Posts by dtech

    menn0

    It looks like a total mess what you're doing on the console. :|

    Even though the binary doesn't start due to an error, the service is probably still running in the background. That's probably why you can't delete it because it's currently in use. First you need to stop that:

    systemctl stop service.tvheadend43

    But why do you want to delete this service file? I don't see any sense in this.

    And, I don't even understand that symlinking. This will in no way lead to results.

    Installing, deleting and updating binary addons without the GUI is a quite problematic.

    ill install it later tongiht and report back, I hope it all works, or else my wetek simply imploded

    Mainly because of the distance, I also recommend another addon, which is also available in my repo, and this is the VNC addon.

    With this, you will also see the GUI, provided that there is a display on the box:

    https://libreelec.dtech.hu/packages/9.2.0/WeTek_Play/arm/service.system.aml-vnc/service.system.aml-vnc-9.2.0.106.zip

    Ai, since I run the whole operation through cli, some 60 miles away, that is not an option unforutnately. What is the source of this 4.3 addon? Can I manually mine this repo? A link would be appreciated :))

    Unfortunately, the directory index on my server is hidden under the package repositories (due to the search robots), so I can only provide a direct link to the binary at most:

    https://libreelec.dtech.hu/packages/9.2.0/WeTek_Play/arm/service.tvheadend43/service.tvheadend43-9.2.0.225.zip

    Found this chip es7134lv in my carte mere wp2 I think it has to do with the sound . Could you be the cause of the problem? Because I removed it and turned on the device without it. The device works normally without it, nothing happened. I mean, as before, the sound and display work from Android et LE 9.2 but LE 11 screen DVI and not sound

    This is a DAC chip, which is responsible for the stereo analog sound on the A/V output (4-pin 3.5mm Jack), so it was completely unnecessary to desolder it.

    I think you should simply change the type of the input in the menu of your display (from DVI to HDMI) and there would be sound immediately.

    I also ran into something similar on my 22" LG Personal TV, and when I set it to HDMI, the sound came on as well.

    To me, hanging at 30-40% processor usage doesn't mean that the SoC is weak, but that the decoder is poorly optimized.
    Maybe the newer version will bring more performance improvements, or maybe rav1d.

    Unfortunately, it is no coincidence that dav1d v1.0.0 was included, and not a newer version. Even for v1.0.0 it was not easy to backport and test the necessary ffmpeg modifications, but for v1.3.0 a lot of things would have to be modified. Since Leia's ffmpeg is an old version, even if I wouldn't call it an impossible task, it's not worth spending so much time, because I don't think the situation would improve much.

    However, dav1d v1.3.0 is already present in the LE12 AMLGX image, so if you want to give it a try, I recommend you try that.

    I want to keep using the old device so it doesn't end up in the trash, not buy a new one, if software solution is possible.

    No one suggested throwing it into the trash, but compared to H.265, I don't know how much better the AV1 codec really is, especially because the hardware acceleration of HEVC is still available on S905. By the way, software decoding always causes unnecessary heat dissipation for the SoC, and knowing this, I no longer see the point, because it is no longer energy efficient anymore.

    Many thanks for your work. Unfortunately the playback is no better, maybe even worse, like a slide show. During playback, the hardware is not fully utilised at all, with the processor cores running at 30-40%. I suppose it's not govenor dependent, clock speed is constantly at 1500MHz.

    In the meantime, I also downloaded a couple of AV1 test videos, and the end result is quite miscellaneous. There are videos that have become smoother with the new version, but there are also videos that are much worse. Unfortunately, updating the software does not help, this SoC is really weak for this task.

    Tartarus

    Unfortunately, you should not rely on the results of these bencmarks, because these do not say anything about internal architectural changes, such as memory controller, cache size and clock, internal instruction set, etc.

    AnTuTu and similar tests will never really show the differences between the functions you need, these will only tell you a ratio value, which is an average of everything. Meanwhile, software media encoding functions may be up to 2 times, but even up to 10 times as fast, if some other function degrades this ratio, which has not changed, or has become weaker compared to the previous one.

    I made a snapshot image with the dav1d 1.0.0 and the necessary ffmpeg patch updates, please check if there is any change in AV1 decoding performance:

    Index of /snapshots/20231012_dav1d-update_a74c19c/

    Under all my images, only amremote (remote.conf) works, meson-ir is not supported.

    I modified the config saved from the MiniMXIII stock Android image, I hope it will work.

    Please extract the attachment, copy the remote.conf file to the /storage/.config/ directory and then reboot.

    I would definitely like some feedback, whether it works or not. If it works, also about whether all buttons are good.

    (The function of the mouse button is different, it opens the info.)

    I will investigate the possibility of updating the AV1 (dav1d) codec, currently the original v0.3.1 is in my images.

    Although it is worth keeping in mind that none of the SoCs supported by me are capable of AV1 hardware decoding, unfortunately.

    Excuse me, but what are you talking about? You wrote at the beginning of your first post:

    i am run libreelec 9.2 on wp2

    But it never officially released, only I made community images from that branch, see in this topic:

    dtech
    February 7, 2021 at 2:33 PM

    Please don't be offended, but based on your answers, sometimes I feel like we don't even talk to each other in English.

    Umm, okay. You didn't actually answer my question, but still okay. ^^

    I am aware of this method, because I also use it, but I tried to approach it from a diagnostic point of view.

    But, I don't know anything about CE either (but neither does anyone on this forum), I thought you were using my LE 9.2.8 community version.

    However, 120Hz really makes no sense, because HDMI 2.0 is not enough for that, and the S905 only has this HDMI version.

    Thank you for this my fellow Hungarian :)

    Halihó! ;)

    The essence of the factory recovery method would be that after pressing the reset button, although the internal uboot starts, if it is not damaged. After that, instead of the kernel it calls the recovery.img from the SD card and then from the USB, finally from the eMMC. (Obviously, it boots what it finds first from the previous sequence.)

    It seems that this is either not implemented in upstream uboot for some reason, or maybe some condition is not met, IDK... :/

    There might be a workaround, please rename the recovery.img file on the "unbrick" SD card to KERNEL and try booting it again.

    Sasser

    ...

    AMLOGIC

    ...

    To set expectations, here’s the ‘No’ list:

    • No support for updates from older LE releases (clean install is mandatory)
    • No support for internal eMMC install except WeTek Hub/Play2 and SBC boards
    • No support for SSV6501 and S908CS WiFi chips
    • No drivers for in-box DVB tuners
    • No hardware deinterlacing
    • No HDR to SDR tonemapping
    • No formal support for newer S905X2/D2/Y2, S905X3, S922X, A311D devices (read below)

    chewitt

    Question: Is there any valid reason why the stock bootloader must be replaced for the LE11+ image?

    I'm only asking this because it seems that the box image can run from an SD card under the factory bootloader without any problems, so I see this as an unnecessarily dangerous thing.

    This is exactly why I don't dare to use emmctool, because at the moment it doesn't seem like any kind of "disaster recovery" would work for WP2 and the Hub. One thing is certain, that the number of problematic cases is much more than zero. Knowing this, I consider it a bit risky to release not sufficiently tested images that overwrite the factory bootloader.