Sony Xperia Z5c (and Z3c) Video on Smartphones

VidThreeNorth

No longer a newbie, moving up!
Joined
Oct 21, 2016
Messages
1,300
Reaction score
269
Can others edit my Photos
Photos NOT OK to edit
I am preparing to buy a new 5G phone, and actually, I have hit a snag, but I will not be discussing that "here", at least not right now. But while researching my 5G phone, I ran across something else that I found interesting: I found I could buy a "new, boxed" Sony Xperia Z5c (the "c" is the Compact version), for around $180 US, which I think was around 1/2 price.

Context:

Around 2014 to 2015-ish, there were a couple of key developments going on in the "smart phone" market. The first was the earliest support of 4K video, and the second was the coming of 64-bit main processors. If you want to research exactly who did what and when, you could try searching "GSMarena.com" which is the most thorough and accurate database of which I know. But really I don't care about knowing it all in detail. I know that I bought the Sony Xperia Z3c around 2014-15 and it still used a quad-core 32-bit processor (Qualcomm) but it was one of the first Sony phones to support 4K video. There were probably a couple of other phones with 4K video that year, but Sony was a camera company, so I had more confidence buying the Z3c. It worked out well enough. The next year (2015), Sony replaced the Z3c with the Z5c which also had 4K video support, but was based on a Qualcomm Octa-core 64-bit main processor. I am pretty sure that I knew the key specs of the Z5c before I bought the Z3c, but it does not really matter. From what I saw, I decided early on that I wanted the Z3c and NOT the Z5c. For the past few years I have had no regrets about having the Z3c instead of a Z5c. I have not done much video work with the Z3c. I had other uses for it than video work, but it did fulfill its purpose for video work, when I absolutely needed 4K coverage, until I bought my Yi-M1 M4:3 camera. After that, I did not even think of using it as a video camera -- or even a still camera, until now.

Ironically, I did not really buy the Z5c for video work either. But since I had it, I thought I should at least do some testing, and compare it against my Z3c for now. It is important that I do this testing immediately because I have some vague plans for what I want to do with the Z5c, and it might not have a working camera after I am done with it.

WARNING! The "Sony Xperia Z5c" is not what I would call a "good product". Depending on what you do with it, it can overheat, and that overheating can result in a damaged display. I had heard this before and I did in fact permanently damage my display. The damage is not currently so bad that the phone is effectively not usable, but I do not feel that the circumstances causing the damage are outside anticipated "common" usage. If you want a better Sony cellphone with 4K video capability, I think that something like the later= Sony Xperia XZ might be a better choice. I have NEVER USED a Sony Xperia XZ, but looking at the changes in the technology (particularly the 14nm process) I think that this later device should experience less problems than I encountered.

According to GSMarena.com, the Sony Xperia Z5c uses a Qualcomm MSM8994 Snapdragon 810 (20 nm) chipset that includes an Octa-core (4x1.5 GHz Cortex-A53 & 4x2.0 GHz Cortex-A57) CPU and an Adreno 430 GPU.


Basic 4K Video on the Z3c and the Z5c:

According to GSMarena.com, the main camera specs for the Z3c are: 20.7 MP, f/2.0, 25mm (wide), 1/2.3", 1.12µm, AF. By comparison, the main camera specs for the Z5c are similarly: 23 MP, f/2.0, 24mm (wider), 1/2.3", PDAF. Looking at these specs, one would assume that the results would be almost identical, except that the Z5c would have a bit wider view. That is not quite what I found. We'll see in the samples that sensor cropping makes a shamble of expectations.

Starting from my Sony Xperia Z3c, I mainly had two notes about using it for 4K video. The first note was to "keep the sun at your back" because the sun in the picture, or even substantially in front of you, resulted in flare. How much flare was variable, but my first test had the sun a few degrees outside of the frame and that give me milky flare beyond what I wanted. The second note was "use a tripod" -- emphatically. The Z3C had no stabilization in UHD video. For Full HD video it has digital stabilization, which I felt worked fairly well, but for UHD video it is not implemented, and the small size and light weight of the Z3c made it worse than any real camcorders I had previously used.

Neither of these issues were solved with the Z5c.

Samples:

About the Sizes:


I am using JPEG at C1 which is the most detailed, and largest file size per pixel resolution. I had hoped to use 1920 (longest side) as a standard, but I found that some files resulted in file sizes too large to upload. So for today, I am targeting a longest side of 1800 pixels. This is not a "standard practice" for me, so while the images in this report will be fairly comparable, other reports I have posted will not necessarily be so.


