Samsung S20 ultra comparison

Started 7 months ago | Discussions
Lee Jay Forum Pro • Posts: 56,227
Re: Samsung S20 ultra comparison

Panino Manino wrote:

You people are talking about single pixels and single photons?
Really?

It's a way to demonstrate the falsity of his argument against smaller pixels. And photon counting is a proven way to increase image quality.

-- hide signature --

Lee Jay

 Lee Jay's gear list:Lee Jay's gear list
Canon PowerShot SX50 HS Canon EOS 7D Mark II Canon EF 70-200mm F2.8L IS II USM Canon EF 8-15mm f/4L Fisheye USM Canon Extender EF 2x III +10 more
RDM5546
RDM5546 Senior Member • Posts: 1,415
Re: Samsung S20 ultra comparison

Jostian wrote:

RDM5546 wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Downsizing destroys the advantage of more pixels. Why not downsize to one pixel?

More pixels is great for cropping yes not for comparing the quality of output though.

That's the entire point of doing the comparison!

you are much mistaken to think more megapixels equals better IQ!

Then why does the inventor of our modern CMOS sensors (Eric Fossum) agree with me?

https://www.gigajot.tech/

"Being developed at Gigajot, QIS is the next generation of image sensors where high-speed single-photon detection is used to unlock new image capture capabilities for consumers and professionals not possible with today’s devices."

"Jot" is the specialized pixel that is sensitive enough to detect a single photon of light."

"A QIS may contain hundreds of millions to billions of jots."

so a 12mp sony FF has poorer quality IQ than a 108mp Xiaomi??

Sensor size also matters because of total light collected at a given f-stop and shutter speed.

Obviously... but you specifically said MPs are all that matters!

Show me where I said that they were ALL that mattered.

apologies, you didnt say ALL but you said the more megapixels equals better IQ which is not true, unless (theoretically) all else is equal. But the review section is full of reviews (even cameras with same size sensors) where cameras with lesser MPs deliver better IQ.

However, if the 0.8micron pixels from the 108MP sensor were scaled up to full-frame,

But they are not, they are tiny 0.8 micron, you could say the same about any smaller sensor...

No, I'm saying leave the pixel size alone and scale the sensor up.

but you cannot do that above 100% crop 1:1, thats the point and thats why DPReview always use the sensor with the least MP as a base to which other sensors with more MPs get resized to in their studio scene!!

then you'd have a 1.35GP sensor that would out-perform that 12MP Sony in all but the most light-starved situations.

any idea how much light you'd need,

The same. The total light captured doesn't change.

exactly, but when that same amount of light needs to be spread across 12 mp vs 108 mp it makes a huge difference.

in anything but ultra bright light the 12MP Sony will be better

Not true. It's a total myth that larger pixels are better in low light, except for a very extreme situation where each pixel is only gather a couple of photons (where the read noise dominates the shot noise).

You are wrong, a 0.8 micron pixel will NOT gather the same amount of light as a 3.8 micron pixel.

, but again that goes against what you stated earlier saying that more MP's equals better IQ

Because, all else being equal, it does.

but camera's and phones are not, you are overly theoretical, there are many reviews where lesser MP cameras delivered better IQ than same size sensors with more MPs...

But ok we are hijacking this thread, lets just agree to disagree

There is a shot noise signal to limit in a pixel equal to the square root of the maximum capacity of electrons that can be stored in a pixel. If a pixel can store 1,000,000 electronics there is a maximum signal to noise ratio of 1000:1 (120 decibels). The storage capacity of a pixel increases with the surface area of the the pixel. So if the pixel with a 1,000,000 electrons capacity is doubled is height and width stores 4,000,000 electrons then this 4 fold increase in size will improve the signal to noise ratio to 2000:1.

Exactly! THANK YOU!

It always possible to not reach the shot noise limit by a worse sensor design but the the shot noise limit is the final barrier in photo dectectors. Small pixels have more shot noise. Smartphones with headline numbers claiming 108mp like the Samsun S20 Ultra actually default using that sensor with a binning method to produce a 12 megapixel image.

