Samsung Galaxy S20 Ultra's selfie camera ties for second in DxOMark's tests

05 May 2020
That's a better showing than the rear camera, which placed 6th.

Sort by:

  • D
  • AnonD-754814
  • 6p}
  • 18 May 2020

Nick Tagataka, 18 May 2020Ah, I see. Thanks for the quotation, it was really nice of ... moreWell, It's true that I haven't won a phone with telephoto.
I think I have told you before that My last Flagship was S7.
Sorry about that. But as I knew Phone periscope can't do zoom like that physically. I'm assuming these are happening digitally.
I also think the camera mode was on pro mode in the video.

About the SoC performance,
I think we have discussed the performance thing previously. Geekbench 4.4 shows exynos better and the latest geekbench shows Kirin better.
For the record I think our latest topic was on SoC efficiency. If you have forgot then read previous replies. You claimed Kirin to be more efficient.

    AnonD-754814, 17 May 2020Speaking about the M4 cores and the CPU efficiency. As Exyn... moreSomehow I missed out on this comment, weird.

    "I can't say that Kirin 980 SoC is more efficient than Exynos 9820"
    Yeah, because that was never really my point. What I was saying is Kirin 980 has better CPU (in both power efficiency and performance as indicated by SPEC2006 and bunch of system performance benchmarks carried out by Anandtech), while GPU is better on Exynos by a significant margin. The whole reason why I started this conversation was because you said something like Kirin was worse than Exynos (or Exynos had been always better than Kirin, I forgot what you exactly said), which implied that Exynos was superiour in BOTH CPU and GPU.

      AnonD-754814, 18 May 2020Here is your proof(I'm including the whole para). You can a... moreAh, I see. Thanks for the quotation, it was really nice of you.

      "proof about Samsung S20 ultra having a slider to zoom"
      At this point I'm starting to suspect that you've actually never used a smartphone camera before and it's deadening my motivation to talk with you. Here's what you're looking for.

      https://www.youtube.com/watch?v=JrWACdh3lrA
      I won't spend any more time on this, it took way too long for me to explain the obvious to you.

        • D
        • AnonD-754814
        • 6p}
        • 18 May 2020

        Nick Tagataka, 18 May 2020"There is no such thing means there is no option like 1.5x,... moreHere is your proof(I'm including the whole para). You can also read anandtech topic by yourself as there are graph and more info on this.

        Anandtech : " Looking at SamsungÂÂ’s scheduler power tables, we see some odd characteristics. According to this, the A75 cores actually arenÂÂ’t more efficient than the M4 cores, except for at the lower frequencies. What is really irking me however is I canÂÂ’t seem to understand why the A55 cores are marked to be as high In performance, reaching 42% of the maximum load scale. This characteristic can also be found on the Snapdragon 855. The source code marks that the performance scale is determined by MIPS, which I think is an extremely weird metric to normalise performance with. The reason could be related to how PELT works on the Exynos 9820, however IÂÂ’ll get back to this topic on the system performance page later in the piece.
        Looking at the same data when normalising the x-axis for perf/W instead of absolute power, we again see in more obvious manner that the M4 cores should very much be quite a lot more efficient than the A75 cores at the same performance, at least for the vast majority of the A75’s upper frequency range. Similarly, the A75’s should be a ton more efficient the A55 cores at the upper performance points of the A55’s. The power figures here consider the active power (load minus idle) of the whole platform, not merely just the CPU cores. This would be a case where race-to-idle at higher power points is actually more efficient than staying on the efficient CPU cores, because of overhead of the rest of the SoC and platform such as memory controllers, DRAM, and PMIC. "


        As I gave the proof, I think you also can give me proof about Samsung S20 ultra having a slider to zoom and there are options ranging 1.1x to 3.9x or whatever.
        The options are these,,,, 0.5x, 1x, 2x, 4x, 30x, 100x.
        So, quiet speaking from your mind and show some proof.

        I also hope you know that telephoto zoom doesn't work like Camera's zoom. They're not continuous like camera's zoom. There are steps of zooming at hardware level.

          • D
          • AnonD-754814
          • 6p}
          • 18 May 2020

          Nick Tagataka, 18 May 2020"There is no such thing means there is no option like 1.5x,... moreI think you really have problem at understanding, even simple things.
          I wasn't talking about GPU performance. I was talking about GPU energy efficiency.
          Exynos GPU is more efficient.
          So, I was saying, as Kirin is more efficient at CPU and Exynos is more efficient at GPU.
          The overall SoC efficiency difference isn't a big deal.

