Your videos are so informative that I couldn't even dream of switching from unraid. If your youtube videos weren't around I am sure I would have given up on unraid long time ago. Thank you so much for all your tutorials. I look forward to the next one.
not the only option tho, using proxmox here.. wouldnt say it's easier, should also work with other hypervisor solutions or even with just a bare metal distro
Informative but not always right, transcoding loses quality, especially gpu encoding which does not place space savings as priority as it is optimized for real time streaming. Also Av1 is better.
All tutes should be like this; clear, concise, great enunciation, polished. As far as the encoding, I once took a pretty deep spelunking into the HEVC/x265 encoding world and eventually threw in the towel out of it being way more complicated than it seems on the surface, and no desire to go any deeper into it; I just leave it up to the ones who love encoding. There are soo many variables that go into a *quality* 265 encoding job. You can easily lose detail, create banding, and other issues if not done properly, though the size will (usually) be smaller (sometimes bigger, but that's another discussion). If you really want *quality* encodes you can (and probably should) do a side by side frame comparison of the 264 and 265 videos; sometimes you won't notice a difference unless you have a discerning eye, sometimes it's blatantly obvious. Also, compared to encoding in 264, 265 encoding is slllloooooooowwwwww.
And the gpu encoding he suggests trades quality for speed, space is last consideration as it is for streaming, it definitely loses quality if he is using it to save space. Av1 is currently the best, but is slow.
This is superb thank you Space Invader 😍. I have been re downloading my collection in h265, actually often quicker than reencoding, but TDARR looks great, means if something isnt available in 265 is now easily convertible. 👌👌👌👍
Finally getting around to setting this up, my TV shows drive of 12TB is 66% full... figured that ominous, superstitious number means I should work on saving space! Of course, deletion is not an option! Thanks so much SpaceInvader! EDIT: Coming back to say thanks again, I've reduced my media library by a whopping 3.67TB!!! (736 movies and 3,750 episodes of shows)
You sir are just great! Amazing! I was truly considering using Staxrip (on Windows) to convert all my media files to H265 (with CPU encoding). Even though I recently aquired a GTX 1650 Super (TU116 GPU does support B-frames), I read that for qualitysake it's still inferior to CPU encoding and the outputfile is somehow smaller compared to GPU encoding. With that in mind, last evening I even watched your video about Unraid dual boot Windows and passing through a NVME controller so I could have baremetal performance for this little project. And here you are, posting this amazing video today and even announced that you will be uploading a quality comparison video. I say this again, you sir are truly great and amazing!
Been using RipBot on multiple machines and working with a friend to build a custom distributed encoder but it's been slow going on the latter. Going to be VERY interested to see how well this works as a possible way to go. Big thing I need is being able to sleep encoding machines when not in use.
@@SpaceinvaderOne I havent finished the video (and I love all your content this included) but I kinda think you were unfair on the topic of CPU vs GPU encoding. I really think you should mention Intel QuickSync encoding/transcoding vs regular CPU transcoding vs Cuda/GPU encoding.
This entire video is a joke in terms of actually comparing pre and post transcoded files. H265 at lower bit-rates looks nowhere near as good as the source file from a blu-ray. It only looks good in a vacuum where you don't compare A to B. And specially true if you use NVEnc encoder which has piss poor efficiency. NVenc vs CPU is about 50% efficiency. NVEnc files are BIGGER in the end (about 40-50% of source file video size) and look WORSE than CPU encoded files at the same bit-rates. So if you use NVEnc, you end up with files 50% the original size, but looking about 2 steps down. If you are going to H265 encode your files, Use the Slow template in Handbrake, and stick with non-HW accelerated codecs. You'll end up around 25% of the source file size, and about 80% the quality.
@@amlucent Very interested in this myself! This video looks great but I'm going to hold off until I Know more, as my imminent server upgrade is based around QuickSync...
Thank you for this! I have tried to get this going for a while but the complexity just kept getting in the way. Hopefully this will finally help me get it going! 👍
Quick tip: if you run Unraid on a low-power server and have a workstation / more powerful server as a node, you should still have a node on your Unraid server to specifically deal with remux-only jobs (e.g. rearranging streams + removing subtitles etc.). Unless you have some really old hardware, it is always way faster to remux on the same Unraid server instead of using a different physical machine as node (and be limited by network speed). Since Tdarr doesn't yet allow pinning a node to a specific library, you can run 2 server dockers, 1 pinning to a single local node to deal with remux-only jobs. And have the other one linking with powerful nodes to deal with transcoding (and not having to worry about streams out of place and stuff like that).
Great advice! I use an Apple TV 4K as my Plex Client, which has a dedicated Encode-Decode Block for h.265. Nice to know I'm taking full advantage of what I already paid for!
Well a month ahead, I had 4TB of movies and series. A month later of CPU transcoding (Since I dont have a compatible GPU, so I had to do it with my CPU) I saved a whopping 850GB of data transcoding all the media. I cannot see any degredation on my 4K TV. As I said in my previous message, this is a great video which everyone should have a look at!
First off, I love all your videos. I can't get enough of them. I've had no issue getting Tdarr up and using community plugins. They aren't what I need. A video on how to get Tdarr to use Handbrake and FFMPEG would be amazing. I cannot for the life of me see what I'm doing wrong.
Finally got around to setting this up today, thank you! It would seem like there is now an internal Node built into the docker, so if you're doing it all on one box you don't need the second install. Just setting up a test library now before I throw it at the big guns. I'd personally like the multiple sources for a library, and the option to exclude folders. For example, I'd like to set up a library to convert all reality shows with an aggressive compression (multiple folders within TV), and then exclude the same folders from a regular TV library.
@@XLucero1 Cheers! Worked first time with me on Unraid - just added the Nvidia variables to the regular app, filled in the internal node IP and all fine.
So, getting the mapping correct between a docker on an ubuntu server and the docker in unraid took me a hot minute, but it's working! Great video, I already have it hard at work, can't wait to see the storage benefits.
