October meeting notes

Skip to first unread message

Aaron Johnson

Oct 10, 2019, 11:34:54 AM10/10/19
topics: How to stream content with Emby, Transcode video to make it more storage efficient and how to find the best content online.

meeting narative:

The meeting started mostly on time and we had a newcomer so welcome Michael.

We went ahead and did introductions before continuing with the meeting topic

Christian presented his topic first which was the "find the best content online" sub narative...

Usenet Originated in 1979 as a newsgroup network system. Similar to a modern web forum without the rich content.
NNTP was then created for Internet transmissions
It uses well-known TCP ports 119 (standard transmission), 563 (TLS), 433 (server-to-server)

Part of the protocol allows you to post binary content (uses yEnc encoding)

Binary files are broken into smaller .rar files typically and are posted as separate articles. Parity rar files exists as well if the file pieces are corrupt or missing.

These binary pieces of files are actually stored on the News Servers so they have had to grow storage capacity over the years to store these file fragments...

How do you know which articles to download?
The .nzb file is an XML file (map) that contains a list of articles that contain the segments of the file you desire.

There are now Newsgroup Indexer Sites that index .nzb files for easy searching

There are multiple NZB downloaders in the wild, sabnzbd (python) and nzbget (C++) are a few of them.

PVRs or indexer websites can use the NZB downloader APIs to send .nzb files

Next you will want a TV Show library management tool; for all of these tools TheTVDB is used for source information

Sonarr (.net - uses Mono still - formally nzbdrone), Sickrage (not sick but healthy), Sick-beard (dead) and Flexget (Python) are a few of the options...

Also Movie management is also a thing with NZBs so you will need another tool for that! All of these tools use TMDb.

Couchpotato (python) or Radarr (.net) are some examples

Other tools that can help you are things like Jackett which is an API proxy/normalizing site scraper (thanks for the description Cooper!) or Lidarr (.net) which is a music collection management tool.

Next up was Aaron's presentation on "How to stream content with Emby and Transcode video to make it more storage efficient"

First Aaron gave a barely coherent explaination of what Emby is and stumbled to find reasons for why someone should use it over Plex.

All in all I'd say he "may" have convinced one person to give Emby a try but I believe most users are perfectly happy with Plex

Emby is confusing as well due to the change in version that made certain parts of Emby no longer open source... These changes are around the hardware accelerated features for certain devices and Emby felt their hands were tied...
Due to this a fork was created called Jellyfin but Aaron hasn't tested it yet...

That said my opinion on Emby is that it is fantastic! It has feature parity with Plex and is just as confusing in its open source/proprietary nature!

Next Aaron explained that ffmpeg could be used to transcode your video files to save space and bandwidth. He advocated that it is best to either offload the transcoding to another device
by making use of shared storage protocols such as sshfs or nfs but also suggested if additional compute power isn't available that transcoding scheduling might be another option (ie schedule the transcoding to happen in the middle
of the night when no one is using Plex or Emby)

He feels the lack of scheduling or offloading offline transcoding to a separate device are missing features in Plex and Emby and could be a nice addtion if someone were to add these features in the future.

Chris mentioned that the solution from these projects is to make use of hardware accellerated transcoding which is true but isn't always a solution depending on cost and if you choose to do colocated Plex/Emby hosting this becomes an even bigger challenge.

After the transcoding discussion Aaron shared his personal ansible repo that contained some yaml that shows how someone could use ffmpeg on a cron job to transcode or "squash" the video files to preserve space and still retain qualiy.

Aaron doesn't use the best codec though as he had h264 in the defaults but will be testing h265/hevc in the near future per Christian's suggestions/claims that H265 is THE BEST CODEC EVER (exact claim word for word)

Finally we got to see a live demo using a Raspberry Pi 4 (4GB model) and see for ourselves how much better CPU performance was than previous rPi models.

To summarize what we were able to see was the emby-server process spawning ffmpeg transcoding threads that would take up just less than 3 out of 4 cores.

We were even able to spawn 2 more ffmpeg threads out of band from Emby and the Pi4 seemed to be able to keep up still with the simultainous transcoding which seemed impressive.

One thing to keep in mind is that the rPi 4 CPU/APU does generate much more heat and so there are aluminum cases that can be purchased that will help distribute the heat and help
reduce thermal throttling so if you are looking for an rPi 4 be on the lookout for what cases distribute heat the best/most effectively.
Reply all
Reply to author
0 new messages