Samsung Galaxy S25 Ultra to pack massive camera upgrades
- ?
- Anonymous
- CbK
- 17 Jun 2024
Himmelskibet, 13 Jun 2024Another samsung device with basically no unique features, a... moreThis is not 2013
- H
- Himmelskibet
- qPK
- 13 Jun 2024
Another samsung device with basically no unique features, and not even something as simple as a headphone jack or sd slot. And they want 1400+ for it. No thanks. I will stick with my Unihertz Tank 3 Pro
- W
- WhoCaresTM
- 0mU
- 03 Jun 2024
.alpha, 03 Jun 2024Nice BS. S24 Ultra isn't even the best video taker in ... moreYes, yes it is. No matter how much copium you inhale, that won't change. Facts are facts.
And it's not even close.
The amount of jittering, lack of stability and cohesion in colours between lenses in Vivo, Xiaomi and Honor phones is disgusting. Oh, and don't even get me started on lense fogging.
- .
- .alpha
- 4}u
- 03 Jun 2024
WhoCaresTM, 02 Jun 2024I've watched countless...and I really mean countless o... moreNice BS. S24 Ultra isn't even the best video taker in the Android camp
- W
- WhoCaresTM
- 0mU
- 02 Jun 2024
yungfishstick, 27 May 2024Samsung has had just good enough cameras for the past coupl... moreI've watched countless...and I really mean countless of videos, comparing the s23 ultra with everything under the sun.
Imho the s23 ultra takes way better pictures than any phone out there, including the x100 pro, the xiaomi 14 ultra, the iphone 15 pro, etc. And it comes a close second to the iphone for video.
Not sure what you chinese fans are even looking at.
- ?
- Anonymous
- 3JU
- 30 May 2024
Can't wait
- ?
- Anonymous
- JH0
- 30 May 2024
RaviSS, 29 May 20241. Effective fill factor is close to 100% if each pixel has... moreFurthermore, regarding my 600 megapixel comment: That's also the reason why smartphone tele cameras (Xiaomi 14 Ultra, Pixel 7 Pro, Galaxy S24 Ultra, Oppo X7 Ultra, Vivo X100 Ultra) have a high resolution.
A Pixel 7 Pro has a (pixel size/micrometer)/f-number ratio of 0.4 with pixel binning, but 0.2 without pixel binning and indeed it gets slightly more detail from objects when it doesn't use pixel binning.
A Panasonic Zs60 delivers slightly more detail from distant objects (at the maximum zoom setting) than a Panasonic Zs200 because the Panasonic Zs60 has a ratio of 0.2, whereas the Zs200 has a ratio of 0.375.
A ratio of 0.2 is quite common.
Some people use the term diffraction-limited for the pixel size / f-number ratio where diffraction just starts(!) to become a tiny bit visible. But the pixel size still needs to be much smaller in order to capture the maximum level of detail from an object regarding diffraction.
The value 0.2 is actually for "monochrome" sensors at 550nm. A value of 0.2 would require 600 megapixel for a 72mm² sensor with an f-number of 1.7. And as the sensor isn't a "monochrome" sensor, but a Tetra²pixel sensor (which also captures blue light), the value should be even significantly lower than 0.2, so even significantly more than 600 megapixel would be necessary in order to capture the maximum level of detail from an object regarding diffraction.
- ?
- Anonymous
- JH0
- 29 May 2024
RaviSS, 29 May 20241. Effective fill factor is close to 100% if each pixel has... more1. I did say that Quad Pixel autofocus sensors could possibly have a slightly worse low light performance, but I doubt that it would be deployed if it was 20% worse... And not every 200 megapixel sensor has Quad Pixel autofocus, which possibly could reduce the 200 megapixel mode sharpness, too.