I picked up a quadro P620 on Kijiji $ 200 CAD after watching this video THANK YOU!!! I have HW trancoding in PLex now and setting up TDARR!! I have about 90TB of video hopefully I can save some space!!! I did have to convert to a different plex container but everything worked great since I followed your plex 2 plex tutorial!!
Well. This couldn't have come at a more perfect time. Running out of space everywhere. Don't have enough physical space in my servers not to mention no cash for more drives. Just did a test on a video. From 640mb down to 120mb. Quality is the same. Plus Plex loads the video instantly. Great! Still gotta do more tests on files to build confidence as well as some low end tablets but so far so good!
Thank you for another awesome video. As someone who is just starting out with Unraid, Docker, and the Arrs, your videos have been extremely helpful! Can't wait to keep watching.
Excellent! Thank you! Followed your guide and took about 10 minutes to successfully set up and begin on my library. Only real difference was my use of a P2000 card. Still set to 3 GPU transcodes for now (about 3800 files to go) until I see if there's an impact to remote user transcodes.
whatever you do begin the node name with a letter. I named my 970 and it would show every 5 seconds then disappeared. Named it Nvidia 970 and now works perfectly
Glad to see there's another method to encoding (h.265) videos. My library has been growing steady and the disk space is getting thin. Thanks for the great video, this worked like a charm, currently encoding movies as I write this. Nicely done.
@@KptnKMan360 Yep, average of about 90FPS processing speed on the P400. Would be faster on a P1000/P2000 etc. Went from 5TB free space to 9TB free, just from Tdarr alone. Literally saved me hundreds of dollars in hard disks.
This is a really, really good instructional video. Thank you. I had downloaded the Tdarr docker containers and spun them up some time ago (in Fedora not Unraid). However, I was lost getting started. Your video just got me through all that, even though the instructions were for Unraid, it was simple to apply them to standard docker compose files. (As an aside: Anyone reading this trying to set up the Tdarr dockers with compose, beyond the instructions given at the Tdarr site, FYI I needed to add the "privileged: true", "runtime: nvidia", and "devices: - /dev/dri" options to the compose file.) In my media server, I have an Nvidia GeForce GTX 1660, and I set up the hardware encoding as this video explains. To start with Tdarr and learn the ropes, instead of tackling my whole media library, I made a temporary small library with a few videos, a mini-series with three episodes. The original 3-episode run totals 2:52:28, 720p, H264/AAC, total original size: 2.48 Gb. I ran Tdarr on this small test library with the default GPU (hw) HEVC encoding plug-in, with b-frames enabled. It took around 15 mins to complete, new size: 1.50 Gb. Then I re-ran Tdarr on the library (with the original files back in-place), this time using the default CPU (sw) HEVC encoding plug-in (the one you remove in this video). It took around an hour-and-a-half, new size: 1.25 Gb. It's interesting to note in my limited test runs that the sw encoding does a better job at compression than the hw encoding. However, the sw encoding being about 6x slower than the hw encoding. I'll need to ruminate on this...is the extra time cost of sw encoding worth the extra 17% space savings over hw encoding?
Thanks for these videos helped me setup my tdarr i have alot more files to do then you did but I've already saved alot of space ..... I have over 160,000 files to encode so this will save me alot of money for buying new drives for a while
Thank you so much for this tutorial, TDARR is encoding right now on my Unraid server and, as promise in your video, I saves disk space. Thanks again !! 3>
tdarr is one of the most baffling web apps I have ever seen.. 80% of the config that should be done automatically by the system has to be done by the user, and if you get it even slightly wrong, nothing will work. Just amazing that anyone gets this thing configured right at all.
i have been looking at tons of reddit post about transcoding to 265 using various programs like unmanic, fileflows and tdarr and one thing that seemed to be very consistant between all the threads was to NOT use a GPU but to instead use the CPU instead. the main comparison being that CPU encoding results in more compression and higher quality over GPU encoding. Can you provide feedback on this? because your videos are confusing and contradicts others real world experiences.
Great video. Your videos are alway extremely helpful. I am a new unraid user and I am following along to setup my server. This video was very clear and detailed and I was able to setup the tdarr app easily with your instructions.
Your tutorials are without a doubt the best for Unraid and it's applications. Thank you. I found with the GTX 1070 that the FPS per transcode was between 150-200 and that I could run between 4 and 6 before I ran into any issues.
@@AvEtheDJ it's a bit finicky, sometimes my fps drops to less than 60 other times it can be as high as 1000. Check your CPU usage, sometimes it ends up CPU transcoding anyway or some files don't work great. Reduce it to one worker and let it go through those slower files and see what codec it's using if after clearing the slow ones it starts new files and they're also slow try restarting the docker, if that doesn't work reboot the whole server and see. If it's still not working delete the worker node and start it all over again and carefully follow the steps.
@Spaceinvader One --- you mention using a GTX/RTX/Quadro card to do the conversion to HVEC/X265 ... in my experience (handbrake/FFMPEG) doing NVENC(although much faster) vs doing software results in a bigger end file with technically more artifacts(if even noticeable) if the priority is on saving file space and not a balance of speed vs space , that should be considered. also using X265 10bit vs standard X265 results in relatively the same file size but less color banding on conversion of non-HDR content.
if you look at the testing they did(handbrake) they could process at 60fps vs 16 for HEVC , however the hardware ended up resulting in considerably larger files and bitrates. where the end desired result is file size reduction and there is no true urgency and it does appear TDARR is using FFMPEG , you may want to consider testing files in TDARR in both software and hardware for end result file sizes and quality since the speed needed for on the fly transcoding isn't really warranted for this project and the increased file size reduction would be very beneficial.
