Sony Xperia 10 iii for Video

I thought that it would be a good idea to make some frame grabs from a run of the GFXbench. This turned out badly, but I did get some sample frames to upload. In general, the image titles indicate fairly clearly from which benchmark series the frame comes. The indicator "C1" shows highest detail level JPEG compression. "C2" and "C3" files required correspondingly higher compression to create upload-able files (under 2,000 KB).

I was surprised that I ran into problems creating these files. Operating the "Screen Capture" while the benchmark was running resulted in a crash and corrupted the card. I am still in the process of recovering data from that card, which is very annoying. But I did manage to create these sample files.

"01-134107-Tesselation-C1.jpg"

"02-134246-ALU2-C2.jpg"

"03-134410--AztecVul-HT-C2.jpg"

"04-134643-CarChase-C1.jpg"

"05-134949-Manhattan-3_1-C1.jpg"

"06-135029-T-Rex-C3.jpg"
 

Attachments

  • 01-134107-Tesselation-C1.jpg
    01-134107-Tesselation-C1.jpg
    319.9 KB · Views: 64
  • 02-134246-ALU2-C2.jpg
    02-134246-ALU2-C2.jpg
    556.2 KB · Views: 56
  • 03-134410--AztecVul-HT-C2.jpg
    03-134410--AztecVul-HT-C2.jpg
    581.2 KB · Views: 81
  • 04-134643-CarChase-C1.jpg
    04-134643-CarChase-C1.jpg
    310.5 KB · Views: 58
  • 05-134949-Manhattan-3_1-C1.jpg
    05-134949-Manhattan-3_1-C1.jpg
    462 KB · Views: 57
  • 06-135029-T-Rex-C3.jpg
    06-135029-T-Rex-C3.jpg
    722.2 KB · Views: 61
Toronto Int. Film Festival:

I wanted to post a video about the "Broker" movie while the Toronto International Film Festival was still on-going. That was a rushed effort and what I wrote about it had some errors which I hope to correct. Meanwhile, I want to take some time and think about the experience.

Making videos with my Sony Xperia 10 iii was an unusual decision. I have been to the festival many times over the years, but in general, I have not made pictures or videos. I attended, saw my movies and went home. After two years "locked down" (not an entirely accurate description), I felt that this year's festival was worth recording.

Using the Xperia 10 iii was a natural choice. I think that if I had pulled out a real camera, I might have been asked to put it away. I do not think anyone would have minded my having the camera, as long as I did not try to record the movie, but it might have "caused some concern".

My target settings were fairly obvious to me. The 10 iii is no flag-ship camera. It only has fairly good controls and capabilities for FHD video, but for 4K video it only has its "main" (wide) camera. My seat was about as far from the stage as you can get in that theatre. I was in the "nose-bleed" high, upper balcony. So that means FHD at 4x zoom, with stabilization on, auto-focus and everything else auto (exposure and colour balance). The camera was held high (above my natural line-of-sight) so pointing the camera in the right direction was hard, but this particular clip turned out acceptably. Some of my other files that night were ok, and some were not. I was lucky to get this particular clip since searches have not shown another clip replicating its content.

After thinking about it more, I think I have changed my mind about one key thing: Is it worth it to buy a better phone/camera? Up till now, my answer has been "no". Now, I am leaning towards saying "yes". In todays prices, the Xperia 10 iv (which replaced my 10 iii) costs about $450 - $500 US, but the new Xperia 5 iv, is around $900 - $1,000 US. In this situation the Xperia 5 iv would have made possible a 4K video, and one that was visibly better than my video. I point this out because, in theory, I could have recorded a 4K video using the Xperia 10 iii, but it would have looked terrible. Because I had enough experience testing this phone (as you can see by me previous posts), I knew better than try the 4K video in this situation.

Overall, the auto-focus was a bit unstable, the stabilization was "float-y", and there was some focus breathing, but it was ok.