These are the first two captures I have ready for comparison. The first surprise is that for UHD video, the Z5c shows a heavier crop. I found that there is a crop from the largest still picture down to the 1080p (Full HD) frame size, and another crop down to the UHD frame size. I still have not gotten around to doing the testing for the Z3c still images, but looking at the these frames, I would not be surprised of the older Z3c actually used the full sensor width, at least for Full HD, and maybe even for UHD. This pair of videos were actually made simultaneously, so the exposure difference is probably caused by the crop which removed most of the sky from the Z5c frame. If I had matched the exposure manually (adjustment is available), the detail and general rendering, aside from magnification, might have been about the same. The data rate for 4K video for both cameras are about the same -- about 60 mbps.

"01a-Z3c-UHD-0005-1x-Rsz-1800-C1.jpg"
- resized to 1800 x 1021

"01b-Z3c-UHD-0005-1x-Crop-1800-C1.jpg"
- Crop from 1160 x 700

"02a-Z5c-UHD-0004-1x-Rsz1800-C1.jpg"
- resized to 1800 x 1021

"02b-Z5c-UHD-0004-1x-Crop1800-C1.jpg"
- Crop from 800 x 650
 

Attachments

  • 01a-Z3c-UHD-0005-1x-Rsz-1800-C1.jpg
    01a-Z3c-UHD-0005-1x-Rsz-1800-C1.jpg
    655.1 KB · Views: 96
  • 01b-Z3c-UHD-0005-1x-Crop-1800-C1.jpg
    01b-Z3c-UHD-0005-1x-Crop-1800-C1.jpg
    555.5 KB · Views: 89
  • 02a-Z5c-UHD-0004-1x-Rsz1800-C1.jpg
    02a-Z5c-UHD-0004-1x-Rsz1800-C1.jpg
    659.2 KB · Views: 91
  • 02b-Z5c-UHD-0004-1x-Crop1800-C1.jpg
    02b-Z5c-UHD-0004-1x-Crop1800-C1.jpg
    507.2 KB · Views: 80
UHD and Digltal Zoom (2x):

To start this report, I am posting the details of the "Z5c" video file from which the 1x and 2x frame samples were taken. Many samples of various zoom magnifications were made on three files each from the Z3c and Z5c. The file stats were similar for the corresponding video clips.

[From Sony Xperia Z5c]:
"MOV_0004.mp4"

Win 8.1 File Info:

Size: 2,082,710,090 bytes
Created June 6, 2021, 09:29:18
Video
Length: 00:05:02
Frame width 3840
Frame height 2160
Data rate 55,011 kbps
Total bitrate 55,168 kbps
Audio
Bit rate 156 kbps
Channels 2 (stereo)
Audio sample rate 48 kHz

VLC:
Stream 0
Type: Video
Codec: H264 - MPEG-4 AVC (part 10) (avc1)
Language: English
Resolution: 3840x2178
Display Resolution: 3840x2160
Frame rate 29.950396
Decoded format Planar 4:2:0 YUV
Stream 1
Type: Audio
Codec: MPEG AAC Audio (mp4a)
Language English
Channels: Stereo
Sample rate: 48,000 Hz


About the sample frame captures:

As I wrote above, I recorded corresponding clips holding the two cameras in my hands simultaneously. Technically, this was a "handheld" setup, though my hands were braced. This was difficult because holding them and operating them was awkward. During the recordings I had problems with the Z5c in particular aborting the camera mode due to overheating. The Globe and Mail newspaper for the following day reported that the high was 33 deg C. That was the probable ambient temperature, and as you can see, cloud cover was less than half the sky, so that direct sunlight was on the cameras often. The result was that the Z5c in particular shut down often while I was trying to record. I do not recall the Z3c having this problem during this test -- it might have, but if it did, it was "rare".

At this point, with the sky essentially outside the picture frame for both cameras, it should be becoming clear that the Z5c exposure looks lighter. I have not checked pixels yet, but my impression was that there was more flare in that camera during this test. In fact, going beyond this test, the flare problem I found in the Z3c does seem to be worse in the Z5c. I believe I know why, but I will try to write that thought up later, after I post even more samples.

The 2x "zoom" seems to be only a little less sharp and less detailed than the 1x basic recording. In general, I believe that what is happening is fairly straight forward. I think that in both cases, the camera starts by saving the sensor data to a buffer and building the frame image (1 sensor per pixel -- based on the crop factor difference). Then it is "enlarged" using their proprietary formula (close to a bi-cubic, but I think slightly modified). Then the frame is processed by the codec (the colour is chroma sub-sampled to YUV 4:2:0) and the result is compressed by the Codec and stored.