Yep I agree. CPU transcoding will give a smaller file and quality wise is better. But for me the higher power usage of cpu transcoding and the longer time it takes puts me off. Electricity where i am is 21 pence per kwh. When using gpu transcoding for 3 media files at once when transcoding using a p2000 my server will use an extra 20 watts whilst working. However cpu transcoding the same, will take far longer and the server uses aout an extra 80 watts when doing same work with the cpu. For me gpu transcoding is about power efficency and speed. thanks for watching :)
@@SpaceinvaderOne did you account for the cost of how much that would be impacted by add'l HD purchases? a proper H264 software configuration would run much faster then HEVC and not have very much size impact as testing showed the same file would potentially be almost as much as twice as big /w Hardware. [466mb HEVC software / 476mb software H.264 (which runs at a much higher FPS transcoding rate) vs NVENC showing 906mb HVEC & 1024MB H.264 respectively ] . but then again with the way the worlds going i imagine energy prices will be on the rise again soon. --- or configure with some form of average bitrate NVENC so long as you're willing to live with minor dips in quality from time to time to get the best of both worlds, with a minor sacrifice to the archive.
Thanks for explaining the Tdarr settings in this video, I'm not using Unraid or a container, but was still helpful. Just doing some testing now has some issues for a few files error. I will setup a dedicated server container on my Synology and use one of my other PC lay around as a node. I also need to research CPU/GPU as the Tdarr subreddit has a thread discussing using CPU over GPU.
Thanks for this great guide! Can you provide some specific details on how to get this to work with QSV? I have an 8300 Intel CPU and transcoding works with Plex just fine. I added a device to the node with /dev/dri as the name/value. I can't get tdarr to transcode using the cpu. Do you need a specific plugin (I'm just using the default)? I only see two QSV plugins but neither meets my need. I'm just looking for a behavior of what the default does (cover to 265 and leave the audio).
Was able to add both GPU's in my unraid server to the same tdarr node. Just had to add a comma after the first GPU id then paste the second GPU id. Verified both worked via watch nvidia-smi on terminal. Hope this helps anyone that was wondering or ran into problems!
You added the comma and second GPUID inside the NVIDIA_VISIBLE_DEVICES variable "value" field? I'm setting this up now and assumed a new variable would need to be added with GPUID value for the second card.
@@ericbaird8694 Yes. I added the second GPUID inside the same NVIDIA_VISIBLE_DEVICES variable. You dont need to add a space after the comma. Ex. GPU-3409dd85,GPU-3409dd86 Hope this helps! Update: I found a better way by just adding a second tdarr_node. Make sure you update the following on the second tdarr_node: - Name of second container at the top of container Ex: 'tdarr_node2' - 'Node ID' - 'Node Port' - 'Configs' & 'Logs' Path - 'NVIDIA_VISIBLE_DEVICES'
Just wanted to say thank you sir. It is not hard once you can see it being worked out. You did go a bit fast but that is what the timeline is for. Thank you. Just wish I knew what plug-in is the best for my particular set up (Just want if 4k make it with no extra language besides ENG best quality so I do not need transcoding for Plex) Of course I put that here in case someone knows and wishes to tell me :)
This is something I have been wanting to do for a while and this is a great way to do it all automatically. It is also making me appreciate the 1650 Super I purchased even more. Easily the best bang for your buck encoding/transcoding GPU. it has full turning encoding and when I bought it at MSRP when it came out for plex it was a great value too. thanks for the video
That card only has a 128 bit memory bus, so it can't transcode 4k blu-ray remux files in Plex realtime without buffering. Just doesn't have the memory bandwidth. You really need at least a 1660 ti/super to support full 4k transcoding in Plex. If you don't have 4k in your collection, that card is probably fine.
@@diablo81588ik im replying after a year but imo you are wrong, im using a 1650 with turing and im able to transcode 4x 4K remuxes real time, i start getting frame drops on 5th transcode, there is no problem with 128bit bus
It looks like the node module is now bundled with the tdar image "(Tdarr_Node included) Tdarr V2 is a distributed transcoding system for automating media library transcode/remux management and making sure your files are exactly how you need them to be in terms of codecs/streams/containers and so on. Put your spare hardware to use with Tdarr Nodes for Windows, Linux (including Linux arm) and macOS. "
GPU transcoding is certainly faster and more efficient as the video is saying, but in my experience, the resulting file size is much smaller with CPU transcoding (H.265 that is)
I set up my Docker containers like yours and there seem to be some glitches with the standard NVENC plugin. At least one test MP4 video had a 75% file size reduction which is WAY beyond the "up to 50%". Viewing this video, every time a new camera angle was used the first several seconds of the video was pixelated garbage. I then re-did that file to see if it was just a fluke and it was not. I then installed the other NVENC plugin you showed but did not enable and disabled the original one and that video came out as expected with no digital artifacts like before. What exactly went wrong and has anyone experienced this before? The video was a 1080p CZcams video. I am definitely glad I duplicated a few videos for testing.
Absolutely a great video. Thank you so much for doing what you do with your videos. I always come to your channel to get stuff installed and working. And this video again is a gem! Unfortunately my crappy video card does not support NVENC and will have to do all using my CPU, but I think it is worth it. This really prevents me from needing to buy more harddisks this coming year! I´ll send you something through paypal! Keep up the great work!!
@@SpaceinvaderOne funny enough, when I did let it rip on my movie and TV collection it corrupted my cache pool (BTRFS made out of 8x NVMe drives) so that was fun... :D At least I got exercise in fixing that. But now I'm looking into, how to set it up better. Looks like that many NVMe drives is not a good idea.
This is really cool and I wish I could justify converting all my media to 265. Although I have a new media server running Plex and an NVIDIA Shield, I still use KODI with a media player from 2012 a lot and on occasion I play media directly on my cell phone or tablet. I have over 50TB of media, for the moment I'll pay the extra few hundred to avoid the hassle.
A month later and I saved 13TB on my collection, thanks for the great video. I also chose the mkv container but Plex is seems is quite fussy when it comes to 4k and HVEC. My transcodes to other devices would fail. I found I had to use mp4 and then my transcodes work, direct play was fine just transcoding to devices like an ipad or iphone.
At 10:29 you mentioned that Plex/Emby would have the transcodes in ram. How do you set that up? I know how to map to the cache drive or a share but how do you map something to ram?