3. No. As I said, a 24 megapixel crop is comparable to a 24 megapixel Canon Aps-c camera at f/10.5, where you barely notice the effect of diffraction. Planetary astrophotographers know that (pixel size/micrometer) / f-number must be lower than 0.2 in order to get the maximum amount of detail that is theoretically possible for a Bayer sensor at 550nm. For a Quad Bayer or Tetra²pixel sensor or blue light, the value should be even lower in order to get rid of all demosaicing artifacts. Demosaicing artifacts reduce the level of detail. In other words: For a 72mm² sensor, Tetra²pixel array and an f-number of 1.7, even a resolution of 600 megapixel wouldn't be sufficient to get the maximum level of detail....
4. Not true. All Quad Bayer sensors use analog binning. Sony even officially confirmed that. Several years ago I myself used a Sony 12 megapixel Bayer sensor, which had a 3 megapixel analog pixel binning mode, which was very special.
- R
- RaviSS
- XSm
- 29 May 2024
Anonymous, 29 May 2024"Now you might assume that there is no harm in having ... more1. Effective fill factor is close to 100% if each pixel has its own microlens. In the case of Samsung 200MP sensors, there are 50 million microlenses, each covering a 2x2 pixel array. (as evidenced by the "Super QPD" functionality, which wouldn't have been possible if there were one microlens for each of the 200 million pixels). Therefore, there is around 20% light which is not getting captured by photodiodes.
2. I am not sure how much additional benefit the higher number of pixels will provide in terms of colour information. Benefit, if any, would be marginal at best by my estimation. If there is substantial benefit, I would like to learn more about how that materialises.
3. No, the sensor at 200MP is definitely diffraction limited. The diameter of Airy disk at f/1.7 is 2.24 microns, which is larger than Maximum circle of confusion (which is 2.5 times the pixel size, i.e. 1.5 microns in this case). In other words, for a f/1.7 aperture, you need at least 0.9 microns pixels to not be diffraction limited. In other other words, for a f/1.7 aperture and 1/1.3" sensor size, the effective resolution based on diffraction is around 90 MP.
4. I don't see the relevance of the comment you have made about 'analog pixel binning' since that is applicable to CCD sensors. We are talking about a CMOS sensor here, which uses 'digital pixel binning'.
- ?
- Anonymous
- 70d
- 29 May 2024
impinas, 28 May 2024Yes, good, but will Samsung ever decide to compete in the I... moreThe iPhone and iPhone Plus are upper midrange devices.
- ?
- Anonymous
- JHk
- 29 May 2024
RaviSS, 28 May 2024The 50MP and 200MP sensors on smartphones are not true 50MP... more"Now you might assume that there is no harm in having more pixels i.e. in the worst case (dimly limit) situations, the data from four 0.6 μm would be combined and it would be equivalent to one 1.2 μm. Well that assumption is also faulty. Pixels are divided by seperators which are slightly larger than 0.1 μm, and these seperators don't collect any light information, so basically you lose around 20% of the effective area of the pixel.
That's why, in most use cases, for the same sensor size, a 50MP sensor would give you a better quality than a 200MP sensor, whether your output is 12.5MP or 50MP."
Not true.
1. Due to microlenses, the light is directed to the light sensitive area. The effective fill factor is nearly 100% according to Wikipedia. So, in analog pixel binning mode, a 50 megapixel sensor still performs like a 12.5 megapixel sensor. Analog pixel binning doesn't combine data, it combines charge.
But 50 megapixel sensors with Quad Pixel autofocus may have a slightly worse low light performance because one large microlens could be less effective in directing the light to the sensitive areas of the four subpixels.
2. The reason why a lower megapixel number can lead to a better signal to noise ratio is read noise. A 50 megapixel sensor in its 50 megapixel mode leads to more read noise than a 50 megapixel sensor in its 12.5 megapixel analog pixel binning mode and also more read noise than a 12.5 megapixel sensor.