I own this Samsung smartphone camera and the images are typically 12 megapixels that typically make 2-3megabyte JPEG files. You can disable the binning which makes the JPEG files 20-30megabytes in file size and the pixel array is 108 megapixels rather than the defaul 12 megapixels. However the 108 milion pixels are noisiers when you pixel peep. Looking a the whole image the noise looks about the same image noise wise as the 12 megapixel default images made from the 3x3 binning.

So why did the chip designer cut the sensor into 120 megapixels. Ignoring the PR value there is value when software digital zooming with that sensor. Digital zooming is a fancy method of cropping images. If you download raw file of the 108mp image and then crop the image in post processing you do get a better cropped image than if you download the 12 megapixel binned image and crop that image. Cropping higher megapixel imagex works better both in reality and theory. The Samsung S20 zooming uses the full 108 mp images when doing digital zooming to make the zoomed images look better.

 RDM5546's gear list:RDM5546's gear list
Canon G5 X II Canon EOS 70D Canon EOS 7D Mark II Canon EOS 5D Mark IV Canon EOS R +34 more
Lee Jay Forum Pro • Posts: 56,227
Re: Samsung S20 ultra comparison

RDM5546 wrote:

Jostian wrote:

RDM5546 wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Downsizing destroys the advantage of more pixels. Why not downsize to one pixel?

More pixels is great for cropping yes not for comparing the quality of output though.

That's the entire point of doing the comparison!

you are much mistaken to think more megapixels equals better IQ!

Then why does the inventor of our modern CMOS sensors (Eric Fossum) agree with me?

https://www.gigajot.tech/

"Being developed at Gigajot, QIS is the next generation of image sensors where high-speed single-photon detection is used to unlock new image capture capabilities for consumers and professionals not possible with today’s devices."

"Jot" is the specialized pixel that is sensitive enough to detect a single photon of light."

"A QIS may contain hundreds of millions to billions of jots."

so a 12mp sony FF has poorer quality IQ than a 108mp Xiaomi??

Sensor size also matters because of total light collected at a given f-stop and shutter speed.

Obviously... but you specifically said MPs are all that matters!

Show me where I said that they were ALL that mattered.

apologies, you didnt say ALL but you said the more megapixels equals better IQ which is not true, unless (theoretically) all else is equal. But the review section is full of reviews (even cameras with same size sensors) where cameras with lesser MPs deliver better IQ.

However, if the 0.8micron pixels from the 108MP sensor were scaled up to full-frame,

But they are not, they are tiny 0.8 micron, you could say the same about any smaller sensor...

No, I'm saying leave the pixel size alone and scale the sensor up.

but you cannot do that above 100% crop 1:1, thats the point and thats why DPReview always use the sensor with the least MP as a base to which other sensors with more MPs get resized to in their studio scene!!

then you'd have a 1.35GP sensor that would out-perform that 12MP Sony in all but the most light-starved situations.

any idea how much light you'd need,

The same. The total light captured doesn't change.

exactly, but when that same amount of light needs to be spread across 12 mp vs 108 mp it makes a huge difference.

in anything but ultra bright light the 12MP Sony will be better

Not true. It's a total myth that larger pixels are better in low light, except for a very extreme situation where each pixel is only gather a couple of photons (where the read noise dominates the shot noise).

You are wrong, a 0.8 micron pixel will NOT gather the same amount of light as a 3.8 micron pixel.

, but again that goes against what you stated earlier saying that more MP's equals better IQ

Because, all else being equal, it does.

but camera's and phones are not, you are overly theoretical, there are many reviews where lesser MP cameras delivered better IQ than same size sensors with more MPs...