I've been meaning to do this for my media set up but the main issue is that I just don't have enough beef in my server to do the transcoding. I'm running my unRAID on an Intel i3-7100 which does not even have the intel UHD graphics. Just a single 1080p transcode takes forever and is not worth it for now. I'll definitely work this into my workflow once I upgrade my specs down the line, whenever that happens.
If you have other machines you can use as encoding servers then the server power isn't an issue. Many hands make for light work! I use an array of machines with RipBot but am eyeing this for sure!
@spaceinvader coud u do a video , on how i use the quicksync function most newer intel cpu come with nowaday there some out there but cant get it to work .
4:45 you need the pro if you watch 4K UHD content raw. It will stuter on the nvida shield (tube). Only the Pro can handle it. Lots of people had this problem, me included.
I avoid any h264 where possible these days. But I still have a lot of h264 material from the usual scene sources 😅 but this will still be worth while I think.
Great video. Question: I have lots of movies (mkv) with multiple audio and multiple subtitle tracks. Will this be maintained in the target file ? If so is this out of the box or does it need any special setting ? tnx
While this would be possible in theory, currently only Nvidia and Intel GPUs are supported. I don't know how it works on Windows, but Win is supported as well.
I'm also interested in this, but I have all nvidia gpus, all used as passthrough. Does anyone know if the Linux controller is compatible with Windows nodes?
@KptnKMan360 I am currently doing this. Running the tdarr server on my unraid machine and the tdarr node on my separate gaming machine running windows with a nvidia 1070. Shouldn't be anything different if you are running a gpu in passthrough.
for most people they need to be able to stream in 264 for some clients as not all are always 265 capable or 4k. for that you need to focus on your realtime encoding abilities which you will need to spec. a lot of cheap nas can't handle it
Thanks for the great video. I surmise that if the internal GPU is currently configured as passthrough to a VM, then Tdarr will not able to use the GPU encoding.
great video, I just have the problems that the newly created transcoded files don't have the permissions to be moved. Its really annoying to have to set the permissions. Any idea how to fix that?
Great video again as usual! I really enjoy and learn a lot from your content. I followed your directions to a T, however I don't have a GPU yet so I used the CPU encoder and found my 1080p vids were being changed to 480p. Any idea why? I'm glad I noticed before it got too far and have backups. Thanks again!
Thank you very much for your video. I have gained a lot. I am using AMD1700 CPU to decode. I am worrying about what graphics card is suitable for me. I really need a suitable graphics card to help me emby decoding, otherwise the cpu will occupy 100% every time I play the video.
loving the tutorials again...you are like the #1 oracle to go to...But I do have a question @spaceinvader One. does it matter on the graphics card you have ie i have a 1050ti would you process better with not only speed but quality if you have an RTX OMG that is expensive card? also i have had 4 transcodes where the file size is slightly larger. Is this coz i have something worng? Although just with 32 transcodes i have save 20GB :-)
Thanks a lot for this video. Really like it and keeping on buying harddrive isn't very efficient anyway. I have a 3080 and want to use it for transcoding my library. Does the Bframes option have any impact on my users e.g. they need to transcode files more often? Would the 10bit encoding have any impact for some users who don't have an HDR tv? I wouldn't like it if users all had to start transcoding.
Your videos are so informative that I couldn't even dream of switching from unraid. If your youtube videos weren't around I am sure I would have given up on unraid long time ago. Thank you so much for all your tutorials. I look forward to the next one.
not the only option tho, using proxmox here.. wouldnt say it's easier, should also work with other hypervisor solutions or even with just a bare metal distro
Informative but not always right, transcoding loses quality, especially gpu encoding which does not place space savings as priority as it is optimized for real time streaming. Also Av1 is better.
All tutes should be like this; clear, concise, great enunciation, polished. As far as the encoding, I once took a pretty deep spelunking into the HEVC/x265 encoding world and eventually threw in the towel out of it being way more complicated than it seems on the surface, and no desire to go any deeper into it; I just leave it up to the ones who love encoding. There are soo many variables that go into a *quality* 265 encoding job. You can easily lose detail, create banding, and other issues if not done properly, though the size will (usually) be smaller (sometimes bigger, but that's another discussion). If you really want *quality* encodes you can (and probably should) do a side by side frame comparison of the 264 and 265 videos; sometimes you won't notice a difference unless you have a discerning eye, sometimes it's blatantly obvious. Also, compared to encoding in 264, 265 encoding is slllloooooooowwwwww.
And the gpu encoding he suggests trades quality for speed, space is last consideration as it is for streaming, it definitely loses quality if he is using it to save space. Av1 is currently the best, but is slow.
This is superb thank you Space Invader 😍. I have been re downloading my collection in h265, actually often quicker than reencoding, but TDARR looks great, means if something isnt available in 265 is now easily convertible. 👌👌👌👍
I agree. The electricity and time for transcoding is easier to the radarr to just grab a 265 variant that is hopefully not a copy of a copy.
Finally getting around to setting this up, my TV shows drive of 12TB is 66% full... figured that ominous, superstitious number means I should work on saving space! Of course, deletion is not an option!
Thanks so much SpaceInvader!
EDIT: Coming back to say thanks again, I've reduced my media library by a whopping 3.67TB!!! (736 movies and 3,750 episodes of shows)
You sir are just great! Amazing! I was truly considering using Staxrip (on Windows) to convert all my media files to H265 (with CPU encoding). Even though I recently aquired a GTX 1650 Super (TU116 GPU does support B-frames), I read that for qualitysake it's still inferior to CPU encoding and the outputfile is somehow smaller compared to GPU encoding.
With that in mind, last evening I even watched your video about Unraid dual boot Windows and passing through a NVME controller so I could have baremetal performance for this little project.
And here you are, posting this amazing video today and even announced that you will be uploading a quality comparison video. I say this again, you sir are truly great and amazing!