It is important to understand that one piece of advice I have heard occasionally is wrong: Some people have said that it is "just as good", or possibly even "better" to do stabilization and/or re-sizing in post. This is not true if chroma and compression have occurred in the sourcing file. Image detail has been lost which cannot be recovered. At this point, any further cropping and compressing will only increase the detail loss over what has already occurred. Having said that, I have often re-stabilized and on occasion, re-cropped camera files during editing. I only do this if it is "unavoidable". It is better to get your cropping and stabilizing done "in camera" correctly. As bad as "in camera digital zooming" is, if you are going to need it, then it is better to as much of it done in camera than leaving it to later. The only time "leaving it to post" can work is when you have "excess detail" that you can afford to lose because you are recording in significantly high detail in camera than you will need.

Z3c Sample Frame, Zoom 2x

"03a-Z3c-UHD-0005-2x-Rsz-1800-C1.jpg"

- resized whole frame

"03b-Z3c-UHD-0005-2x-Crop-1800-C1.jpg"
- detail crop

Z5c Sample Frame, Zoom 2x

"04a-Z5c-UHD-0004-2x-Rsz-1800-C1.jpg"

- resized whole frame

"04b-Z5c-UHD-0004-2x-Crop-1800-C1.jpg"
- detail crop
 

Attachments

  • 03a-Z3c-UHD-0005-2x-Rsz-1800-C1.jpg
    03a-Z3c-UHD-0005-2x-Rsz-1800-C1.jpg
    508.6 KB · Views: 86
  • 03b-Z3c-UHD-0005-2x-Crop-1800-C1.jpg
    03b-Z3c-UHD-0005-2x-Crop-1800-C1.jpg
    423.2 KB · Views: 84
  • 04a-Z5c-UHD-0004-2x-Rsz-1800-C1.jpg
    04a-Z5c-UHD-0004-2x-Rsz-1800-C1.jpg
    471 KB · Views: 85
  • 04b-Z5c-UHD-0004-2x-Crop-1800-C1.jpg
    04b-Z5c-UHD-0004-2x-Crop-1800-C1.jpg
    472.5 KB · Views: 85
4x Magnification:

Exploring these Sony cellphones makes me think of the most basic issues of evaluating cameras in general. What is "acceptable performance"? The obvious answer is that if this is the only camera you have, then you either take the picture (or video clip) or not. "Acceptable" to that extent is irrelevant. On the other hand, if you know your camera is really bad past 2x zoom, then you know how important it is to get close instead of relying on greater magnification. So you decided that you do not want to "accept" the result of the digital zoom if you can move closer. Or maybe I could express it this way: I do not mind 2x magnification, but I have a stronger motivation to move my camera than use 4x or more magnification. At any rate, these are the 4x magnification results for both the Z3c and the Z5c.

In this pair, the Z5c looks just a bit sharper, but if I was really stuck in a situation where I needed this recording, I'd be strongly wishing I had something better to use. No surprises.


Z3c Sample Frame, Zoom 4x

"05a-Z3c-UHD-0005-4x-Rsz-1800-C1.jpg"

- resize x=1800

"05b-Z3c-UHD-0005-4x-Crop-1800-C1.jpg"
- crop from 0,120


Z5c Sample Frame, Zoom 4x

"06a-Z5c-UHD-0004-4x-Rsz-1800-C1.jpg"

- resize x=1800

"06b-Z5c-UHD-0004-4x-Crop-1800-C1.jpg"
- crop from 0,500
 

Attachments

  • 05a-Z3c-UHD-0005-4x-rsz-1800-C1.jpg
    05a-Z3c-UHD-0005-4x-rsz-1800-C1.jpg
    362.6 KB · Views: 81
  • 05b-Z3c-UHD-0005-4x-Crop-1800-C1.jpg
    05b-Z3c-UHD-0005-4x-Crop-1800-C1.jpg
    325.1 KB · Views: 78
  • 06a-Z5c-UHD-0004-4x-rsz-1800-C1.jpg
    06a-Z5c-UHD-0004-4x-rsz-1800-C1.jpg
    395.5 KB · Views: 86
  • 06b-Z5c-UHD-0004-4x-Crop-1800-C1.jpg
    06b-Z5c-UHD-0004-4x-Crop-1800-C1.jpg
    314.7 KB · Views: 80
Sony Xperia Z5c: 6x and 8x

The maximum digital zoom on both the Z3c and Z5c cameras for 4K video is 8x. I am only posting samples from the Z5c. Really, I am just posting these to be complete. The higher the magnification, the worse they look, and the progression is fairly predictable. I am not bothering with the Z3c because the progression in deteriorating quality is essentially the same. There are no surprises.


Sony Xperia Z5c 4K Samples:

6X:

"07a-Z5c-UHD-0004-6x-Rsz-1800-C1.jpg"

- resize x=1800

"07b-Z5c-UHD-0004-6x-Crop-1800-C1.jpg"
- start 1050,2177

I am not sure if I deliberately stopped the 0004 clip for the sake of keeping the file a reasonable length (probably) or whether it aborted due to the phone overheating. The Z5c did overheat quite often that day. I do not think that the Z3c overheated. It might have, but it certainly had less problems than the Z5c.