            AnonD-754814, 17 May 2020You seem to frequently ignore my logic. There is no such th... more"There is no such thing means there is no option like 1.5x, 2.5x or 3.5x zoom. Is there?"
            Yes, there are, and I've said this for a hundred times already. Just because you don't see the buttons that directly jumps into those magnifications it doesn't mean they don't exist. In the camera app, use your fingers to slowly pinch into a distant subject and you'll see a smooth digital zoom in action. I won't repeat this twice.

            And why do you keep talking about the GPU performance on Exynos? The point I've been making so far is that Exynos has an edge in GPU while Kirin has better CPU cores, latter of which you seem to disagree with me.

            "Also Anandtech mentioned Mongoose M4 is more efficient than Cortex A75"
            Could you quote the part they specifically said it? And how are you going to explain that Kirin 980 devices beat S10 series in multiple benchmarks that measure system performance, which, as I already said, heavily depends on CPU performance?

              • D
              • AnonD-754814
              • 6p}
              • 17 May 2020

              AnonD-754814, 17 May 2020You seem to frequently ignore my logic. There is no such th... moreSpeaking about the M4 cores and the CPU efficiency. As Exynos has older A75 and the low power A55 running at 150MHz more than the Kirin 980.
              We can say overall the Exynos CPU isn't as efficient as Kirin 980. In other words Kirin 980 has more efficient CPU.

              But while taking GPU efficiency into account as well, I can't say that Kirin 980 SoC is more efficient than Exynos 9820.

                • D
                • AnonD-754814
                • 6p}
                • 17 May 2020

                Nick Tagataka, 17 May 2020Yes, cropping from FHD+ to FHD means it's less taxing for G... moreYou seem to frequently ignore my logic. There is no such thing means there is no option like 1.5x, 2.5x or 3.5x zoom. Is there ?
                So why're you even talking about it ?

                Well if the overall endurance rating doesn't prove anything then you can read anandtech's GPU power efficiency tables.
                Exynos 9820 GPU has much better performance/watt result than Kirin 980.
                Actually the Kirin 980's performance/W is on the par with old Exynos 9810.

                Also Anandtech mentioned Mongoose M4 is more efficient than Cortex A75.
                So, the power efficiency difference between cortex A76 and M4 is very minor if not none.

                Exynos has poorer battery life than Qualcomm because Qualcomm implements the ARM core better.
                So, qualcomm does it better doesn't mean Kirin can do it better to.

                  AnonD-754814, 15 May 2020Well. My bad. The video I was talking about turns out to ... moreYes, cropping from FHD+ to FHD means it's less taxing for GPU to render graphics, but obviously it's not as drastic as the difference between FHD and HD that you initially suggested (only around 18% less pixels to render).

                  "Do you even realize there is no such thing"
                  What?

                  "Last years Note 10+ has more battery endurance rating"
                  I literally explained in my previous comment how you can't use that "battery endurance rating" to measure the SoC efficiency as it's affected by other factors, with the most significant one being the display power consumption. Seriously, don't reply to me if you can't be bothered to read and understand my comments. I'm getting tired of this conversation.

                    • D
                    • AnonD-754814
                    • 6p}
                    • 15 May 2020

                    Nick Tagataka, 10 May 2020That crop indicates the reduction in vertical resolution, w... moreDo you realize that Last years Note 10+ has more battery endurance rating than any Huawei flagship according to my knowledge ?

                    You know Huawei had big time chance on making a much more efficient SoC than both Qualcomm and Exynos ? Because they had an integrated modem.
                    But for some reasons they've found it more beneficial not to research the new Kirin 990 with new ARM core and GPU.
                    Which is definitely a bad move.

                      • D
                      • AnonD-754814
                      • 6p}
                      • 15 May 2020

                      Nick Tagataka, 10 May 2020If that's the case then it's fine, but I do suggest you to ... moreWell, there is only 1 option more then.
                      2x .

                      You kept saying 1.1x, 3.9x, 2.5x, 3.5x.
                      Do you even realize there is no such thing ?

                        • D
                        • AnonD-754814
                        • 6p}
                        • 15 May 2020

                        Nick Tagataka, 10 May 2020If that's the case then it's fine, but I do suggest you to ... moreWell. My bad.
                        The video I was talking about turns out to be Kirin 970 vs older Exynos 8895 . I kinda mixed it inn my memory as it was a newer review than those.
                        The exynos CPU did better even it was older.

                        About the cropping thing. The Kirin has to render less as it crops.
                        Right or not ?

