All Activity

Intel Arc B580 (Battlemage) - ffdetect crashes with exit code 135 (SIGSEGV) - Hardware transcoding not working on Unraid Hardware: Dell R740xd Intel Xeon Gold 6138 Matrox G200eW3 (onboard, card0) – native on motherboard Intel Arc B580 (PCIe, card1/renderD128) Unraid 7.3.1, kernel 6.18.33 Resizable BAR enabled in BIOS Problem: Emby 4.9.5.0 does not detect the Intel Arc B580 for hardware transcoding. Hardware Detection page shows no GPU adapters. Both VAAPI and QuickSync detection fail. Key finding from hardware_detection log: ffdetect version 5.1-emby_2023_06_25_p4 crashes with exit code 135 (SIGSEGV) for both VAAPI and QuickSync detection ExitCode: 135 (VAAPI) ExitCode: 135 (QuickSync) This is a segmentation fault – not just a "device not found" error, but a full crash of ffdetect when attempting to initialize VA-API with the B580. Additionally, the /dev/dri devices inside the container show group owner dialout instead of video: crwxrwxrwx 1 root dialout 226, 0 Jun 6 card0 crwxrwxrwx 1 root dialout 226, 1 Jun 6 card1 crwxrwxrwx 1 root dialout 226, 128 Jun 6 renderD128 What I have tried: Extra Parameters --device=/dev/dri:/dev/dri --group-add 18 – no GPU detected Installed intel-media-driver (iHD_drv_video.so) manually inside container – still crashes vainfo --display drm --device /dev/dri/renderD128 inside container returns vaInitialize failed with error code -1 Tested with official emby/embyserver and emby/embyserver:beta images – same result, exit code 135 Tested with Ubuntu 24.04 intel-media-va-driver 24.1.0 – iHD_drv_video.so init failed error code 18 Tested with Ubuntu 25.04 intel-media-va-driver 25.1.2 – iHD hangs at Found init function __vaDriverInit_1_22, never returns Tested with Jellyfin-ffmpeg 7.1.3 – same iHD init hang for B580 chmod 777 /dev/dri/* and setfacl -m u:99:rw /dev/dri/* – no improvement What works on this system: FileFlows with Jellyfin-ffmpeg 7.1.4 successfully uses hevc_qsv with Intel Arc A310 (Alchemist) Plex successfully detects B580 but transcoding fails at stream start Conclusion: The root cause appears to be that ffdetect 5.1-emby_2023_06_25_p4 crashes (SIGSEGV, exit 135) when trying to initialize VA-API for Intel Arc B580 (Battlemage/Xe2). This is likely because the iHD driver bundled with ffdetect does not support Battlemage architecture. Question: Is there a planned update to ffdetect/ffmpeg with B580 support? The Emby forum thread #135987 mentions a fix was added in 4.9.2.1-beta but we cannot confirm this is present in the current release. Is there a workaround to manually replace the ffdetect binary or iHD driver? EDIT Root cause identified: The SIGBUS crash is a confirmed bug in Intel's iHD driver for Battlemage/B580. A bug report exists directly at Intel's media-driver GitHub (issue #1927 and #1893): "Running vainfo on Intel Arc B580 (Battlemage G21) with the iHD driver results in a SIGBUS crash during heap initialization in RenderHal_Initialize. The issue appears in memset() during state heap interface setup." The fix requires intel-media-driver version 25.3.2 or later (commit e46f78c35dfd11d7d95addf7b0b68685a2010fae). None of the tested Emby containers include this version. References: https://github.com/intel/media-driver/issues/1927 https://github.com/intel/media-driver/issues/1893 Updated question: Does Emby plan to update the bundled iHD driver to version 25.3.2+ to fix the SIGBUS crash on Intel Arc B580 (Battlemage)? Is there a workaround to mount the host iHD driver into the Emby container in the meantime?

@Asdepique777I made it work with the 2:3 aspect, not done yet as i basically just forced the aspect. Picture from my test server: Might have something tonight or tomorrow.

Hi Luke Yes, I'm on 4.9.5 but the issue is still there. It did it again last when starting to play a recorded movie. Thanks, Warren.

Tu es sur que le serveur tourne toujours ? il a pas changé d'ip ? a-t'il une IP fixe dans ton réseau local ou DHCP ?

I’m currently wrapping up the first release version. As soon as it’s ready, I’ll post it here.

This is an ongoing problem for me, to the point that Emby media player has always failed on the first start attempt but comes right on the second. However it won't complete a full start at all today just the black logo screen. Was there ever a resolution for it? My Emby sever is on a NAS and all devices accesss it by either wifi or cabled LAN. Emby runs perfectly fine on all other devices, including remote streaming, except on this particular Windows 11Pro PC (which is running NVDIA graphics card, which looking over posts seems to be a recurrent thread).

@nypauliehas this helped?

OK if you can provide screenshots and sample data then I can probably give a better answer. Thanks.

Hi. Can you try searching for our standard android app (Just "Emby" on Amazon and "Emby for Android on Google) on the same device's app store and see how that compares? Thanks.