Hi SandWraith0, thanks for the kind comments and thankyou for watching :)
Been using RipBot on multiple machines and working with a friend to build a custom distributed encoder but it's been slow going on the latter. Going to be VERY interested to see how well this works as a possible way to go. Big thing I need is being able to sleep encoding machines when not in use.
@@SpaceinvaderOne I havent finished the video (and I love all your content this included) but I kinda think you were unfair on the topic of CPU vs GPU encoding. I really think you should mention Intel QuickSync encoding/transcoding vs regular CPU transcoding vs Cuda/GPU encoding.
This entire video is a joke in terms of actually comparing pre and post transcoded files.
H265 at lower bit-rates looks nowhere near as good as the source file from a blu-ray. It only looks good in a vacuum where you don't compare A to B.
And specially true if you use NVEnc encoder which has piss poor efficiency. NVenc vs CPU is about 50% efficiency. NVEnc files are BIGGER in the end (about 40-50% of source file video size) and look WORSE than CPU encoded files at the same bit-rates.
So if you use NVEnc, you end up with files 50% the original size, but looking about 2 steps down.
If you are going to H265 encode your files, Use the Slow template in Handbrake, and stick with non-HW accelerated codecs. You'll end up around 25% of the source file size, and about 80% the quality.
@@amlucent Very interested in this myself! This video looks great but I'm going to hold off until I Know more, as my imminent server upgrade is based around QuickSync...
Thank you for this! I have tried to get this going for a while but the complexity just kept getting in the way. Hopefully this will finally help me get it going! 👍
I hope you get it working. thanks for watching :)
Quick tip: if you run Unraid on a low-power server and have a workstation / more powerful server as a node, you should still have a node on your Unraid server to specifically deal with remux-only jobs (e.g. rearranging streams + removing subtitles etc.). Unless you have some really old hardware, it is always way faster to remux on the same Unraid server instead of using a different physical machine as node (and be limited by network speed). Since Tdarr doesn't yet allow pinning a node to a specific library, you can run 2 server dockers, 1 pinning to a single local node to deal with remux-only jobs. And have the other one linking with powerful nodes to deal with transcoding (and not having to worry about streams out of place and stuff like that).
Great advice! I use an Apple TV 4K as my Plex Client, which has a dedicated Encode-Decode Block for h.265. Nice to know I'm taking full advantage of what I already paid for!
Well a month ahead, I had 4TB of movies and series. A month later of CPU transcoding (Since I dont have a compatible GPU, so I had to do it with my CPU) I saved a whopping 850GB of data transcoding all the media. I cannot see any degredation on my 4K TV. As I said in my previous message, this is a great video which everyone should have a look at!
Awesome! Thanks to YOU I am halfway through ONE library and have saved almost 6TB!
Saving space with gpu transcoding is the most lossy, its not meant for that, the space wasn't free.
Like your explanation style! great work on the video!
First off, I love all your videos. I can't get enough of them. I've had no issue getting Tdarr up and using community plugins. They aren't what I need. A video on how to get Tdarr to use Handbrake and FFMPEG would be amazing. I cannot for the life of me see what I'm doing wrong.
Tdarr has changed alot since this video. Any updated guides?
Finally got around to setting this up today, thank you! It would seem like there is now an internal Node built into the docker, so if you're doing it all on one box you don't need the second install. Just setting up a test library now before I throw it at the big guns.
I'd personally like the multiple sources for a library, and the option to exclude folders. For example, I'd like to set up a library to convert all reality shows with an aggressive compression (multiple folders within TV), and then exclude the same folders from a regular TV library.
I will be working on this today! Glad to know it still works.
@@XLucero1 Cheers! Worked first time with me on Unraid - just added the Nvidia variables to the regular app, filled in the internal node IP and all fine.
So, getting the mapping correct between a docker on an ubuntu server and the docker in unraid took me a hot minute, but it's working! Great video, I already have it hard at work, can't wait to see the storage benefits.
I picked up a quadro P620 on Kijiji $ 200 CAD after watching this video THANK YOU!!! I have HW trancoding in PLex now and setting up TDARR!! I have about 90TB of video hopefully I can save some space!!! I did have to convert to a different plex container but everything worked great since I followed your plex 2 plex tutorial!!
Ive been waiting for this videofor ages! Thank you SI1
Hope you enjoyed it!
@@SpaceinvaderOne I wasnt sure of the variables and this is the most up to date tutorial.
Well. This couldn't have come at a more perfect time. Running out of space everywhere. Don't have enough physical space in my servers not to mention no cash for more drives. Just did a test on a video. From 640mb down to 120mb. Quality is the same. Plus Plex loads the video instantly. Great! Still gotta do more tests on files to build confidence as well as some low end tablets but so far so good!
You are insanely good at this man, thank you so much.
Thankyou XDefinitive. I'm glad you like the vids 😀
@@SpaceinvaderOne I can’t imagine the time this takes to prepare, how can I support you?
Thank you for another awesome video. As someone who is just starting out with Unraid, Docker, and the Arrs, your videos have been extremely helpful! Can't wait to keep watching.
I am really looking forward to your video comparing transcode options. Thank you :)
Just went through your video again and once again, very helpful. You the man!
Thank you. Could never figure out how to properly set up until now
First time watching your tutorials, really clear and informative. New sub here
Awesome, thank you!
Thanks so much for this video! With just a used Quadro 400, I've saved at least 4TB of space with no perceivable dip in quality.
What a great option! I can't wait to see the GPU comparison video, so I know which card to shop for (but probably can't afford lol)
Excellent! Thank you! Followed your guide and took about 10 minutes to successfully set up and begin on my library. Only real difference was my use of a P2000 card. Still set to 3 GPU transcodes for now (about 3800 files to go) until I see if there's an impact to remote user transcodes.
whatever you do begin the node name with a letter. I named my 970 and it would show every 5 seconds then disappeared. Named it Nvidia 970 and now works perfectly
Glad to see there's another method to encoding (h.265) videos. My library has been growing steady and the disk space is getting thin. Thanks for the great video, this worked like a charm, currently encoding movies as I write this. Nicely done.