                          AnonD-754814, 09 May 2020Well, I don't have anyone here using Huawei flagships. So, ... moreIf that's the case then it's fine, but I do suggest you to go to the store once the lockdown is over and you'll know what I'm saying.

                          "1x zoom"
                          Didn't I tell you multiple times that there's no point of bringing it up here because it's not actually a zoom at all? Why do you think I kept saying "1.1x - 3.9x" instead?

                          If I list a few reasons why I don't use iPhones;
                          -I'm used to Android's UI
                          -There are many apps I always use that are available free on Android but not on iOS
                          -I can expand the storage on my S10 and P30 in case I need to
                          -I'd like to keep my personal and work phones separate, so it's more logical for me to get two not-so-expensive Android flagships than buying a single super-expensive iPhone.

                          I do understand the appeal of using iPhone with Mac, like instant file transfer or automatic backup over wifi, but iPhone isn't just for me all things considered.

                            AnonD-754814, 09 May 2020This isn't the comparison that I was talking about. The tes... moreThat crop indicates the reduction in vertical resolution, which means it was most likely running at FHD whereas S10+ was rendering graphics in FHD+ (I stand corrected for the part where I said P30P used FHD+ - it should've been FHD) since there's nothing that indicates that the resolution was lowered in both x and y directions.

                            The battery life is heavily affected by a display panel being used and OS optimisation so you can't just see the battery endurance score from GSMArena and say which SoC is more efficient than the other, unless they are in exactly the same phone (e.g. Exynos & SD variants of S10). Historically Samsung's top-of-the-line OLED panels have always consumed less power than ones from BOE, hence allowing the battery to last fairly long despite having double the number of pixels. Anandtech just does it more accurately in this regard since they directly measure SoCs' power consumptions while they run the benchmarks.

                              • D
                              • AnonD-754814
                              • 6p}
                              • 09 May 2020

                              Nick Tagataka, 09 May 2020Well, here's the thing. I have tested out P30 Pro thoroughl... moreWell, I don't have anyone here using Huawei flagships. So, can't have it practically.
                              As an armature, peoples taste varies. So, can't take your words.
                              Why don't you do a side by side comparison of 1x zoom. According to you it's bad too.

                              Like I said I really don't care which pc you use. But as you speak so much about value thing about phones I thought you have the same taste for PC.
                              With that price you could have bought a great PC with great monitor.
                              As you talked about MaC OS. Why don't you use iPhone with it ?
                              A Huawei device with iMac is definitely a stupid thing to choose.

                                • D
                                • AnonD-754814
                                • 6p}
                                • 09 May 2020

                                Nick Tagataka, 09 May 2020"3rd party info doesn't have any value over real one" I do... moreThis isn't the comparison that I was talking about. The test was on Sped Test G 2.0 . I will provide the link when I find it again.
                                I must say, you jumped to conclusion too early. Did you watch the video carefully ?
                                Because even if it wasn't that video. It proved one of my point. Kirin 980 runs the GPU test at HD resolution, not FHD.
                                Did you see the crop on the P30 pro's screen ? That's called running at HD, not FHD.

                                And don't lecture much about the Kirin 980's power efficiency. Even having 100MAH bigger battery and a Lower resolution(practically almost half of S10+) the battery endurance of P30 pro is 100 hours compared to 91 hours of S10+. Practical proof is Mate 20 pro. Having the same SoC and battery it's 85 endurance rating.

                                Also sorry about the Geekbench score. It's actually not my fault. The latest Geekbench shows The kirin 980 has better multi-core score(I didn't know about that). But that wasn't the case on the previous version of geekbench. Exynos 9820 had better multicore score on that version. When both phone were released that version of Geekbench was the latest then. And in that version Exynos 9820 has better single core and better multi core score than Kirin.
                                All kirin SoC has got massive multicore score jump on the latest Geekbench. I don't know why.

                                  AnonD-754814, 09 May 2020Did you read the GSMARENA's review comparison that I sent ?... moreWell, here's the thing. I have tested out P30 Pro thoroughly when it came out, and after getting myself the S10 and the regular P30, I compared the latter's camera to my friend's P30 Pro to see how much better the pro's camera is over the regular version. I tested its camera out in both the demo and the retail versions, and let me say this: The image fusion between the main and the telephoto lens takes place from 3x. So like I said - if you doubt me, before complaining about it here, TRY THE ACTUAL DEVICE YOURSELF. 2x photos look sharp across the frame whereas you see a noticeable jump in quality in the middle of the frame at 3x and beyond. And yes I think GSMArena's article was incorrect about that particular statement. I know what I saw and I don't want to waste my time explaining this to you either.