The first sample frame is a repeat of 4x magnification as a "control". It turned out particularly useful because the composition emphasized the horizontal tree I have included in the earlier samples. The last 8x sample featured a seagull that happened to land as I started that part of the test. My "detail crop" emphasizes the tree again, but the seagull also adds a reference. The lightest pixels in the seagull did not completely clip. The RED channel clipped, but the BLUE and GREEN were "in bounds". Really, the semi-clipped pixels might be recoverable, though I would probably throw away this clip as an exposure error if this had been a serious project.

4X:

"08a-Z5c-UHD-0005-4x-Rsz-1800-C1.jpg"

- resize x=1800

"08b-Z5c-UHD-0005-4x-Crop-1800-C1.jpg"
- from 450,240

8X:

"09a-Z5c-UHD-0005-8x-Rsz-1800-C1.jpg"

- resize x=1800

"09b-Z5c-UHD-0005-8x-Crop-1800-C1.jpg"
- Detail crop from 1100,0
 

Attachments

  • 07a-Z5c-UHD-0004-6x-Rsz-1800-C1.jpg
    07a-Z5c-UHD-0004-6x-Rsz-1800-C1.jpg
    347.4 KB · Views: 82
  • 07b-Z5c-UHD-0004-6x-Crop-1800-C1.jpg
    07b-Z5c-UHD-0004-6x-Crop-1800-C1.jpg
    268.3 KB · Views: 84
  • 08a-Z5c-UHD-0005-4x-rsz-1800-C1.jpg
    08a-Z5c-UHD-0005-4x-rsz-1800-C1.jpg
    380.4 KB · Views: 83
  • 08b-Z5c-UHD-0005-4x-Crop-1800-C1.jpg
    08b-Z5c-UHD-0005-4x-Crop-1800-C1.jpg
    315.1 KB · Views: 81
  • 09a-Z5c-UHD-0005-8x-Rsz-1800-C1.jpg
    09a-Z5c-UHD-0005-8x-Rsz-1800-C1.jpg
    199.7 KB · Views: 84
  • 09b-Z5c-UHD-0005-8x-Crop-1800-C1.jpg
    09b-Z5c-UHD-0005-8x-Crop-1800-C1.jpg
    204.7 KB · Views: 84
Z5c: Field of View (aka Sensor Cropping)

The crop factors varied noticeably between various imaging modes leaving the question of what exactly we were looking at, so I decided to do a specific test of field of view at various settings. I thought this would be a quick and fairly simple test. Of course it ended up with a fairly good indication of what is going on, but it does leave some questions.

In general though, I am disappointed that I ran into problems conducting the test which leaves it a bit in-accurate. I had to make a slight adjustment to the position of the camera in the middle of the test cycle. I am satisfied that the result is less than 1% error in my results, but I would have preferred to start again and redo the photography. Unfortunately, I do not have time to do that, so I am posting my findings as-is.

One discrepancy which I will note is that I am doing all my calculations relative to X-axis measurements, which is theoretically "incorrect". For any photography both the X and Y axes should be considered. In fact, theoretically, the diagonal is the basic measurement for all non-anamorphic lenses. But for most video composition, the X axis is generally the most important, so that is all I am calculating. If you want more accurate numbers, well, nobody is stopping you from doing it yourself.


My "control" is a Sony a5000 mirror-less camera with a Sigma 30mm F2.8 lens. I have not verified the focal length, but is it a widely used and highly regarded lens.

The Target:

This was the main failure of my test. I had intended to use the tape measure, but it was only sharp enough in a few images. So I reverted to the "Color Chart" as a standard target. This turned out to be the right scale, but it was slightly off-center in the early images. I had to re-center it late in the test. I should have started at the "longer effective focal lengths" to ensure I started with a good composition. But as I wrote before, I am confident that the discrepancy is less than 1%.

As for my calculation of the "width" of the Chart in the pictures, the Chart was tilted slightly. I decided to only measure the top edge (upper left corner to upper right corner) and use the differences in the X values without calculating the actual "length" of the upper edge. I felt this was "close enough".