@WDPR4100

One doesn't need to - but if you (personally, I'm quite anal when it comes to curating my collection, and I suppose I ain't the only one) want infinitely more control, more providers and more granulation in metadata/artwork fetching (to name just a few), TMM is leagues ahead. And that's OK, it is purposely built software for that exact functionality - Emby has a lot more on its plate.

Just out of curiosity, why would one need to use TMM in addition to Emby?

换成了4.9.5.0就好了

I'd approach this with a hybrid model instead. collections can be server wide, but with a flag to make it private, then a list specify users. two fields in the database required. `private = true/false` `collection_users = [List]` -- i.e. `["user1","user2","user3"]` these can be replicated and pushed down to the local .nfo's as well for permanent storage outside the DB, similar to how <locked> works. making them exactly like playlists (only personnel to each users), hurts the uniqueness of collections, being that that are curated by the server owner. above recommendation lets collections continue to be server admin managed, but allow for user targeting.

getting the same issues. there really is a simple fix here. just give us download start, stop, pause buttons. for direct transfer, original quality, content, there really should be zero issues with this. transcoded/converted content. do that on the server and cache, when ready, try to push to device. if it fails, wait until user opens app and hits "download NOW", then force start the transfer. easy solve. and most of peoples issues will go away. we don't care about background sync or whatever. we just want to be able to start a download NOW, and be able to monitor it's progress. triage this first. worry about background sync and transcoding later.

Serious, its what I was able to afford, i am on ssdi disability

Radeon RX 9060 XT ?? Are you serious, or just trolling?

Some logs from when I was noticing problems as it is playing fine right now, note I have a VPN(ProtonVPN) and I have a tunnel setup to go to emby but it shows the ip for the vpn. When I turn off VPN and stop qbittorrent and then restart emby it plays fine, not sure if its the vpn, downloads, etc embyserver (13).txt ffmpeg-transcode-13195652-ad0b-49fe-b74c-d5aea8a61e49_1.txt

Here are my latest logs hardware_detection-63916280929.txt ffmpeg-directstream-38ecaa04-7d0f-4d17-b878-516cd6635bf8_1.txt embyserver (14).txt ffmpeg-transcode-1f0050f6-e3b6-4218-9490-1139d15ca793_1.txt

It could also be that the "F" in your processor name doesn't support HW acceleration. Software transcoding may also be your problem. Before you buy new equipment, please upload your logs...

Your drive has bad sectors and is starting to fail. Instead of buying from Amazon I would suggest goharddrive or ServerPartDeals. They both sell refurbished drives with 2-5 year warranties. 12TB with 5 year warranty "only" $320 (they used to be $75 before the AI/Surveillance Centre Apocalypse) 24TB with 5 year warranty for $560 I have personally bought from goharddrive in the past, and lots of users on reddit do as well. If you can come to Canada, you can get drives for cheaper shipped out of the states..24TB external for $660 CDN, or $474 USD..

I am dealing with issues with a old USB HDD where some MKV files have been having stuttering when playing on my Google TV 4K box. System Info: in My Gaming PC file Intel i5-13400F with 32 GB DDR4 C Drive 1 TB NVME D Drive 4 TB NVME E Drive 1 TB SSD F Drive(where my media is) USB External WD WD50NDZM 5 TB HDD WD MyPassport 260D USB External 5 TB HDD Model WD50NDZM I am looking to get a new HDD as I feel this one might be having issues due to age and the casing has a crack in it sadly. Looking at these two HDD's and wondering which one would be the better option: https://www.amazon.com/dp/B0DW8ZW47C/?coliid=I1XTYFCNVVVQN3&colid=2LCEWMVHVJCO6&psc=1&ref_=list_c_wl_lv_ov_lig_dp_it or https://www.amazon.com/dp/B084ZTSMWF/?coliid=I3D0ZS2HARWF80&colid=2LCEWMVHVJCO6&psc=1&ref_=list_c_wl_lv_ov_lig_dp_it Or is the problem that I am using a HDD instead of a faster SSD? My Gaming PC.LOG

@radeon Feature Request: Separate Theme Video for Trailers (e.g., theme-trailer.mp4) Description Currently, Emby supports a single theme.mp4 (or theme.mkv) file inside a movie or series folder, which plays background music or a backdrop video when viewing the item's details page. I would love to see an official plugin or native support for a dedicated Trailer Theme Video. This would allow us to separate standard background ambient videos from actual movie trailers, playing a specific trailer intro or preview automatically when browsing a movie. Proposed Structure Example Ideally, the system could recognize a specific naming convention or a dedicated subfolder structure like this: Plaintext Avatar (2009)/ ├── Backdrops/ │ └── theme.mp4 <-- Plays as the standard background video ├── Trailers/ │ └── theme.mp4 <-- Plays specifically when highlighting or previewing trailers ├── Avatar.mkv └── poster.jpg

That is most likely because you are not following the proper naming conventions: Movies, TV, Music ...

Aparece en el sitio Emby Server como traducido al 100% (https://translate.emby.media/projects/embyserver/) ¿como explico que faltan cosas?.