                                  "But I have to say that you're pretty stupid to use iMac because you're not in the Apple ecosystem"
                                  ....I know you meant well but I gotta say that's still pretty insensitive to call me stupid just because I use iMac. I've been a mac user for a long time and simply prefer MacOS over Windows. Because of that my laptop is MBP as well. Uh well, I digressed.

                                    • D
                                    • AnonD-754814
                                    • 6p}
                                    • 09 May 2020

                                    Nick Tagataka, 09 May 2020It will start behave similarly ONCE YOU HIT 3X MAGNIFICATIO... moreDid you read the GSMARENA's review comparison that I sent ?
                                    They clearly did say that P30 pro does the same as S20 ultra. Middle part super clear and corner noticeably soft. Or will you just ignore that to continue your arrogance ?
                                    I think I'm just wasting my time with you. Just the earlier comment you said P30 pro does it differently and when I bring proof that both P30 pro and Ultra does it the same way and the quality was same at 2x zoom. I'm not saying it, GSMARENA said it.


                                    I don't really care which device you use. But I have to say that you're pretty stupid to use iMac because you're not in the Apple ecosystem. It's just a waste of money if you use iMac only.
                                    Also I mentioned about PC because you gave me a standalone photo to look at and I told you to compare side by side with P30 pro. The quality is same at low level. ALso you can't use DXOMARK's sample as proof because we're arguing about them.

                                      AnonD-754814, 08 May 2020I didn't put any pressure on you. I was just reminding you ... moreBack to the original point that I was trying to make,
                                      The processes that take place on each phones are as follows:
                                      S20 Ultra:
                                      1-1.9x - Regular digital zoom
                                      2-3.9x - Image fusion (with edge softness and some aligning errors)
                                      4x and beyond - Excellent quality from telephoto lens

                                      P30 Pro:
                                      1-2.9x - Super res zoom
                                      3-4.9x - Image fusion (with edge softness but fewer aligning errors AFAICT)
                                      5x and beyond - Very good quality from the telephoto lens

                                      To summarise my point: I'm guessing DxOMark gave S20 Ultra a lower score for zoom because they measure the image quality and its consistency throughout the zoom range, and deduct scores significantly if there is a noticeable artifacts or there is a loss of detail at lower zoom magnification. This is also obvious from the fact that V30 Pro, a phone with a "mere" 3x telephoto lens, is pretty far ahead of P30 Pro in zoom score (95 vs 100).

                                        AnonD-754814, 08 May 2020I didn't put any pressure on you. I was just reminding you ... more"3rd party info doesn't have any value over real one"
                                        I don't understand what you mean. Geekbench is one benchmark that shows Exynos 9820 is superiour in single-core CPU performance but inferiour in multi-core. Antutu says Kirin 980 performs better in CPU section of the test than Exynos 9820, and the suite of 10+ benchmarks results on Anandtech report better system performance (which is mostly CPU dependent) and power efficiency from Kirin as well. Just give up on your fantasy and quit whatever the logic bending you're doing in your head already, it won't help you.

                                        As for the Speed Test G: Smaller number = shorter time taken to process the same tasks = faster, which means it shows - as I made obvious several times in the past - CPU is better on the Kirin. Gary himself literally says that in the video lol. Also, P30 Pro wasn't running at HD+ (Honestly where did you get this information from?) but rather at FHD+ resolution. But as the second Speed Test G link shows, Exynos does have an upper hand in overall GPU performance.
                                        https://www.youtube.com/watch?v=lWI7i85oWhM&t=188s
                                        https://www.youtube.com/watch?v=6MSZxiviLTw

                                        "I think I referred you the same photo, just versus P30 pro"
                                        In the GSMArena's article I believe you referred, 4x test was carried out on both devices but due to difference in native only S20U switched to the telephoto lens which obviously allowed it to yield better results. However, at 2x they didn't add P30P into the comparison (which makes no sense), and there was no 3x zoom test either where both devices would be forced to use the image fusion.

                                        "This is a smartphone periscope, not a 2000$ telephoto lens"
                                        ..I'm talking about the quality of the digital zoom that takes place before the camera fully switches to the telephoto lens, not the actual image quality from the telephoto lens.

                                        "Also you just ignored the point that you mentioned, 1.5x, 2.5x, 3.5x, 3.9x"
                                        ......Huh??

                                          1 of 4