[See also: Cannes and The Tower of Babel: Movie "Broker"]

[caps @ 4:25]
 

Attachments

  • 04-Broker-18-00h09m04s566.jpg
    04-Broker-18-00h09m04s566.jpg
    200 KB · Views: 34
Last edited:
More Technical Notes From the "Broker" Movie Experience:

My previous comments after the event were inconsistent. My first statement was that better phone/cameras could give better results but "not by much." After I thought about it more, I changed my mind. Here is a problem -- no matter how much better the phone/camera, sitting high in the 2nd balcony, at a distance of about 70' - 80' (I worked this out by testing the camera later), hand-held, with that lighting, with people bumping into me and blocking my view, well, I just cannot do that much better. If you think that you can, then show us.

But what can one expect at, say, the "next price level"? That gets interesting. As I mentioned, that "next level" of around $900 - $1,000 US holds Sony's newest "Xperia 5 iv" ("5iv") which is a good representative of that category. And yes, it would have been best used recording at 4K, probably 30 fps. But how different is it? According to GSMarena.com, the "5iv" telephoto camera has a 12 mp 1/3.5" sensor and a fixed focal length f/2.4, 60mm (35mm equivalent), 1/3.5" lens with Dual Pixel Phase detect auto-focus and optical image stabilization, HDR and added 24 and 25 fps frame rates (as well as slow motion, which I would not have used in this situation). My "10iii" has phase detect auto-focus, but it is not clear whether I have "dual pixel" auto-focus, which is a specific technology (I probably have it, but I am not sure). But right now, the biggest missing data is the video bitrate of the cameras. I can report the information I have from my camera file:

"MOV_0009.mp4"
1,102,680,848 bytes
Created Sep 13, 2022, 17:39:42
Length 08:19
Frame width 1920
Frame height 1080
Data rate 17,515 kbps
Total bitrate 17,671 kbps
Frame rate 30 frames/second
Audio
Bit rate 156 kbps
Channels 2 (stereo)
Audio sample rate 48 kHz

This particular file was about 17.5 mbps. So far, my other FHD files have been roughly in that range too. But I still have not found any other test reports that give video bit rates -- anywhere. I could criticize lens reports for camera/phones as well. The top line Sony products apparently have "all glass" lenses. It has never been clear to me whether my "10iii" lenses are "all glass" too.

But get back to the sensor data and the bit rate. At "4x zoom" for my video, I was using the "tele" camera at 2x zoom. That means that my sensor was effectively a "2 megapixel" sensor (8 / (2 x 2)).

If I used the "5iv", the comparison is hard. At "4x zoom", I would effectively have a "3 megapixel" sensor (12 / (2 x 2)). But wait, would I be using it at "4x"? According to the GSMarena data, to get the same composition, my zoom setting on the "5iv" would be "x / 60 * 54". So instead of "4x" zoom, I would have probably used "3.6X zoom", which in turn probably works out to about 3.7 MP. This is much better than my 2.0 MP, but far from the upper line Xperia 1iv (currently about $880 US per GSMarena, but I have doubts about that) which has a true zoom, and thus uses a theoretical 12MP sensor at all these focal lengths.

So having thrown around all these numbers, (and again, still missing the "5iv" bit rate), I can conclude that the "5iv" should give a noticeably better image, but in the end, I would still just have to test it and see the result myself.

Would I recommend buying the "5iv" or a "1iv" today? The most this all moves me to is "maybe". I still look at it all and note that it took over a year before I found a situation where it would have paid off with a better publicly posted "critical" video clip. All for (currently) about 1.7x the price.

[2022-09-22 10:00

While I was preparing this "comment" I eventually realized that I had failed to compensate for the difference between sensor formats (generally 4:3) and movie image formats (16:9). So, all the "effective" megapixel numbers are higher than they should be, but proportionally about right. Since the comparisons are proportionally correct, I decided to finish it with the above numbers. However, if you really need to be more accurate, then you will have to "finish" the calculations yourself.
 
Last edited:
"Battery Care":

This is not directly a photographic issue, but indirectly it affects usability. Not having as full a charge as I feel I should does reduce my confidence in using the cameras, more-so for video, but for stills as well.