Can confirm. Used tdarr and gained back 40% of space on the array using an Nvidia Quadro P400. Took about a week for everything.
Great. thanks for sharing your results
Damn, that's nice.
Im definitely looking into this.
@@KptnKMan360 Yep, average of about 90FPS processing speed on the P400. Would be faster on a P1000/P2000 etc. Went from 5TB free space to 9TB free, just from Tdarr alone. Literally saved me hundreds of dollars in hard disks.
@@KaldekBoch oooh, that's nice. I have a 3090 and 1080Ti. This sounds like a very nice project. 🤔
This is a really, really good instructional video. Thank you. I had downloaded the Tdarr docker containers and spun them up some time ago (in Fedora not Unraid). However, I was lost getting started. Your video just got me through all that, even though the instructions were for Unraid, it was simple to apply them to standard docker compose files. (As an aside: Anyone reading this trying to set up the Tdarr dockers with compose, beyond the instructions given at the Tdarr site, FYI I needed to add the "privileged: true", "runtime: nvidia", and "devices: - /dev/dri" options to the compose file.)
In my media server, I have an Nvidia GeForce GTX 1660, and I set up the hardware encoding as this video explains. To start with Tdarr and learn the ropes, instead of tackling my whole media library, I made a temporary small library with a few videos, a mini-series with three episodes. The original 3-episode run totals 2:52:28, 720p, H264/AAC, total original size: 2.48 Gb.
I ran Tdarr on this small test library with the default GPU (hw) HEVC encoding plug-in, with b-frames enabled. It took around 15 mins to complete, new size: 1.50 Gb.
Then I re-ran Tdarr on the library (with the original files back in-place), this time using the default CPU (sw) HEVC encoding plug-in (the one you remove in this video). It took around an hour-and-a-half, new size: 1.25 Gb.
It's interesting to note in my limited test runs that the sw encoding does a better job at compression than the hw encoding. However, the sw encoding being about 6x slower than the hw encoding. I'll need to ruminate on this...is the extra time cost of sw encoding worth the extra 17% space savings over hw encoding?
A poor soul like me is using a Synology and Plex. All of a sudden nothing plays. Yet can play on my PC. Just getting turned off from Plex now.
Thanks for these videos helped me setup my tdarr i have alot more files to do then you did but I've already saved alot of space ..... I have over 160,000 files to encode so this will save me alot of money for buying new drives for a while
You make the best Unraid guides. Thank you for this.
Thank you so much for this tutorial, TDARR is encoding right now on my Unraid server and, as promise in your video, I saves disk space. Thanks again !! 3>
Excellent video, thanks for making the time it literally saved me lots of time playing around with tdarr to get it up and running.
tdarr is one of the most baffling web apps I have ever seen.. 80% of the config that should be done automatically by the system has to be done by the user, and if you get it even slightly wrong, nothing will work. Just amazing that anyone gets this thing configured right at all.
i have been looking at tons of reddit post about transcoding to 265 using various programs like unmanic, fileflows and tdarr and one thing that seemed to be very consistant between all the threads was to NOT use a GPU but to instead use the CPU instead. the main comparison being that CPU encoding results in more compression and higher quality over GPU encoding. Can you provide feedback on this? because your videos are confusing and contradicts others real world experiences.
Fantastic video, thank you so much! Saved me from having to buy another HDD just yet.
I am super interested in how this works with subtitles. That would be the only thing holding me back.
Awesome video Spaceinvader One. Really appreciate you sharing this knowledge!
Wow. Did this save me a ton of time and drive space. Great tutorial!
Great video. Your videos are alway extremely helpful. I am a new unraid user and I am following along to setup my server. This video was very clear and detailed and I was able to setup the tdarr app easily with your instructions.
Superb as ever Ed. Going to convert my media. Great tutorial.
Your tutorials are without a doubt the best for Unraid and it's applications. Thank you.
I found with the GTX 1070 that the FPS per transcode was between 150-200 and that I could run between 4 and 6 before I ran into any issues.
I must have something wrong set... my GTX 1070 goes 4-7fps :(
@@AvEtheDJ it's a bit finicky, sometimes my fps drops to less than 60 other times it can be as high as 1000. Check your CPU usage, sometimes it ends up CPU transcoding anyway or some files don't work great. Reduce it to one worker and let it go through those slower files and see what codec it's using if after clearing the slow ones it starts new files and they're also slow try restarting the docker, if that doesn't work reboot the whole server and see. If it's still not working delete the worker node and start it all over again and carefully follow the steps.
@Spaceinvader One --- you mention using a GTX/RTX/Quadro card to do the conversion to HVEC/X265 ... in my experience (handbrake/FFMPEG) doing NVENC(although much faster) vs doing software results in a bigger end file with technically more artifacts(if even noticeable) if the priority is on saving file space and not a balance of speed vs space , that should be considered. also using X265 10bit vs standard X265 results in relatively the same file size but less color banding on conversion of non-HDR content.
if you look at the testing they did(handbrake) they could process at 60fps vs 16 for HEVC , however the hardware ended up resulting in considerably larger files and bitrates. where the end desired result is file size reduction and there is no true urgency and it does appear TDARR is using FFMPEG , you may want to consider testing files in TDARR in both software and hardware for end result file sizes and quality since the speed needed for on the fly transcoding isn't really warranted for this project and the increased file size reduction would be very beneficial.