"10-a5k-DSC01335.JPG" [Not Uploaded]
Size: 7,176,192 bytes
Created: June 28, 2021, 15:23:26
Program name ILCE-5000 v1.10

Dimensions 5456x 3632
Bit depth 24
Resolution unit 2
Color representation sRGB
Compressed bits/pixel 3
F-stop f/4
Exposure time: 1/1000 sec
ISO speed ISO-100
Exposure bias: -0.3 step [to preserve white, according to zebra]
Focal length 30 mm
Max aperture 2.96875
Metering mode Center Weighted Average
35mm focal length 45
Brightness: 9.1148375
Exposure program Normal
EXIF: version 0230

"10a-a5k-Rsz-1800-C1.JPG"
"10b-a5k-Crop-1800-C1.JPG"


Control: Sony a5000 w/Sigma 30mm F2.8 lens
Full Frame equivalent 45mm lens field of view

Width of Chart 3162 - 1754 = 1408 pixels (top edge)
Width ratio of Chart to picture = 1408 / 5456
~ 45mm


"11-Z5c_Still-4x3-0001.JPG" [Not Uploaded]
Size: 14,566,079 bytes
Created: June 28, 2021, 15:04:40
Program name 32.4.A.1.54_0_f500

Dimensions 5520 x 4140
Bit depth 24
Resolution unit 2
Color representation sRGB
Camera maker Sony
Camera model E5823
F-stop f/2
Exposure time: 1/1000 sec
ISO speed ISO-40
Exposure bias: 0
Focal length 4 mm
Metering mode Pattern
EXIF: version 0220

"11a-Z5c_Still-4x3-16-9-Rsz-1800-C3.JPG"
"11b-Z5c_Still-4x3-16-9-Crop-1800-C1.JPG"


NOTE: Resizing the original 4:3 to 1800 wide resulted in a file too big to upload, so I have cropped it to a 16:9 file before resizing to 1800. Even then the C1 compressed file was too big, so I dropped the detail level to C3.

Original size: 5520 x 4140
Width ratio of Chart to picture = 691 / 5520
~ 22mm


"12-Z5c_Still_16x9-0020.JPG" [Not Uploaded]
Size: 12,359,997 bytes
Created: June 10, 2021, 12:30:52
Program name 32.4.A.1.54_0_f500

Dimensions 3366 x 5984
Bit depth 24
Resolution unit 2
Color representation sRGB
Camera maker Sony
Camera model E5823
F-stop f/2
Exposure time: 1/1000 sec
ISO speed ISO-40
Exposure bias: 0
Focal length 4 mm
Metering mode Pattern
Digital zoom 2.64 [NOTE: This was zoomed]
EXIF: version 0220

I did not take this picture intending to use it for this report. I took it simply because I wanted a picture of this group of flowers.

The original composition was in 3:4 ratio, but I thought it would look better if it were narrower. Unfortunately, the Z5c does not support a 3:2 format, so I decided to take it as a 16:9 which I could crop later. The composition looked better from a slight distance, so I moved the camera to where the various shapes and positions were best, and then I used the zoom to crop it (EXIF reports that the digital zoom was 2.64x). The size of the original JPEG was the surprise. The 5984 x 3366 ratio implies a longer horizontal coverage than the supported 4:3 image size 5520 x 4140. It implies to me that the sensor array might actually be 5984 x 4488 if it is 4:3 or maybe 5984 x 3990 if it is 3:2, which gives a total sensor counts of 26,856,192 (4:3) or 23,897,160 (3:2). The "4:3" estimate seems more likely.

"12a-Z5c_Still_16x9-Rsz-1800-C4.JPG"
"12b-Z5c_Still_16x9-Crop-1800-C2.JPG"


In 16:9 format the Z5c produces 5984 x 3366 stills. The 5984 width is probably using the whole sensor array at a 1 sensor = 1 pixel ratio. When I was taking these samples, I did not realize this and so, I did not take a 16:9 still for this test at all. However, based on "picture 12", I am estimating the theoretical results:

Theoretical ratio = 691 / 5984
~ 20mm


"13-Z5c-0005-FHD.png" [Not Uploaded]

"13a-Z5c-0005-FHD-rsz-1800-C2.jpg"
"13b-Z5c-0005-FHD-Crop-1800-C2.jpg"


Width ratio of Chart to picture = 232 / 1920
~ 21 mm


"14-Z5c-0007-UHD.png" [Not Uploaded]

"14a-Z5c-0007-UHD-Rsz-1800-C1.jpg"
"14b-Z5c-0007-UHD-Crop-1800-C1.jpg"