This might be part of Android or a Sony add-on. In "Settings" it is in the 5th category from the top labelled "Battery". According to its screen text "Reduces the amount of time the battery is close to fully charged to increase battery lifespan." It has an on/off toggle switch and 3 option settings "Auto", "Custom" and "Always". So far, I have been running with it on "Auto".

Results:

It has been sort of "mixed". I do not use my Xperia 10 iii's alarm. I have a separate alarm. Actually, I prefer to wake to an AM radio, but lately that is not working. My 2nd choice would be to wake to an FM radio. That would be an interesting possibility if my Xperia 10 iii had FM radio, but my unit does not. My alarm clock is set for around 7:00. I think that it usually is charged to 100% when I wake up, but fairly often it is less. For example:

Sunday Oct 9 morning at about 7:30 100%

However, its current charging regime does not "fully" charge the battery. I have found that it drops down to 99% almost immediately if I am doing something like reading some news. On this day, at about 8:00 down to 95%, I plugged in the charger. On Mon Oct 10, at wakeup, charge was 94%. By about 8:15 it was up to 95%. I did not check again until around 10:00 and it was 100%.

So despite having at least a couple of months now (since at least the August firmware update) to learn my usage pattern, including my fairly regular wake up of around 07:00, sometimes the charge is not "topped up" and ready.

This "smart" system is getting in my way. I think I might try the "manual" ("Custom") setup eventually and see how that works out.

2023-02-13 00:56
Update re "Battery Care", Android 12, Build No 62.1.A.0.675 Release Keys

I eventually used the "Custom" settings which includes adjustments for starting and ending the "Battery Care" system and a target percentage level. When "Battery Care" is active, the maximum charge is reduced to a level a bit over its "100%" charge level, which is not actually full charge.

My current Battery Care level is 90%. Mostly, during "Battery Care" it will only charge to that 90% level I set. Near the end of "Battery Care" it charges to it's special 100% limit (below the real "full charge" level) and does not charge further.

Due to my sleep habits, I still sometimes wake up before it reaches its "100%" level. This does not bother me. Partly, since I have set it "manually" and understand the results, I find it acceptable. The way it works conforms to everything I think I know about lithium batteries. I do not really know how much good it is doing for the battery, I find it "acceptable".
 
Last edited:
I decided to try Cyberlink "Power Director" version 10.5.0, build number 120064 for Android. This is a quick report. I tried building a video from a couple of unimportant clips and found that once I added them to the "project" there does not seem to be a way of removing them. Since screen space is small on a cellphone, I do not want a bunch of irrelevant pieces cluttering up my work space. After spending, what I felt was a long time trying to figure out a way of removing the pieces, I have given up. I will be uninstalling it.

Update: I have uninstalled it. This ends this report.
 
A Firmware update became available and I have applied it as of 2022-12-15 23:11. Any photos or videos made after this will have used this firmware. It is another Android 12 version, so I am not anticipating changes.

Current Firmware: (Dec 29, 2022)

Buil number: "62.1.A.0.675 release-keys"
 
Last edited:
Sony Xperia 10iii and Android 13

This is my latest video recorded on my Sony Xperia 10 iii (see below) and combined with my other uploads, from the earliest images I have posted, it almost represents the full development story of this camera/phone (about 4-6 more months and I should have the last firmware update, but it should mainly be "bug fixes"). The obvious questions are "how much did it change over time" and "did the camera phone progress or regress? And along with those is the general questions "do I like the phone?"

I have been thinking about these questions for months now, and they are hard questions, mainly because I did not really pay attention to the cameras during the early period. Also, it is easy when you look at the top line products to say something like "this is the best on the market". When you look at something that you know is not "the absolute best", then you have to find a way to say something that is relative to the market, and accurate.

For example, in HDR stills mode, it often takes about 1 - 2 seconds from the time I tap the screen till the picture is taken - when it works. Sometimes I tap the screen and nothing happens. Is that the fault of the way I tap the screen? I don't feel that I know for certain, and so, up till now, I have not mention it.

What does the latest video I posted show? Looking back to the beginning and thinking about it, it does seem to me that the exposures and maybe the colors are a bit better now. Among my earlier posted images (see comment #5 from Aug 26, 2021 above) I found that often, exposure needed to be reduced by about 1 stop. This was also seen in samples that I posted in the "stills" topic. I have never been satisfied that I knew what caused it, but it did seem somewhat consistent in that it I cannot recall it under-exposing except in obviously dark situations.