"The 50MP and 200MP sensors on smartphones are not true 50MP / 200 MP sensors. The color resolution in both cases is the same"
The effective color resolution is not completely the same. Due to having more pixels, it should be still easier to get slightly better color values after demosaicing. Especially at the edges of a 4x4 pixel group.
"the additional count of pixels will be very limited given the very small pixel pitch, which has its own set of challenges such as diffraction limits"
The main camera of the S24 Ultra has an f-number of 1.7 and a pixel pitch of 0.6 micrometer. The diffraction is quite low, a 24 megapixel crop would be comparable to a 24 megapixel Canon Aps-c camera at f/10.5, where a reduction in sharpness is just slightly or barely visible at pixel level.
- ?
- Anonymous
- EU0
- 29 May 2024
Wtf, 28 May 2024Wtf you been saying? Regular S24 is very much on par with I... morePlease kindly relax sir.
- S
- Ss
- u7V
- 29 May 2024
200mp old main sensor with bad color and blurry low light image then. Certainly hope for actual improvement such as variable aperture, large pixel size instead of useless high megapixel, and decent colour reproduction but that mean Samsung can't reuse their three year old sensors.
- W
- Wtf
- 7jW
- 28 May 2024
impinas, 28 May 2024Yes, good, but will Samsung ever decide to compete in the I... moreWtf you been saying? Regular S24 is very much on par with Iphone 15 pro... Better processor, better camera... You sound so immature, have you even compare the specs?
- T
- Thara
- U{%
- 28 May 2024
it it has 150x zoom or more,i will buy it
- T
- Tech
- fCZ
- 28 May 2024
Lamith, 28 May 2024Just give me a high end smartphone that doesn't have h... moreAbsolutely
- G
- Guru
- X@T
- 28 May 2024
RaviSS, 28 May 2024The 50MP and 200MP sensors on smartphones are not true 50MP... moreIt's called pixel binning.Cheap phone uses this technology to upscale their photo details. But flagship phone companies like Samsung uses 200mp sensor.
- i
- impinas
- 3Af
- 28 May 2024
Yes, good, but will Samsung ever decide to compete in the Iphone pro segment? Regular S/S+ can barely compete with regular Iphone/plus. In other words, I'm wondering why they don't produce a smaller "Ultra".
- L
- Lamith
- Fvd
- 28 May 2024
Just give me a high end smartphone that doesn't have half of its production budget wasted on camera's that will always be more shitty than a dedicated digital camera. There simply is no 8 Gen 3 phone that doesn't have multiple camera's, telezoom, optical zoom blablabla.
- R
- RaviSS
- XSm
- 28 May 2024
Anonymous, 28 May 202450MP and 200MP sensors aren't made the way they for &q... moreThe 50MP and 200MP sensors on smartphones are not true 50MP / 200 MP sensors. The color resolution in both cases is the same, i.e. 12.5MP (50 MP uses one color filter for 4 pixels while 200MP uses one color filter for 16 pixels). So you only get more data only in respect of the light information.
Let's take for example the main sensors of S24 Ultra and Pixel 8 series. Both are 1/1.3" sensors. The former has 0.6 μm and the latter has 1.2 μm. Basically each pixel in the latter sensor has been sub-divided into 4 more pixels.
But except in perfectly lit situations, the actual amount for additional detail resolved with the additional count of pixels will be very limited given the very small pixel pitch, which has its own set of challenges such as diffraction limits, not to mention the computation load for demosaicing will be much higher given the 16x difference between color resolution and number of pixels.
Now you might assume that there is no harm in having more pixels i.e. in the worst case (dimly limit) situations, the data from four 0.6 μm would be combined and it would be equivalent to one 1.2 μm. Well that assumption is also faulty. Pixels are divided by seperators which are slightly larger than 0.1 μm, and these seperators don't collect any light information, so basically you lose around 20% of the effective area of the pixel.
That's why, in most use cases, for the same sensor size, a 50MP sensor would give you a better quality than a 200MP sensor, whether your output is 12.5MP or 50MP.