But ok we are hijacking this thread, lets just agree to disagree

There is a shot noise signal to limit in a pixel equal to the square root of the maximum capacity of electrons that can be stored in a pixel. If a pixel can store 1,000,000 electronics there is a maximum signal to noise ratio of 1000:1 (120 decibels). The storage capacity of a pixel increases with the surface area of the the pixel. So if the pixel with a 1,000,000 electrons capacity is doubled is height and width stores 4,000,000 electrons then this 4 fold increase in size will improve the signal to noise ratio to 2000:1.

Exactly! THANK YOU!

It always possible to not reach the shot noise limit by a worse sensor design but the the shot noise limit is the final barrier in photo dectectors. Small pixels have more shot noise.

Well, technically it's less, but I know what you meant.

Regardless, shot noise at the pixel level is largely irrelevant for almost everything.

Smartphones with headline numbers claiming 108mp like the Samsun S20 Ultra actually default using that sensor with a binning method to produce a 12 megapixel image.

I own this Samsung smartphone camera and the images are typically 12 megapixels that typically make 2-3megabyte JPEG files. You can disable the binning which makes the JPEG files 20-30megabytes in file size and the pixel array is 108 megapixels rather than the defaul 12 megapixels. However the 108 milion pixels are noisiers when you pixel peep.

Look at the same image scale, and I'll bet not.

Here's my example from long ago to demonstrate the advantage - even in low-light - of using small pixels.

Looking a the whole image the noise looks about the same image noise wise as the 12 megapixel default images made from the 3x3 binning.

That's because you're looking at a down-sampled image unless you have a 108MP monitor.

So why did the chip designer cut the sensor into 120 megapixels.

108.

Cropping, in short.

Ignoring the PR value there is value when software digital zooming with that sensor.

Yes.

Digital zooming is a fancy method of cropping images.

It's just another name.

If you download raw file of the 108mp image and then crop the image in post processing you do get a better cropped image than if you download the 12 megapixel binned image and crop that image.

Bingo.

Cropping higher megapixel imagex works better both in reality and theory. The Samsung S20 zooming uses the full 108 mp images when doing digital zooming to make the zoomed images look better.

I think of it as 3x digital zoom at 12MP.

-- hide signature --

Lee Jay

 Lee Jay's gear list:Lee Jay's gear list
Canon PowerShot SX50 HS Canon EOS 7D Mark II Canon EF 70-200mm F2.8L IS II USM Canon EF 8-15mm f/4L Fisheye USM Canon Extender EF 2x III +10 more
RDM5546
RDM5546 Senior Member • Posts: 1,415
Re: Samsung S20 ultra comparison

RDM5546 wrote:

Jostian wrote:

RDM5546 wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Downsizing destroys the advantage of more pixels. Why not downsize to one pixel?

More pixels is great for cropping yes not for comparing the quality of output though.

That's the entire point of doing the comparison!

you are much mistaken to think more megapixels equals better IQ!

Then why does the inventor of our modern CMOS sensors (Eric Fossum) agree with me?

https://www.gigajot.tech/

"Being developed at Gigajot, QIS is the next generation of image sensors where high-speed single-photon detection is used to unlock new image capture capabilities for consumers and professionals not possible with today’s devices."

"Jot" is the specialized pixel that is sensitive enough to detect a single photon of light."

"A QIS may contain hundreds of millions to billions of jots."

so a 12mp sony FF has poorer quality IQ than a 108mp Xiaomi??

Sensor size also matters because of total light collected at a given f-stop and shutter speed.

Obviously... but you specifically said MPs are all that matters!

Show me where I said that they were ALL that mattered.

apologies, you didnt say ALL but you said the more megapixels equals better IQ which is not true, unless (theoretically) all else is equal. But the review section is full of reviews (even cameras with same size sensors) where cameras with lesser MPs deliver better IQ.

However, if the 0.8micron pixels from the 108MP sensor were scaled up to full-frame,

But they are not, they are tiny 0.8 micron, you could say the same about any smaller sensor...