What I will say is that it seems to me now, that the stills and videos I have shown seem to be getting better. But the difference seems to be inconsistent and hard to tell. I cannot say for certain whether the difference is my getting better using the cameras, or whether the firmware is getting better, or both. I believe that the difference is, at least in part due to a slight improvement to the firmware. That much is good. Unfortunately, I would also say that I feel that Sony has not developed the camera apps (or maybe the system level firmware) as much as I would have liked. I apologize for being vague, but the most I can say for now is that I would have preferred Sony to have moved the interface a bit closer to the upper line Xperia 5 and Xperia 1. But I have to acknowledge that the Xperia 10 series is aimed at a less "enthusiast" type market.

This latest sample video is a fairly hard test for the camera with some dark shadows (under tree branches and even a fairly lengthy tunnel) and bright highlights (a small waterfall and some rocks in open bright sunlight).

Do I "like" the phone/camera? Well, "sort of . . ." The latest "Xperia 10v" seems to be aimed at a slightly lower price bracket, just as the "Xperia 5iv" which was introduced a few months back. If I could buy one now, I suppose I would buy the Xperia 5iv. But it is still a higher priced product than my 10iii, and I cannot currently afford it. So I am happy that I got what I have. Make what you can of that.

Would I buy an Apple or Samsung at the same price range? No, I can say that I prefer image files I get from the Sony over what I have seen from those phones. The Pixels? I am not sure enough about the prices I could get for the Pixels, but for now, I would say they are competitive.

"[FHD] 2023 05 29 Black Creek Sony Xperia 10 iii Android 13"
posted June 06, 2023
""
 
I made a sample frame grab from the Black Creek clip and checked some pixels in the highlights. The Xperia 10iii's HD setting did better than I thought. The highlights, particularly the waterfall, had "component clipping" where a pixel might read something like "255,250,246" which would indicate that the "red" component had clipped, but I did not find a fully clipped pixel. I might find one if I check more thoroughly. Still, as a test, this clip was set to exposure EV=0, and I think I would have preferred EV=-1. That might have clipped some of the shadows, but I prefer keeping the highlights contained if I can.

Major Correction:

I have been misunderstanding the settings in my Xperia 10 iii. The "HDR" setting only affects STILL photos. There is NO HDR setting for video. On the one hand, I never felt it was doing really good HDR, but now, I am maybe a bit more satisfied with this phone. The dynamic range of the main sensor seems to be pretty good. This was a fairly difficult lighting case, and yet, both the shadows and the highlights ended up reasonably good. The highlights were the real problem area and as I wrote above, an exposure of -1 would have been better, but checking the pixels showed there was still just a bit of detail left at EV= 0.
 
Last edited:
The Black Creek short video had some interesting technical issues. The sunlight that day was fairly contrasty, though not "maxed". The fellow walking the dog was wearing black and the dog was black, and in the tunnel, they are somewhat in shadow. Then again, as I pointed out, the waterfall was almost overexposed to the point of clipping.

I returned on the 11th hoping for the same conditions to see if I could do a bit better. Unfortunately, on the second trip, the sky was a bit overcast, dropping the contrast. I did reduce the exposure 1 stop and made a short recording. Unfortunately, I do not know the fellow with the dog, so I could not repeat that aspect of the clip. Also, I decided to change the composition and come closer to the waterfall, and record it in UHD. The waterfall itself had less water. This frame capture was compressed at C8 which is much more compressed than I usually use, but still retains pretty good detail. The frames I captured were about 9MB each in PNG.

The second recording was technically better (among other things, the horizontal level looks almost perfect), but without the fellow walking his dog, I don't find the whole video interesting enough to post, so here is just a single frame.
 