Width ratio of Chart to picture = 690 / 3840
~ 31 mm
 

Attachments

  • 10a-a5k-Rsz-1800-C1.JPG
    10a-a5k-Rsz-1800-C1.JPG
    588.2 KB · Views: 79
  • 10b-a5k-Crop-1800-C1.JPG
    10b-a5k-Crop-1800-C1.JPG
    284.6 KB · Views: 84
  • 11a-Z5c_Still-4x3-16-9-Rsz-1800-C3.JPG
    11a-Z5c_Still-4x3-16-9-Rsz-1800-C3.JPG
    863.6 KB · Views: 74
  • 11b-Z5c_Still-4x3-16-9-Crop-1800-C1.JPG
    11b-Z5c_Still-4x3-16-9-Crop-1800-C1.JPG
    688.2 KB · Views: 77
  • 12a-Z5c_Still_16x9-Rsz-1800-C4.JPG
    12a-Z5c_Still_16x9-Rsz-1800-C4.JPG
    1,000.8 KB · Views: 78
  • 12b-Z5c_Still_16x9-Crop-1800-C2.JPG
    12b-Z5c_Still_16x9-Crop-1800-C2.JPG
    749.2 KB · Views: 79
  • 13a-Z5c-0005-FHD-rsz-1800-C2.jpg
    13a-Z5c-0005-FHD-rsz-1800-C2.jpg
    795.5 KB · Views: 77
  • 13b-Z5c-0005-FHD-Crop-1800-C2.jpg
    13b-Z5c-0005-FHD-Crop-1800-C2.jpg
    861.2 KB · Views: 80
  • 14a-Z5c-0007-UHD-Rsz-1800-C1.jpg
    14a-Z5c-0007-UHD-Rsz-1800-C1.jpg
    768.8 KB · Views: 74
  • 14b-Z5c-0007-UHD-Crop-1800-C1.jpg
    14b-Z5c-0007-UHD-Crop-1800-C1.jpg
    474.3 KB · Views: 76
Z5c Flare

I have mentioned that even the earlier Z3c had a problem with flare, but the Z5c was worse. But I have never quantified the problem, and I am not even sure if a meaningful measurement of flare is possible, but the images I have posted so far only begin to show the problem. I have more.

About The Samples:

The sample video clips that I recorded accidentally showed a bit of the problem, and I almost decided to skip them and start over. But the problem needs to be considered. Think about this: If it is a sunny day, and you stand in a spot, and point your camera level, how many degrees around you do you expect to be able to record? You might have some doubts about having the sun in a picture, but that will only be whatever angle of view your lens provides. If you are covering, say, 30 degrees wide, then that should leave about 330 degrees around you recordable. If the sun is 50 degrees off-axis, you might expect not to have a problem with flare. What if you are limited to pointing your camera in the arc of about 180 degrees when the sun is literally behind the camera? If you have used a well designed and well made modern camera, you might find such a restriction absurd, but that is pretty much what I found on both the Z3c and Z5c cameras, and in the newer Z5c, it was worse.

I am posting frame grabs from only the Z5c. The time of day is about 15:00 - 16:00 -- the sun is still near maximum elevation and a bit towards the western sky. I am estimating the camera directions is about 70 - 80 degrees west of north, and maybe around 15 degrees downward. While the Sun is "in front of the camera", it is so far off axis of the actual frame that most cameras would not have any significant problem with it, especially with the angle of view of a 30mm focal length lens (35mm equivalent). But not only is there distinct flare, these particular frames show clear "bands" of flare

"01-Z5c-0003-Flare-C1.jpg"
- the way the frame looked as captured

"02-Z5c-0003-Flare-Adjusted-C2.jpg"
- same frame processed to enhance the flare, mainly by adding contrast

"03-Z5c-0005-Flare02-C1.jpg"
- the way the frame looked as captured

"04-Z5c-0005-Flare02-Adjusted-C1.jpg"
- same frame processed to enhance the flare, mainly by adding contrast
Why are there such distinct bands of green flare (yes they are specifically green, not having to do with the trees in the background)?

Looking closely at the camera, I found that there is a bare metal ring sitting below the filter/protector glass on the back of the camera. It might be a "washer", or it could have something to do with the construction of the lens case. I do not know, but it is bright and probably bare aluminum alloy -- with no paint or attempt to reduce reflection.

"05-Xperia-Z5c-Cam-Sketch01-C50.jpg"

The overall construction looks like my sketch, and it shows up clearly in a photo of the back of the phone.

"06-Z5c-Camera-Crop1920-C1.jpg"

I have never seen anything like this in another phone camera. Not visible in my close-up of the camera, the lens coating is green, and of low quality (it reflects a relatively high amount of light compared to other camera lenses I have).

Light comes into the camera area, reflects off the lens coating, towards the metal ring, and re-reflects back towards the lens and gets into the picture. More light can also come in at a steeper angle and hit the metal ring directly, and be reflected into the lens, and the picture.

Can this be fixed? I have fund reference for disassembling the phone in general, but nothing showing had to disassemble the camera. An "easy fix" would be to paint the ring matte black. Another partial fix would be to add a lens hood. It is unlikely that I will ever try to paint the metal black -- it is too much work and too risky for me to attempt. I might try to make a lens hood.

Considering that this is not an important camera for me (I have better), you can expect that I will not do anything about it. If you run into a similar situation on some phone, good luck. . . .