No, I'm saying leave the pixel size alone and scale the sensor up.

but you cannot do that above 100% crop 1:1, thats the point and thats why DPReview always use the sensor with the least MP as a base to which other sensors with more MPs get resized to in their studio scene!!

then you'd have a 1.35GP sensor that would out-perform that 12MP Sony in all but the most light-starved situations.

any idea how much light you'd need,

The same. The total light captured doesn't change.

exactly, but when that same amount of light needs to be spread across 12 mp vs 108 mp it makes a huge difference.

in anything but ultra bright light the 12MP Sony will be better

Not true. It's a total myth that larger pixels are better in low light, except for a very extreme situation where each pixel is only gather a couple of photons (where the read noise dominates the shot noise).

You are wrong, a 0.8 micron pixel will NOT gather the same amount of light as a 3.8 micron pixel.

, but again that goes against what you stated earlier saying that more MP's equals better IQ

Because, all else being equal, it does.

but camera's and phones are not, you are overly theoretical, there are many reviews where lesser MP cameras delivered better IQ than same size sensors with more MPs...

But ok we are hijacking this thread, lets just agree to disagree

There is a shot noise signal to limit in a pixel equal to the square root of the maximum capacity of electrons that can be stored in a pixel. If a pixel can store 1,000,000 electronics there is a maximum signal to noise ratio of 1000:1 (120 decibels). The storage capacity of a pixel increases with the surface area of the the pixel. So if the pixel with a 1,000,000 electrons capacity is doubled is height and width stores 4,000,000 electrons then this 4 fold increase in size will improve the signal to noise ratio to 2000:1.

Exactly! THANK YOU!

It always possible to not reach the shot noise limit by a worse sensor design but the the shot noise limit is the final barrier in photo dectectors. Small pixels have more shot noise. Smartphones with headline numbers claiming 108mp like the Samsun S20 Ultra actually default using that sensor with a binning method to produce a 12 megapixel image.

I own this Samsung smartphone camera and the images are typically 12 megapixels that typically make 2-3megabyte JPEG files. You can disable the binning which makes the JPEG files 20-30megabytes in file size and the pixel array is 108 megapixels rather than the defaul 12 megapixels. However the 108 milion pixels are noisiers when you pixel peep. Looking a the whole image the noise looks about the same image noise wise as the 12 megapixel default images made from the 3x3 binning.

So why did the chip designer cut the sensor into 120 megapixels. Ignoring the PR value there is value when software digital zooming with that sensor. Digital zooming is a fancy method of cropping images. If you download raw file of the 108mp image and then crop the image in post processing you do get a better cropped image than if you download the 12 megapixel binned image and crop that image. Cropping higher megapixel imagex works better both in reality and theory. The Samsung S20 zooming uses the full 108 mp images when doing digital zooming to make the zoomed images look better.

I can not load the 108mp raw file here but to give you a rough idea of the S20 Ultra's camera image quality the best I can do is to load the 12mpixel standard JPEG image from the S20 ultra smartphone of a pine tree in my front yard.

12 mpixel standard format image of a pine tree automatically made by 108mpixel image sensor in the Samsung S20 Ultra Smartphone using 9X9 binning. The Image is store in a 6.587 MB JPEG file.

 RDM5546's gear list:RDM5546's gear list
Canon G5 X II Canon EOS 70D Canon EOS 7D Mark II Canon EOS 5D Mark IV Canon EOS R +34 more
Jostian
Jostian Veteran Member • Posts: 4,187
Re: Samsung S20 ultra comparison

RDM5546 wrote:

RDM5546 wrote:

Jostian wrote:

RDM5546 wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Lee Jay wrote:

Jostian wrote:

Downsizing destroys the advantage of more pixels. Why not downsize to one pixel?

More pixels is great for cropping yes not for comparing the quality of output though.

That's the entire point of doing the comparison!

you are much mistaken to think more megapixels equals better IQ!

Then why does the inventor of our modern CMOS sensors (Eric Fossum) agree with me?

https://www.gigajot.tech/

"Being developed at Gigajot, QIS is the next generation of image sensors where high-speed single-photon detection is used to unlock new image capture capabilities for consumers and professionals not possible with today’s devices."