Yep I agree. CPU transcoding will give a smaller file and quality wise is better. But for me the higher power usage of cpu transcoding and the longer time it takes puts me off. Electricity where i am is 21 pence per kwh. When using gpu transcoding for 3 media files at once when transcoding using a p2000 my server will use an extra 20 watts whilst working. However cpu transcoding the same, will take far longer and the server uses aout an extra 80 watts when doing same work with the cpu. For me gpu transcoding is about power efficency and speed. thanks for watching :)
@@SpaceinvaderOne did you account for the cost of how much that would be impacted by add'l HD purchases? a proper H264 software configuration would run much faster then HEVC and not have very much size impact as testing showed the same file would potentially be almost as much as twice as big /w Hardware. [466mb HEVC software / 476mb software H.264 (which runs at a much higher FPS transcoding rate) vs NVENC showing 906mb HVEC & 1024MB H.264 respectively ] . but then again with the way the worlds going i imagine energy prices will be on the rise again soon. --- or configure with some form of average bitrate NVENC so long as you're willing to live with minor dips in quality from time to time to get the best of both worlds, with a minor sacrifice to the archive.
great video!! I just added TB’s to my NAS… but I’m also going to deploy this
Thanks for explaining the Tdarr settings in this video, I'm not using Unraid or a container, but was still helpful. Just doing some testing now has some issues for a few files error. I will setup a dedicated server container on my Synology and use one of my other PC lay around as a node. I also need to research CPU/GPU as the Tdarr subreddit has a thread discussing using CPU over GPU.
Good video Ed! Thank you for always sharing with us!💖👍😎JP
Thanks alot. i am glad that you enjoyed it :)
@@SpaceinvaderOne 😎
Is it possible to use my quicksync on my intel cpu for this
Another brilliant video! Thanks SpaceInvaderOne.
😀
Your welcome John. thanks for watching :)
Great Video Ed, you explain yourself so well, well done
Just added this to my unraid, works very nicely!
Great AK. Glad its working for you :)
Thanks for this great guide! Can you provide some specific details on how to get this to work with QSV? I have an 8300 Intel CPU and transcoding works with Plex just fine. I added a device to the node with /dev/dri as the name/value. I can't get tdarr to transcode using the cpu. Do you need a specific plugin (I'm just using the default)? I only see two QSV plugins but neither meets my need. I'm just looking for a behavior of what the default does (cover to 265 and leave the audio).
without you i would have given up on Unraid a longtime ago.
another great video just when i needed it!
Solid walkthrough, thank you!
Was able to add both GPU's in my unraid server to the same tdarr node. Just had to add a comma after the first GPU id then paste the second GPU id. Verified both worked via watch nvidia-smi on terminal. Hope this helps anyone that was wondering or ran into problems!
You added the comma and second GPUID inside the NVIDIA_VISIBLE_DEVICES variable "value" field? I'm setting this up now and assumed a new variable would need to be added with GPUID value for the second card.
@@ericbaird8694 Yes. I added the second GPUID inside the same NVIDIA_VISIBLE_DEVICES variable. You dont need to add a space after the comma.
Ex. GPU-3409dd85,GPU-3409dd86
Hope this helps!
Update:
I found a better way by just adding a second tdarr_node. Make sure you update the following on the second tdarr_node:
- Name of second container at the top of container Ex: 'tdarr_node2'
- 'Node ID'
- 'Node Port'
- 'Configs' & 'Logs' Path
- 'NVIDIA_VISIBLE_DEVICES'
Just wanted to say thank you sir. It is not hard once you can see it being worked out. You did go a bit fast but that is what the timeline is for. Thank you. Just wish I knew what plug-in is the best for my particular set up (Just want if 4k make it with no extra language besides ENG best quality so I do not need transcoding for Plex) Of course I put that here in case someone knows and wishes to tell me :)
This is something I have been wanting to do for a while and this is a great way to do it all automatically. It is also making me appreciate the 1650 Super I purchased even more. Easily the best bang for your buck encoding/transcoding GPU. it has full turning encoding and when I bought it at MSRP when it came out for plex it was a great value too.
thanks for the video
That card only has a 128 bit memory bus, so it can't transcode 4k blu-ray remux files in Plex realtime without buffering. Just doesn't have the memory bandwidth. You really need at least a 1660 ti/super to support full 4k transcoding in Plex. If you don't have 4k in your collection, that card is probably fine.
@@diablo81588ik im replying after a year but imo you are wrong, im using a 1650 with turing and im able to transcode 4x 4K remuxes real time, i start getting frame drops on 5th transcode, there is no problem with 128bit bus
@@vyktriiAlso doesn't matter either way as gpu transcoding is a bad way to save space as it optmized for speed not space.
Another great tutorial!!! Before I start on my library does this have any impact on audio codec (DTD, Atmos, Etc), Or 10bit HDR Movies?
audio doesnt change, seems like passthru everything.
Any news on HDR? Thanks :)
nice vid it helps us all keep your great vid coming thanks
It looks like the node module is now bundled with the tdar image
"(Tdarr_Node included) Tdarr V2 is a distributed transcoding system for automating media library transcode/remux management and making sure your files are exactly how you need them to be in terms of codecs/streams/containers and so on. Put your spare hardware to use with Tdarr Nodes for Windows, Linux (including Linux arm) and macOS.
"
GPU transcoding is certainly faster and more efficient as the video is saying, but in my experience, the resulting file size is much smaller with CPU transcoding (H.265 that is)
Or as it turns out with a Qualcomm encoder on a smartphone, because Qualcomm are smarter than Nvidia and Intel.
Greeting from Brazil! Another great video! Thank you for all :)
Hi there Leonardo. Thanks for watching over there i Brazil. i hope you have a great day :)
simply..and always...the best!
I set up my Docker containers like yours and there seem to be some glitches with the standard NVENC plugin. At least one test MP4 video had a 75% file size reduction which is WAY beyond the "up to 50%". Viewing this video, every time a new camera angle was used the first several seconds of the video was pixelated garbage. I then re-did that file to see if it was just a fluke and it was not. I then installed the other NVENC plugin you showed but did not enable and disabled the original one and that video came out as expected with no digital artifacts like before. What exactly went wrong and has anyone experienced this before? The video was a 1080p CZcams video. I am definitely glad I duplicated a few videos for testing.
Absolutely a great video. Thank you so much for doing what you do with your videos. I always come to your channel to get stuff installed and working. And this video again is a gem! Unfortunately my crappy video card does not support NVENC and will have to do all using my CPU, but I think it is worth it. This really prevents me from needing to buy more harddisks this coming year! I´ll send you something through paypal! Keep up the great work!!