"Xperia Z5 Compact LCD Display Replacement",
Posted Jun 6, 2017 by "MovilOne Phone Repair" [Length 19:31]
""

The video I have linked above shows disassembly of the Z5 Compact, but it does not show anything about disassembly or repair for the camera module.

[2021-07-21 09:36
I have watched the above video again and it looks like I was wrong about the construction of the camera. It looks like the metal ring is probably part of the filter assembly attached to the back. So if I take off the back, then I "fix" this by painting the inner surfaces of the metal ring with matte black paint. I still have to plans to actually do it though. I do not want to try the disassembly because I have worked with hot air guns before and it can cause damage. Just disassembling it would take practice attempts. I do not have anything to practice on.]
 

Attachments

  • 01-Z5c-0003-Flare-C1.jpg
    01-Z5c-0003-Flare-C1.jpg
    416.8 KB · Views: 72
  • 02-Z5c-0003-Flare-Adjusted-C2.jpg
    02-Z5c-0003-Flare-Adjusted-C2.jpg
    537.8 KB · Views: 74
  • 03-Z5c-0005-Flare02-C1.jpg
    03-Z5c-0005-Flare02-C1.jpg
    310.4 KB · Views: 76
  • 04-Z5c-0005-Flare02-Adjusted-C1.jpg
    04-Z5c-0005-Flare02-Adjusted-C1.jpg
    424.8 KB · Views: 73
  • 05-Xperia-Z5c-Cam-Sketch01-C50.jpg
    1.9 MB · Views: 71
  • 06-Z5c-Camera-Crop1920-C1.JPG
    06-Z5c-Camera-Crop1920-C1.JPG
    191.8 KB · Views: 73
Last edited:
Last Comment (for now):

I have been trying to get motivated to write this comment for quite a while now. When I started making notes on the Xperia 10 iii, I was not sure how much work I wanted to put into that, but I knew that I would take that time from the time I worked on the Z5 Compact. Since neither was a "primary concern" for me, I did not want this work to get in the way of more important stuff. Moreover, I had not doubt that I did not want this work to expand much beyond just looking at the video capabilities, and to an extent, it has to go beyond that much, at least a little.

What Happened?

I should finish this (current) look at the Z5c by going back to the beginning. At the top, I warned that this was not a "good product", so I should state why:

In roughly the first week I had the Z5c, I updated the firmware to the last version which I will call "7.1.1/2". The reason I am calling it this is because if you check the phone's "setup" to check the version, it says "7.1.1". But I do remember the update sequence getting to "7.1.1" and then apparently receiving another small update. Moreover, if you check some of the "CPU-Z" type programs, they report the version as "7.1.2". It seems that there was an extra, small update, which was probably a very small security update, which the third party programs see and report as 7.1.2, but it fails to change the version that is reported by the main firmware. This is "no big deal", but it was nice to find Sony doing something like this, again, probably as a security update, and probably going beyond what they originally planned. That much is good.
[2021-0928 16:07 Correction: the Z5c, when fully upgraded, runs Android "7.1.1/2" not "8.1.1/2" which I originally wrote.]

But after that last update, having a new, pristine phone, I started installing and trying a few programs on it. And at first, everything was fine. But then I installed "Call of Duty". As I have often written, I am not a gamer. But I thought it was important that I test a wide range of programs to see what the phone could do.

The program ran for a while, and the phone got, what I would call "warm". Keep in mind that I was holding it mainly around the edges and only touching the screen enough to control it. I did this specifically to allow maximum air contact to cool the phone. I foresaw that it might have problems if it "over-heated". But the problems I anticipated were "throttling" or possibly a crash. What happened is that the screen touch function became unreliable. I found that I could turn the phone off with the power switch and then on again with the power switch, re-entering the password, and then continue, and the touch screen would work correctly again.

I did this a couple of times, and then quit because it was annoying. When I let the phone cool off and restarted it, I noticed something wrong. Along the top edge of the screen, white light was visible. It looked to me like the parts of the screen had separated. What could cause that? I had a few ideas, but mainly it was likely that heat had caused the parts of the screen to de-form and separate ("warp"). If a physical change in the shape of parts occurred, then that was "damage" and the only way I could be certain of it being fixed would be replacement of the parts. Even if it were done under warranty, I do not have the time to deal with it.

So as far as I knew, I had a phone that had "damaged itself". Note that as far as I was concerned, *I* did NOT do it. It did it to itself. As far as I was concerned I had used it reasonably, and within "normal" expectations, and it had caused parts within itself to become damaged -- probably permanently. Call it what you will -- defective design, defective manufacture, whatever else, it was in my mind "not a good product".