"Jot" is the specialized pixel that is sensitive enough to detect a single photon of light."

"A QIS may contain hundreds of millions to billions of jots."

so a 12mp sony FF has poorer quality IQ than a 108mp Xiaomi??

Sensor size also matters because of total light collected at a given f-stop and shutter speed.

Obviously... but you specifically said MPs are all that matters!

Show me where I said that they were ALL that mattered.

apologies, you didnt say ALL but you said the more megapixels equals better IQ which is not true, unless (theoretically) all else is equal. But the review section is full of reviews (even cameras with same size sensors) where cameras with lesser MPs deliver better IQ.

However, if the 0.8micron pixels from the 108MP sensor were scaled up to full-frame,

But they are not, they are tiny 0.8 micron, you could say the same about any smaller sensor...

No, I'm saying leave the pixel size alone and scale the sensor up.

but you cannot do that above 100% crop 1:1, thats the point and thats why DPReview always use the sensor with the least MP as a base to which other sensors with more MPs get resized to in their studio scene!!

then you'd have a 1.35GP sensor that would out-perform that 12MP Sony in all but the most light-starved situations.

any idea how much light you'd need,

The same. The total light captured doesn't change.

exactly, but when that same amount of light needs to be spread across 12 mp vs 108 mp it makes a huge difference.

in anything but ultra bright light the 12MP Sony will be better

Not true. It's a total myth that larger pixels are better in low light, except for a very extreme situation where each pixel is only gather a couple of photons (where the read noise dominates the shot noise).

You are wrong, a 0.8 micron pixel will NOT gather the same amount of light as a 3.8 micron pixel.

, but again that goes against what you stated earlier saying that more MP's equals better IQ

Because, all else being equal, it does.

but camera's and phones are not, you are overly theoretical, there are many reviews where lesser MP cameras delivered better IQ than same size sensors with more MPs...

But ok we are hijacking this thread, lets just agree to disagree

There is a shot noise signal to limit in a pixel equal to the square root of the maximum capacity of electrons that can be stored in a pixel. If a pixel can store 1,000,000 electronics there is a maximum signal to noise ratio of 1000:1 (120 decibels). The storage capacity of a pixel increases with the surface area of the the pixel. So if the pixel with a 1,000,000 electrons capacity is doubled is height and width stores 4,000,000 electrons then this 4 fold increase in size will improve the signal to noise ratio to 2000:1.

Exactly! THANK YOU!

It always possible to not reach the shot noise limit by a worse sensor design but the the shot noise limit is the final barrier in photo dectectors. Small pixels have more shot noise. Smartphones with headline numbers claiming 108mp like the Samsun S20 Ultra actually default using that sensor with a binning method to produce a 12 megapixel image.

I own this Samsung smartphone camera and the images are typically 12 megapixels that typically make 2-3megabyte JPEG files. You can disable the binning which makes the JPEG files 20-30megabytes in file size and the pixel array is 108 megapixels rather than the defaul 12 megapixels. However the 108 milion pixels are noisiers when you pixel peep. Looking a the whole image the noise looks about the same image noise wise as the 12 megapixel default images made from the 3x3 binning.

So why did the chip designer cut the sensor into 120 megapixels. Ignoring the PR value there is value when software digital zooming with that sensor. Digital zooming is a fancy method of cropping images. If you download raw file of the 108mp image and then crop the image in post processing you do get a better cropped image than if you download the 12 megapixel binned image and crop that image. Cropping higher megapixel imagex works better both in reality and theory. The Samsung S20 zooming uses the full 108 mp images when doing digital zooming to make the zoomed images look better.

I can not load the 108mp raw file here but to give you a rough idea of the S20 Ultra's camera image quality the best I can do is to load the 12mpixel standard JPEG image from the S20 ultra smartphone of a pine tree in my front yard.

