kimovil for most phone give a 0.5MP.
LIDAR is a ToF sensor as it measure the time the light need to travel back and forth, a laser measurement (like typical laser autofocus) is ToF too, but doesn't give any map, only point distance, while a LIDAR also use ToF tech but give a map, which is basically an image with distance/delay rather than color (to simplify).
ToF doesn't cost more than a regular camera, it take even less as you don't even need a bayer for each color, you only need an IR filter (to filter out other light frequencies), the speed at witch it work is basically the same as what any regular camera sensor work, as they are analog, they aren't limited by a tick frequency so the timing is easily measured.
Also since you are outputting periodic images rather than continuous, the analog to digital converter isn't overloaded, making it perfectly fine with regular camera hardware.
Where it would be expensive would be in a camera build for ToF imaging (LIDAR) where you need a really fast converted to continuously output an image.
And it won't take more energy than a regular camera sensor as it basically is a regular camera sensor but with 3 time less real pixels (as you don't have RGB to handle), which also make it less expensive.
I know how it work.
Because 12MP is basically the photography industry standard since a while, and only recently DSLR of more than 12MP emerge in more than few occasional models, which I think is because for the same sensor size (APS-C mostly) the tech have evolved to have better light collecting since the time we make those sensors, giving the same per pixel light quality as bigger fewer ones, allowing higher resolution sensors without loosing quality.
As I said, most smartphones anyway output 12MP, the goal is to have the best grain/details at 12MP, multi bayer tech is not that much better than 12MP camera (though it have its advantages), but that said I'd prefer an IMX700 or ISOCELL GN1 over a 12MP for their added tech despite higher resolutions, which isn't a drawback on them as they output amazing quality.
I explained it, a wide 16MP or ultrawide 19MP is logical because since you have a wider angle, for the same subject, you have less pixels on a wide/ultrawide than a regular camera, the total image have more degrees, meaning you need more pixels to compensate if you want that the same subject, at the same distance, have the same resolution than through the main shooter.
Same with telephoto which have for goal to look far, and on top of the optical zoom, the hybrid/numerical zoom is quite often used too, so if you want a good resolution despite a non optical zoom, an higher resolution (but not too much as it would impact light gathering capabilities which are important in optical zoom) is logical.
Also, combining telephoto with portrait is a smart choice, as face shot require a high focal length to be good looking, which in turn require to be further of the subject than through a regular shooter.
This image give a great demonstration of how a longer focal length give a much better result :
It work universally with all faces.
And with a higher resolution sensor AND a high resolution ToF sensor, able to give a really accurate depth map, we can digitally create a better looking face like a portrait camera do without requiring the telephoto sensor to have too long focal length or having a dedicated camera, which is one of the multiple amazing capabilities a ToF sensor with higher resolution would offer, and having a more than 12MP quality help to compute a 12MP final image without loosing quality through transformations.
I am not biased on any brands, maybe I could say I don't like Lenovo, but that's all and it is more about what they did as a brand than their products.
I could also say I like Huawei ? But funnily enough, despite the P40 Pro+ being the phone that match the most my tastes (well, if it had a pop up with 3D facial recognition or no front camera at all, a flat display and jack), except their IMX700 and wide angle, I would rethink the camera setup if I could anyway, so...
That being said though, Sony and Google are great exemple of brand known for having really good picture quality on their phone and who are using 12MP main camera, they are also the one who aren't using lesser resolution for wide/ultrawide and telephoto, which is observable in the Xperia 1 II (even though it also have those 0.5MP ToF) and Pixel 4, for me that's a big sign, but this isn't from that that I got my logic.