Rearranging the letters for tdarr can be fun
Thanks for a great video. Very informative
Thanks, this was super useful for me, :) I just can't stop marveling what I can all do with Unraid so easily :)
Yep Unraid is certainly very flexible. Thanks for watching :)
@@SpaceinvaderOne funny enough, when I did let it rip on my movie and TV collection it corrupted my cache pool (BTRFS made out of 8x NVMe drives) so that was fun... :D At least I got exercise in fixing that. But now I'm looking into, how to set it up better. Looks like that many NVMe drives is not a good idea.
This is really cool and I wish I could justify converting all my media to 265. Although I have a new media server running Plex and an NVIDIA Shield, I still use KODI with a media player from 2012 a lot and on occasion I play media directly on my cell phone or tablet. I have over 50TB of media, for the moment I'll pay the extra few hundred to avoid the hassle.
A month later and I saved 13TB on my collection, thanks for the great video. I also chose the mkv container but Plex is seems is quite fussy when it comes to 4k and HVEC. My transcodes to other devices would fail. I found I had to use mp4 and then my transcodes work, direct play was fine just transcoding to devices like an ipad or iphone.
so why is everyone going to HVEC when MP4 works just as fine? I am so curious.
@@KINGLIFERISM x265 HVEC is a codec, MP4 is a container
At 10:29 you mentioned that Plex/Emby would have the transcodes in ram. How do you set that up? I know how to map to the cache drive or a share but how do you map something to ram?
I've been meaning to do this for my media set up but the main issue is that I just don't have enough beef in my server to do the transcoding. I'm running my unRAID on an Intel i3-7100 which does not even have the intel UHD graphics. Just a single 1080p transcode takes forever and is not worth it for now. I'll definitely work this into my workflow once I upgrade my specs down the line, whenever that happens.
If you have other machines you can use as encoding servers then the server power isn't an issue. Many hands make for light work! I use an array of machines with RipBot but am eyeing this for sure!
@spaceinvader coud u do a video , on how i use the quicksync function most newer intel cpu come with nowaday there some out there but cant get it to work .
thanks for the nice tutorial my friend
Awesome video. I always learn something watching your channel. Quick question, do you need to re-scan your Plex library after this is complete?
4:45 you need the pro if you watch 4K UHD content raw. It will stuter on the nvida shield (tube). Only the Pro can handle it. Lots of people had this problem, me included.
Thank you!!! It worked flawlessly
Your welcome. Im glad its working :)
I avoid any h264 where possible these days. But I still have a lot of h264 material from the usual scene sources 😅 but this will still be worth while I think.
You can save even more space if you AV1. But it requires newer hardware, and a very powerful one at that.
Great video.
Question: I have lots of movies (mkv) with multiple audio and multiple subtitle tracks. Will this be maintained in the target file ?
If so is this out of the box or does it need any special setting ?
tnx
So useful! Thank you!
Just got my tdarr re encoding my whole movies collection. Is it possible to setup a node on windows and use a Amd GPU?
While this would be possible in theory, currently only Nvidia and Intel GPUs are supported. I don't know how it works on Windows, but Win is supported as well.
I'm also interested in this, but I have all nvidia gpus, all used as passthrough. Does anyone know if the Linux controller is compatible with Windows nodes?
@KptnKMan360 I am currently doing this. Running the tdarr server on my unraid machine and the tdarr node on my separate gaming machine running windows with a nvidia 1070. Shouldn't be anything different if you are running a gpu in passthrough.
@@pettsonktm yesssss, thank you kindly, sir. This is perfect. 🤔
@@pettsonktm what did you do to get the windows node working correctly?
Thanks for the tutorial! No GPU in the server for me, time to absolutely SPANK the CPU for a while :)
thank you for this video, it saved my ass since i forgot how to do this and tdarr for some reason decided to nuke itself from my server
Any idea what the variable would be for AMD GPU's? I have a XFX RTX580
for most people they need to be able to stream in 264 for some clients as not all are always 265 capable or 4k. for that you need to focus on your realtime encoding abilities which you will need to spec. a lot of cheap nas can't handle it
Thanks for the great video. I surmise that if the internal GPU is currently configured as passthrough to a VM, then Tdarr will not able to use the GPU encoding.
great video, I just have the problems that the newly created transcoded files don't have the permissions to be moved. Its really annoying to have to set the permissions. Any idea how to fix that?
Awesome!
My Hero
Enjoyed this a lot, hopefully it will spawn mass adoption of HEVC (wishful thinking)
I'm glad you liked it and thanks for watching :)
Awesome intro!!
Great video again as usual! I really enjoy and learn a lot from your content. I followed your directions to a T, however I don't have a GPU yet so I used the CPU encoder and found my 1080p vids were being changed to 480p. Any idea why? I'm glad I noticed before it got too far and have backups. Thanks again!
Thank you very much for your video. I have gained a lot. I am using AMD1700 CPU to decode. I am worrying about what graphics card is suitable for me. I really need a suitable graphics card to help me emby decoding, otherwise the cpu will occupy 100% every time I play the video.
Great cameo!
Absolutely EPIC!
Thanks for watching Colby :)
Amazing video, any chance this would work with an intel integrated gpu? 10th gen
loving the tutorials again...you are like the #1 oracle to go to...But I do have a question @spaceinvader One. does it matter on the graphics card you have ie i have a 1050ti would you process better with not only speed but quality if you have an RTX OMG that is expensive card? also i have had 4 transcodes where the file size is slightly larger. Is this coz i have something worng? Although just with 32 transcodes i have save 20GB :-)
Thanks a lot for this video. Really like it and keeping on buying harddrive isn't very efficient anyway. I have a 3080 and want to use it for transcoding my library. Does the Bframes option have any impact on my users e.g. they need to transcode files more often?
Would the 10bit encoding have any impact for some users who don't have an HDR tv?
I wouldn't like it if users all had to start transcoding.