12 mpixel standard format image of a pine tree automatically made by 108mpixel image sensor in the Samsung S20 Ultra Smartphone using 9X9 binning. The Image is store in a 6.587 MB JPEG file.

thats my point the IQ (see crop) is no better than std 12mp phones, still overprocessed & smeary.

compared to a grass bush taken with my Note 9, the note 9 shot is better, no so over-processed

-- hide signature --
 Jostian's gear list:Jostian's gear list
Sony a9 Sony a6600 Tamron 28-75mm F2.8 III Tamron 20mm F2.8 Di III OSD 808 +2 more
Panino Manino
Panino Manino Regular Member • Posts: 320
Re: Samsung S20 ultra comparison

What garbage processing... can't you use another camera app to shot the same tree to see how different, and better, the processing can be?

RDM5546
RDM5546 Senior Member • Posts: 1,415
Re: Samsung S20 ultra comparison

Panino Manino wrote:

What garbage processing... can't you use another camera app to shot the same tree to see how different, and better, the processing can be?

These are good suggestions.  I do have better cameras and lenses to shoot the tree with.  I will try that if the sun comes out of the clouds today.   I will use  my EOS 5D Mk14 and my EOS R with lenses set for the same field of view.  I will retake the S20 Ultra photos as well.

 RDM5546's gear list:RDM5546's gear list
Canon G5 X II Canon EOS 70D Canon EOS 7D Mark II Canon EOS 5D Mark IV Canon EOS R +34 more
RDM5546
RDM5546 Senior Member • Posts: 1,415
Re: Samsung S20 ultra comparison

RDM5546 wrote:

Panino Manino wrote:

What garbage processing... can't you use another camera app to shot the same tree to see how different, and better, the processing can be?

These are good suggestions. I do have better cameras and lenses to shoot the tree with. I will try that if the sun comes out of the clouds today. I will use my EOS 5D Mk14 and my EOS R with lenses set for the same field of view. I will retake the S20 Ultra photos as well.

First i will show the two less common camera outputs. The following image is a 108 megapixel (12000x9000) (39.7 MB in size) JPEG image using with the jpg processing done within the smartphone camera app. The large file size makes this format troublesome in handling.

12000x9000 JPEG File made in the Samsung S20 Ultra

Next is another shot as using 4000x3000 raw ouyput files of the Samsung S20 Ultra. The raw image was created as a DNG file, then exported and moved to my computer where the DNG file was opened and converted to JPEG using 100% quality Jpeg settings:

The Samsung 20 Ultra raw image 4000x3000 in size made by a 3x3 binning of the 108 mp sensor with the Samsung S20 Ultra hardware circuitry and then stored in DNG format.

The next image is from a Canon EOS R mirrorless camera with a RF 24-70mmf2.8 IS USM zoom lens

Picture  from a Canon EOS R mirrorless camera with a RF 24-70mmf2.8 IS USM zoom lens

The last image is from a EOS 5D MK IV DSLR using a EF24mmf1.4L II ISM USM lens

Using Canon EOS 5D MK IV DSLR with  a EF24mmf1.4L II ISM USM l prime lens

 RDM5546's gear list:RDM5546's gear list
Canon G5 X II Canon EOS 70D Canon EOS 7D Mark II Canon EOS 5D Mark IV Canon EOS R +34 more
Edgar_in_Indy
Edgar_in_Indy Regular Member • Posts: 427
Re: Samsung S20 ultra comparison

On the picture from the DNG file, it almost looks like the focus is on the bushes in the foreground.

Is there a way for you to share the DNG file so we can play with it ourselves in our RAW developer of choice?

 Edgar_in_Indy's gear list:Edgar_in_Indy's gear list
Olympus XZ-1 Panasonic LX10 Pentax K-30 Panasonic Lumix DMC-GX7 Sony a6500
RDM5546
RDM5546 Senior Member • Posts: 1,415
Re: Samsung S20 ultra comparison

I have sent you a DP Private Message with my email so that I may send you a return email with with the 1510033.dng file. It is 23.6 MB and I do not know a better way.

