I have a home workstation that I use for a ton of things including hosting, gaming, Oculus Rift, work, file server, and more. It is also my primary Plex Server (specs below). One of the devices that uses it is the Plex app on a Roku we have in the bedroom where my wife watches Gilmore Girls almost exclusively which are smaller 640 x 352 AVI files. Whenever she does, I notice that Plex uses a large chunk of the CPU to transcode.
My Question
With the Roku connected via 5Ghz WiFi (great signal), the server having a 1Gbps ethernet connection, and the files being played being such a low resolution; why it be transcoding at all?
I did some testing today and in Plex’s setting on the Roku, I set it to Force a direct connection and for local connections to be Original. After doing so, I played one of the 640 x 352 (357mb) videos again then walked a room over to the server. After opening task manager, I saw PlexTranscode was using 63.9% of my CPU. The % being used did change frequently, but I assume it did as things were buffered.
The primary way I watch my media is via Kodi which accesses the files over SMB and does all encoding locally therefore besides a touch of network & disk overhead, requires no other server resources. I’d think with Plex set to play directly / original for something low-res, I would have the same experience. Is that not true? If not, would there be any way to perhaps have an option to transcode with my video card instead?
Therefore plex has to step in and transcode the file or you wouldn’t be able to play them.
Your best bet is to either resource your Gilmore Girls files or us any number of tools (see my sig) that will help you convert your files to AVC video in MP4 or MKV formats. If you convert them outside of Plex then your server won’t have to do anything but serve up the files which uses hardly any cpu cycles.
Ahh - that makes sense. Didn’t think Roku wouldn’t support the type. Will look to find an app to convert them. Will look at your thread, but any specific Windows app you would suggest to do them in batch? Thanks!
Hadn’t heard of those before so downloaded, installed a couple of scripts via PIP I found were required, then added a single 608x352 (360mb) avi into the Process folder to test conversion with default configuration. Even with ffmpeg using +80% of my CPU, the conversion was only at 33% 4 minutes in. Using the same source file in Handbrake finished in 94 seconds, though.
Obviously, the settings could be adjusted within the script since I’m going for a low definition, but despite that am still surprised it was going to take as long as it did using the script. Given all of this plus the fact that I forgot to ask something else, let me rephrase my question(s), please…
Are there any benefits to the script suggested over adding them to a queue in Handbrake?
What’s the best way to normalize audio across all of the episodes for this show (using the same or different procedure)? I tried an app a while back (forgot the name) and waited while it processed all episodes from this show which took a while. When she started watching them again, I found that the audio / video sync was not only off, but its time between the sync varied throughout each episode
I can answer them @bzowk
Chances are you probably have Handbrake setup to use hardware compression which is much faster. The scripts can use HW as well but defaults to CPU since it’s still the best overall quality. Unless you changed Handbrake you are using a quality of 23 and the scripts use 17 which takes longer as well. With the scripts you can surely adjust them to your needs.
Some of the major difference is that the scripts do a lot more. They can d/l subtitles, pull subtitles out of the files and convert them to SRT. You can remove audio tracks and subtitles not in a language you care about. It creates a 2 channel audio track as the default to make the file as direct playable as possible on all devices. It does audio normlization, etc
If you have an MKV with h.264 video in it then chances are it will copy the video while remuxing other things.
So it looks at each file individually and does what’s need to that file vs just using a set of presets which may reprocess the video when not needed. The outcome is that every file you convert should be quite consistent across the board.
