yes, now i'm not worried about a pc reboot. (still need to test though) this is my biggest issue with docker. I have to manually make sure portainer properly started, restart ah4c, refresh cdvr sources everytime.
edit: @mackid1993
must run ch4c service stop under administrator for the service to stay stopped.
without admin the service will restart automatically.
so even with docker shut down I'm unable to get 2 encoders working with the built in dual hdmi on the nano a1. I think the 2nd chrome instance just takes too long to actually load the stream and something times out.
I may try a usb hdmi later, but I'm going to try and leave my setup alone for a bit.
I moved my setup to an old Kaby Lake system i7 7700 with 16 GB of RAM and my old Quadro P400 and it's working much better. This does not do well under VirtIO.
i've been using ch4c to record news since quality doesn't matter (although it's not bad most of the time)
the browser window does remember if CC was on, and does remember the CC settings, but for me it always displays large font, until I navigate to CC settings, then just by loading that popup, the CC reverts to the saved preference.
not sure if there's anything that can be done with that as this is the first chrome grab project i've used.
@mackid1993
running your latest commit. 
@dravenst
might be time for some type of versioning. update when idle like cdvr would be great!
I've been using admin console to service stop, uninstall, replace ch4c exe, reinstall. then I use a user ps1 to stop the service and let it self restart.
I would have thought you'd be passing the video card through, so that your Windows VM would be using the Windows drivers for that card, rather than VirtIO drivers.
What I discovered is Widevine decryption is CPU bound so passing through the GPU thru works but any other activity on the server creates latency. Confirmed on two Unraid boxes with my old Quadro P400. Eventually video will go slow on and audio will sound like Darth Vader.
What did help was isolating a few cores and pinning them but it still wasn't a great experience for stream stability. I moved to an old Kaby Lake mini itx build my grandparents recently retired (that I built for them nearly 10 years ago) and CH4C is running super well with the P400 in that on bare metal.
Virtualized Alder Lake is worse than bare metal Kaby Lake for this workload. I also noticed high dpc latency in this VM which likely contributed. The nice thing about this machine is 5 display outputs in one SFF box and Windows 11 IoT Enterprise LTSC runs great on it with a decade of patches. I do wonder if anything can be done to make CH4C less sensitive to latency or if it's Chrome.
Edit: I was passing the card through, the issue was widevine software decoding.
bhwaaahhh! I finally had that problem I've read so much about...
DTV times out after 4 hours... I wonder if we can simulate a keystroke to send a heartbeat.
i know on dtv devices the 'exit' key will always back out to the full screen channel.
I'd actually be fine with sending the channel# for the start of any recording even if that channel has already been tuned.
I send a media play which does nothing but keep it alive with nothing on screen with my Osprey boxes.
not sure what i'm doing differently, but my boxes on tv's (no adb) never sleep. even free guest youtube rarely asks on my cat's tv.
I do have the default 4h in ah4c whatever it's sending, but i've never had to worry about it.
DTV just rolled out a setting on the app that changes the time from 4 to12 hours. I checked on the Firestick App and it’s there under Settings, Preferences, Who's watching, select your option (4 or12).
This is only for apps and should not affect our Osprey's. That setting may be there on the web interface but haven’t looked for it.
Ospreys have always had this option however it will keep the device awake and streaming if it's accidentally turned on hence leaving the 4 hr timeout and using my keep watching PR for ah4c.
I did post a new release (0.5.3) with the latest updates and PR's. Thanks @mackid1993 for the contribution on running as a native windows service. I made a minor modification to it to be compatible with running via an RDP session too.
glad it was minor to you. kinda big for me since I had been relying on chromeRD. just yesterday i had an extended outage and chrome doesn't work without internet. lucky I always have a liberkey handy.
Is there any mechanism where CH4C could have hammered my DirecTV credentials in some way and left my account in some sort of weird state? I noticed that CH4C stopped working at some point because I was mainly using my osprey tuners. I noticed one browser session had logged out, and the next day after fixing it I started having just major issues with my account that I've been trying to track down. I was going in many directions as the issues were really confusing and DirecTV support was clueless. They said my boxes were buffering on a symmetrical gigabit line with everything hardwired.
Finally resetting my DirecTV password security questions and resetting my password a second time and allowing all of my devices to deauthorize server side seems to have hopefully hopefully helped. It seemed to be a DRM sort of licensing issue related to Widevine. My streams were tearing down five minutes to the dot after tuning and retuning themselves. The logging I pulled indicated it to be a DRM related issue. This was on my Osprey boxes after I had shut CH4C down.
Is there any mechanism or bug that was fixed potentially after late June that could have led to my account getting maybe hammered with login requests or something like that? Something that could have looked like automation or account sharing. It was on 0.5.3 and 0.5.4 I believe. I'm just curious if this could have been the cause. I have a correlation in timing but not necessarily causation.
I find it interesting that clearing out my devices on my account and resetting my password seemed to fix it and that it seemed to happen when I set CH4C back up again after not using it for a bit.
Nothing would be designed in like that on purpose. If you suspect there's something going on in that regard, I'd suggest forking it and having a look. The DTV portion of CH4C was largely based on what was done in Prismcast.
I did not mean to insinuate anything would be designed like that on purpose please do not take it that way. 
I just noticed a weird correlation in timing. I'm actually genuinely afraid to test any further because my account was so broken. And I think this could have been maybe been the cause based on timing. I am just so afraid to break anything else. It's not easy for me to get a new account from them if I do really break things because I have a promotional offer right now. So if I had to cancel my account and start a new one I would lose promotions.
I was just curious if anyone else has noticed anything weird like this, and also if a bug has been patched since I last paid attention.
What I could do is have maybe Fable take a look at the codebase and just run an analysis against it. I would not want to test anything on my live account but I am happy to burn tokens for the good of the community.
No worries -- we're good. Off the top of my head, I can't think of anything in ch4c or Prismcast that would fall into the category you're describing by design. So if there is an issue like that, I feel reasonably confident it would be a bug.
That would be great. If there is a bug, it could very well be present in Prismcast's DTV code too.