However, I have put the pictures on my google drive which you should be able to link on. It has the DNG and JPEG files I used as well as a tiff file I made from the DNG file that I used to make the high quality JPEG file of the raw data.

https://drive.google.com/drive/folders/1OrGFzWZcv5pmf5OHULXbk0eO1r9m6NQV?usp=sharing

 RDM5546's gear list:RDM5546's gear list
Canon G5 X II Canon EOS 70D Canon EOS 7D Mark II Canon EOS 5D Mark IV Canon EOS R +34 more
SHood Veteran Member • Posts: 5,620
Pixel 4 comparison
1
 SHood's gear list:SHood's gear list
Olympus E-M1 II Olympus OM-D E-M1X Olympus M.Zuiko 300mm F4 IS Pro Olympus 12-100mm F4.0 Olympus 17mm F1.2 Pro
ultimoInfierno Forum Member • Posts: 62
Re: Samsung S20 ultra comparison

It would be really interesting to know, which image was nearest reality. Differences are tremedous, don’t you Think?

Regards

RDM5546
RDM5546 Senior Member • Posts: 1,415
Re: Samsung S20 ultra comparison

ultimoInfierno wrote:

It would be really interesting to know, which image was nearest reality. Differences are tremedous, don’t you Think?

Regards

I think the two Canon camera images were closer to real life.  The sky was not blue to the extent shown in the S-20 image.   The green colors were not as rich as the S20 image

So the DSLR cameras were more high fidelity in reproducing the colors of the scene.

 RDM5546's gear list:RDM5546's gear list
Canon G5 X II Canon EOS 70D Canon EOS 7D Mark II Canon EOS 5D Mark IV Canon EOS R +34 more
OP albero Regular Member • Posts: 237
Re: Pixel 4 comparison

SHood wrote:

Full size files can be downloaded.

https://www.google.ca/amp/s/www.androidauthority.com/samsung-galaxy-s20-ultra-vs-google-pixel-4-xl-1091116/amp/

Interesting comparison for the "river night shot"...if you download and check the photo of the Samsung S20 ultra is much more detailed and bright compared to the Goigle pixel 4 XL!

 albero's gear list:albero's gear list
Nikon Z6
teemodk Senior Member • Posts: 1,287
Re: Samsung S20 ultra comparison

Lee Jay wrote:

Bingo.

Here is a real world comparison. Same sensor, same camera, different use of the sensor. I think the images speak for themselves.

Left to right: 12mp binned and gcam processed - 48mp gcam processed - 48mp raw (gcam).

RDM5546
RDM5546 Senior Member • Posts: 1,415
Re: Samsung S20 ultra comparison

teemodk wrote:

Lee Jay wrote:

Bingo.

Here is a real world comparison. Same sensor, same camera, different use of the sensor. I think the images speak for themselves.

Left to right: 12mp binned and gcam processed - 48mp gcam processed - 48mp raw (gcam).

WOW!! This is an amazing picture from 108M sensor in the S20 Utra. It very convincing than the original 12p binned is not as good as the other two.

I must apologise that I am not familiar with gcam or gcam processing. I assume the image on the left is a cropped image from a Samsung rear 108mp camera configured to create a 12mp standard 3x3 binned image such as those the Samsung standard camera app creates.

I know how to create a 108mp JPEG image using the standard Samsung camera app but the largest image I can make in RAW mode is 12mp. In the Samsung app I do this by turning the raw on/off switch in preference to the ON position. Then I have to use pro mode of the app to get this RAW switch to work. While in the camera app's pro mode I can not select a format large that 12mp.

I am thinking this 48mp saving of a raw image in some attribe of gcam. I need to learn more about gcam and easiest way to sideload it on my Samsung S20 Ultra.

Can you have both gcam and the Samsung Camera app installed on you phone at one time?   I want to be able to use both apps.

How would you recomend to the the install the easier way?

