Just to help clarify the issue, my video does not stop as well. This video shows Edge on left and Pale Moon on right. Both are using the HTML5. Notice that even though the error message pops up the video keeps going.

OK,The videos so far have told me it probably is not i-frame related because it breaks after a different number of frames each time.
I checked myself, Blue Iris did not suddenly start supporting Direct-to-Wire in clips.
Lets check on a few more basics. These questions are for everyone repeatedly having the issue with HTML5 player stalls being detected, FPS readout dropping to 0, playback progress bar stopping moving, etc (all these things happen as a consequence of the same problem).
1. Are you using hardware accelerated decoding in Blue Iris on any of the cameras? (check status window > Cameras to be sure. The "HA" column should be all hyphens.
View attachment 242197
2. Are you using hardware accelerated encoding? (in BI Settings > Web server > Advanced > Streaming 0 (or 1 or 2) > Configure.
3. Is direct-to-disk being used on affected cameras?
4. Does it only affect some cameras (and not others)?
5. Does it affect group streaming (multi-camera views)?
6. Did anyone try downgrading Blue Iris to a version that was not affected, to see if it is still not affected? This would help rule out if Blue Iris is doing something different to cause the issue. Please do not feel compelled to perform this test if you haven't done so already.
Mine does it on both live or playbackSo you're experiencing this on LIVE as well? So far, I've only seen it on playback - and then only if it's 1x. If play back on anything other than normal, it doesn't happen. Interesting.
Sorry for delayed response to your questions, Bryan...The videos so far have told me it probably is not i-frame related because it breaks after a different number of frames each time.
I checked myself, Blue Iris did not suddenly start supporting Direct-to-Wire in clips.
Lets check on a few more basics. These questions are for everyone repeatedly having the issue with HTML5 player stalls being detected, FPS readout dropping to 0, playback progress bar stopping moving, etc (all these things happen as a consequence of the same problem).
1. Are you using hardware accelerated decoding in Blue Iris on any of the cameras? (check status window > Cameras to be sure. The "HA" column should be all hyphens.
View attachment 242197
2. Are you using hardware accelerated encoding? (in BI Settings > Web server > Advanced > Streaming 0 (or 1 or 2) > Configure.
3. Is direct-to-disk being used on affected cameras?
4. Does it only affect some cameras (and not others)?
5. Does it affect group streaming (multi-camera views)?
6. Did anyone try downgrading Blue Iris to a version that was not affected, to see if it is still not affected? This would help rule out if Blue Iris is doing something different to cause the issue. Please do not feel compelled to perform this test if you haven't done so already.

I did not have direct to wire checked there.





I'm pretty sure the D2W avoids the problem by sending the stream straight from the camera to the browser - bypassing all encoding that BI does. As in my case, with D2W enabled, all single-camera live streams works great. But even with that enabled, any playback (or composite/group cameras) will still have the issue, as those are re-encoded by BI in a way that doesn't play nice with the HTML5 player in chromium (which apparently includes Edge).Ok, I think I have found something. I went in and checked under : see pic
View attachment 242241 I did not have direct to wire checked there.
I did have it checked each camera's settings here:
View attachment 242242
When i do check it enabled in the BI settings I cannot not produce the error. But, I know i have been running this unchecked before as I always see the time in UI3, with direct to wire I do not as I don't use the camera's time overlay.
Have to get yard mowed, but just wanted to share that. Also here are just a few of the cams encoder settings :
View attachment 242243View attachment 242244View attachment 242245View attachment 242246View attachment 242247
When I get back I will try the UI3 bp2008 want us to try.
I'll switch to a non D2W profile and report back momentarily. In the interim, I applied the 320 files (though my browser info page still shows 319, so I'm not entirely sure the update took). I've been hitting my playback pretty hard, and so far I've not seen the stalling. Maybe it's a placebo effect
ETA: I didn't wait long enough. It's still doing it. And to your point, when I disable D2W, my live stream almost immediately begins to stall. But, as before the video itself seems unaffected.
<video> element to a canvas at every screen refresh and comparing all the pixels with the previous frame to see if they changed. This would only tell me that the frame changed and not which frame was rendered, which breaks UI3's ability to compute player delay and makes the timestamp and seek/scrub bar inaccurate at the very least. I'm not going to do that.
Ok, I think I have found something. I went in and checked under : see pic
I did not have direct to wire checked there.
I did have it checked each camera's settings here:
View attachment 242242
Man, I wish I had just a fraction of your understanding and patience to explain this to people like me
Thanks for testing. I'm not sure what to do at this point.
It is possible I could detect frame rendering by (very inefficiently) copying the<video>element to a canvas at every screen refresh and comparing all the pixels with the previous frame to see if they changed. This would only tell me that the frame changed and not which frame was rendered, which breaks UI3's ability to compute player delay and makes the timestamp and seek/scrub bar inaccurate at the very least. I'm not going to do that.
I think I never answered your earlier question about adding an option to suppress the stall detection and reload behavior. I'm not sure that is a good idea. Because when the video player is in this bad state, all kinds of stuff is broken. Including the seek/scrub bar. I'll add it if necessary but it feels wrong.
An AI suggested setting the playback position to 1 millisecond after its current position, as it might force the video player to do a seek operation which could (maybe) get it to resume sending progress updates. But it would also cause a stutter I am quite sure. Not just a little stutter either. More like rewinding to the point where the player stopped reporting progress updates, if we do that as an alternative to restarting the stream the way it does now.
Here's something else you could try. Find the Enable Max Delay Limit setting in Video Player (Advanced), enable it, and try setting the max delay to low values both above and below where your player delay would normally sit. This will trigger when UI3 thinks the player delay has exceeded your set value. With luck it could get the video player back out of its bad state. We can hope. It might also cause a pretty bad stutter; you'll have to try to find out.
View attachment 242250
Another possibility. I suppose it is possible that keyframes arriving in the stream will cause the player to resume reporting progress correctly again (it could explain why direct to wire is preventing the issue for you).
BI web server streaming normally uses a very long keyframe interval for efficiency. But if a short keyframe interval works, it could be the best workaround for this issue right now.
Create a new streaming profile in UI3:
- Variable Bit Rate or Constant Bit Rate. Doesn't really matter which.
- Assign a very high max bit rate like 24000 Kbps to mitigate the quality loss of having a low Keyframe Interval.
- Set Keyframe Interval to 10 or something else low like that.
I hope this issue is affecting bigger fish than Blue Iris and UI3. Someone with the resources to get it fixed (like Hulu, Netflix, Youtube).