Anyway, I had not intention of sending it in for repair, nor to disassemble it or try to fix it myself. So for the most part, it sat unused. I did run it a occasionally, but handling it delicately, so I would not cause further damage. Then one day, I noticed that the upper left corner of the screen seemed "high" and I pressed down on it gently. The glass pushed in easily. I turned it on and the white light leak decreased. I got bolder and pressed it down -- still gently, and I could see the light leak reduce.

I searched YouTube for repair videos of the Z5c (and other Sony models around that era) and found out that the phone was held together by heat glue. That was no surprise, but after I thought about it, it occurred to me that what probably happened was the glass of the screen had "unglued" from the frame due to overheating. A manufacturer designing something like this has to pick a glue that does not release at too low a temperature, and use enough of it. In this case, it looks like Sony used a glue that released at too low a temperature, or did not use enough of it, or both. I tried squeezing the glass back into the phone more firmly, and it looked much better. But after I turned it off and restarted it later, it was separated again. After weeks of doing this, I found that I had reached a point where the screen in that corner will not stay down, but it does not look so bad. The screen performs almost like new. I do not think that the parts have necessarily "deformed". There might be a bit of that, but I do not think I need new parts.

I do think that if I wanted to try, I could probably fix it myself, but it would be risky, so I will not bother trying. It is NOT "good as new". But the only problem I anticipate now is that the "weather sealing" is probably gone away. That corner of the display will probably leak water or dust if I take it outside, or am otherwise careless. But I have concluded that it is "not as bad as I thought -- but bad enough" Either Sony failed to pick a high enough release temperature for the glue, or did not use enough of it, or both, _and_ the failed to program a good temperature throttle afterwards to "fix" the defect. Yes, a software fix should be possible -- at the cost of some performance.

So that is where I am leaving it for now. It is "not a good product", but depending on my mood, it is "not horrible".

What are My Plans?

I bought this to try and run Linux on it. There are "kits" to run "Linux-like" interfaces over Android, without "rooting". I have installed such a kit, but the instructions were not clear, and I have not gotten it to work yet. On the other hand, there are more extensive kits that need "rooting" to operate. If I "root" the phone, Sony has warned that the 4K video will not function. There may be other aspects of the camera software that will also fail. So that is why I am essentially "concluding" my report here. If I "root" the phone, I might not be able to use the camera further, or at least some aspects of the camera.

Or if I root the phone and install a "complete replacement" software kit, I might not even have any phone software available. So in either case, I might not be able to report further about the camera at all.

Nothing is decided at this point, but it is best if I just say "don't expect more reports on this phone's camera for now."

What is sad about this case is that if I fix the camera's flare problem, and re-glue the screen in a way that keeps it from ungluing itself again, it really is not a "horrible" product. I do think that proper "thermal throttling" should have been included in the firmware updates. With a proper firmware fix, and, what looks like a fairly simple fix of the camera (a little matte black paint in the case) I could have ended up liking this phone. Instead, it is a bit higher ranking than a "paper-weight". That's a sad end point.
 
Last edited:
A Bit More Perspective:

While this topic was mainly about the Z5 Compact ("Z5c") did mention the Z3 Compact (aka "Z3c") quite a bit. This was intentional because it showed the large discrepancy in my experiences with two ostensibly similar products. I should state clearly that I have had my Z3c for years now, and though it has never been my "main phone", I have been using it fairly often. This is because I find it more convenient to split my functions onto different devices. I tried the "do everything on one phone" approach years ago and found that everything became unnecessarily compromised, whereas if I just used another device for some purposes, everything was smoother, and sometimes there were specific advantages, like recharging my main phone while doing something through the Internet with another.

What I should say is that while I would have liked the Z3c to be even better (which in theory the Z5c might have been). It's working "just fine", and I might end up using it for years to come. This is particularly true because Sony and other phone companies are converting to OLED screens, which are not a technology good for heavy usage. Also, laughably, the latest Xperia 10 iii, which I use, and the Xperia 5 iii, are considered "small" phones in this era. I have tablet computers, and I like them, but I prefer phones to be smaller. For me, even my Xperia 10 iii is "large and cumbersome". But if you are considering an "old technology" phone, I have no reason to think that Sony's later products are particularly bad. I might be wrong about that, but I do think that the mistakes in the Z5c were addressed in later products. Looking at the next year's X Compact and XZ series phones, Sony used plastic and eventually aluminum backs, presumably to help cooling, and the processors continued to use more advanced technologies.

I would expect that the Z5 series phones probably damaged Sony's reputation for those products, and I think they have probably just started to recover their reputation in recent years. I do not really keep track of phone technologies these days, but my latest Xperia 10 iii is a very different experience. Because it is a lower cost phone model, I am not wildly surprised by its performance. But so far, it has been quite satisfactory. So I would not suggest judging Sony's phones generally by this particular product.
 

Most reactions

Back
Top