 RDM5546's gear list:RDM5546's gear list
Canon G5 X II Canon EOS 70D Canon EOS 7D Mark II Canon EOS 5D Mark IV Canon EOS R +34 more
Lee Jay Forum Pro • Posts: 56,227
Re: Samsung S20 ultra comparison

RDM5546 wrote:

teemodk wrote:

Lee Jay wrote:

Bingo.

Here is a real world comparison. Same sensor, same camera, different use of the sensor. I think the images speak for themselves.

Left to right: 12mp binned and gcam processed - 48mp gcam processed - 48mp raw (gcam).

WOW!! This is an amazing picture from 108M sensor in the S20 Utra.

That middle one is labeled 48, so they seem to all be from the 48MP 19mm periscope camera, not the 108MP 7mm main camera.

-- hide signature --

Lee Jay

 Lee Jay's gear list:Lee Jay's gear list
Canon PowerShot SX50 HS Canon EOS 7D Mark II Canon EF 70-200mm F2.8L IS II USM Canon EF 8-15mm f/4L Fisheye USM Canon Extender EF 2x III +10 more
teemodk Senior Member • Posts: 1,287
Re: Samsung S20 ultra comparison

RDM5546 wrote:

teemodk wrote:

Lee Jay wrote:

Bingo.

Here is a real world comparison. Same sensor, same camera, different use of the sensor. I think the images speak for themselves.

WOW!! This is an amazing picture from 108M sensor in the S20 Utra. It very convincing than the original 12p binned is not as good as the other two.

No, I got this sensor from my days with NASA. Some leftovers from Hubble.

teemodk Senior Member • Posts: 1,287
Re: Samsung S20 ultra comparison

Lee Jay wrote:

RDM5546 wrote:

teemodk wrote:

Lee Jay wrote:

Bingo.

Here is a real world comparison. Same sensor, same camera, different use of the sensor. I think the images speak for themselves.

Left to right: 12mp binned and gcam processed - 48mp gcam processed - 48mp raw (gcam).

WOW!! This is an amazing picture from 108M sensor in the S20 Utra.

That middle one is labeled 48, so they seem to all be from the 48MP 19mm periscope camera, not the 108MP 7mm main camera.

It's from an average cam with the IMX586 sensor. Native 48MP, binned to 12MP. The only thing this picture were ment for, is illustrate the claim that more pixels give more detail. The native camera app does not do any processing for the raw 48MP. gCam does and improves the output slightly without smearing.

The 20 ultra should do way better than this, but that was not the topic of my post.

RDM5546
RDM5546 Senior Member • Posts: 1,415
Re: Samsung S20 ultra comparison

Lee Jay wrote:

RDM5546 wrote:

teemodk wrote:

Lee Jay wrote:

Bingo.

Here is a real world comparison. Same sensor, same camera, different use of the sensor. I think the images speak for themselves.

Left to right: 12mp binned and gcam processed - 48mp gcam processed - 48mp raw (gcam).

WOW!! This is an amazing picture from 108M sensor in the S20 Utra.

That middle one is labeled 48, so they seem to all be from the 48MP 19mm periscope camera, not the 108MP 7mm main camera.

OK.  The 48mp makes sense if the 48mp image came from the telephoto camera which has the 48mp sensor.    However, I have not be able to adjust the Samsung camera app to output either a raw or a  JPEG image from the 48mp sensor. I can output a 12mp image from the 48mp image (presumably binned 2x2)  but the 12mp is the largest array I can output from that telephoto camera and that image is available only as a JPEG and not a RAW file.  It appears the gcam App is much flexible than the native Samsung app in this regard.    Maybe it can output raw files of the telephoto camera as well as the 108mp camera.  I am going to have learn more about the fantastic advance camera app on the S20 ultra.

 RDM5546's gear list:RDM5546's gear list
Canon G5 X II Canon EOS 70D Canon EOS 7D Mark II Canon EOS 5D Mark IV Canon EOS R +34 more
Keyboard shortcuts:
FForum MMy threads