Attachments

  • 01-2023-06-12-15h00m37s468-C8.jpg
    01-2023-06-12-15h00m37s468-C8.jpg
    1.6 MB · Views: 14
i think it is good but previous sony phones have been quite prone to overheating

It isn't just Sony and it isn't just phones. Any time a videographer is pushing the limits of the current technology, heat is a problem. This is why pretty much every real "Pro" video camera has a fan pumping cooling air around. And yeah, if you push a Samsung phone to its limits, well "gee, it's pretty hot too". In fact, it is about as hot as a Sony product using the same chipset and recording equivalent video. The electrons don't care if you are using Sony, or Samsung or whatever. But lately, some of the phone testers are finally figuring it out. "Oh, what's going on here? The Sony video actually looks better. I wonder why?, and what is all this "bit-rate" stuff?"

As for my now "old" Xperia 10 iii, keeping in mind that this was Sony's lowest cost phone that year, with Qualcomm's "cheapo" chipset, I expect it to run "warmer" doing what the Xperia V iii and Xperia 1 iii (which had the top-line chipsets) do easily. That's fine. The point of this exercise was purely a test to see what would happen. It did pretty well. But if I want to do better, I have plenty of cameras that can do better. . . .
 
Last edited:
Stats from The Videos:

It did not occur to me before to post the stats I was writing about. Here they are:

The source clip for the "dog walking" video is 210,355,390 bytes and runs for 1:35. According to Windows 8.1 it is 1920 x 1080, with a video "data rate" of 17,523 kbps and a total bitrate of 17,680 kbps, at a frame rate of 30 frames/second. The audio bit rate is 156 kbps and recorded in 2 tracks (stereo) with an audio sample rate of 48 kHz.

VLC reports it as H264 MPEG-4AVC (part 10)(avc1)
Resolution is 1920 x 1080 at 30.180032 per second
Decoded format is Planar 4:2:0 YUV full scale.

The final output file (the file I posted) was 267,042,816 bytes and plays back at 29.970029 frames per second with a total run time of 1:40.

The second file (no dog walker) which I did not post was 704,863,821 bytes with a run time of 2:13. The frame was 3840 x 2160. Windows 8.1 reports the Data rate was 41,901 kbps, with a Total bitrate of 42,057 kbps. The frame rate is 30 fps, with an audio bit rate of 155 kbps, in 2 channel (stereo) at 48 kHz.

The UHD bit rate numbers are roughly what I was getting right back in my earliest (Android 11) UHD videos, as are the frame rate numbers. So it is possible that literally nothing has changed for the UHD video recording. I have not kept track of the Full HD bitrate numbers, but I expect that if I check back, they are probably also about the same as when I first bought the phone.
 
Nearing End of Life -- NOW I get a Problem?

I have mentioned that I am expecting one last Firmware upgrade this Fall. In fact, I am not certain that this will happen because I have had at least 1 upgrade (after the Android 13 upgrade) already. But after the latest upgrade, I am having a battery charging problem, which is probably being caused by "Battery Care". I have had the Xperia 10 iii on charge for most of the day, checking it every couple of hours, and it has never been higher than 90%. In fact a couple of times it discharged from a higher % downward, while the phone plugged in and "warm". I hope this is not the way Sony is going to leave this product.
 
Xperia 10 iii Frame Grabs

I always have a problem deciding where to put discussion about frame grabs from videos. Is it a "stills" issue or a "video" issue? Anyway, I have mentioned before that the frame grabs from the Xperia 10 iii, though "legal" could not be displayed directly on my Windows 8.1 computer. I still don't know why, though I expect that it is a "big endian/little endian" issue. However, the latest frames I grabbed on the Xperia 10 iii (w/the latest version of Android 13) could be directly viewed under Windows 8.1. Usually I converted them to JPEGs through Corel software. It is not that huge an issue, but it is nice that I can save the time and file space now.
 
According to my version number, I am now running Build Number 62.2.A.0.481 release keys with an "up time" of 270:25:26. I will have to back track that to figure out what update this refers to. It is now 21:05.

2023-06-26 15:08
I did a calculation and the "up time" points back to the Jun 12, 2023 update, which is after the above video test. I am in no hurry to re-test it again. I am not anticipating any significant changes to any video or photographic characteristics.
 
Last edited:

Most reactions

Back
Top