Freezing/Buffering on Windows 11 (NPVR app only)

[sub]

Martin is correct that the problem seems to be the EXT-X-MEDIA-SEQUENCE jumps backwards. This is very unusual behaviour. Normally EXT-X-MEDIA-SEQUENCE would only ever increase.

It seems to be happening because the IPTV service is providing multiple servers with the stream, but the streams are at different points for each, so when NextPVR is asking for the next part of the stream, it's sometimes pointed at a server that is still at an older point in the content. NextPVR does have some logic EXT-X-MEDIA-SEQUENCE to cope with this type of stuff, but I think because the different servers are also using different filenames for the same bits of content, it's confusing matters. Realistically, without having access to a stream that is behaving this way, there is too much guess work involved to try making changes. If you want to temporarily provide access, I could see if there is anything I can do about it.

There is a slight chance that the next version already handles this scenario, since there have been improvements around this logic.

Code:

#EXTM3U
#EXT-X-VERSION:3
#EXT-X-TARGETDURATION:12
#EXT-X-MEDIA-SEQUENCE:1318
#EXTINF:12.278933,
3070082_1318.ts
#EXTINF:12.278933,
3070082_1319.ts
#EXTINF:12.312300,
3070082_1320.ts
#EXTINF:12.278933,
3070082_1321.ts
#EXTINF:10.243567,
3070082_1322.ts

Code:

#EXTM3U
#EXT-X-VERSION:3
#EXT-X-TARGETDURATION:12
#EXT-X-MEDIA-SEQUENCE:1318
#EXTINF:12.278933,
3030995_1318.ts
#EXTINF:12.278933,
3030995_1319.ts
#EXTINF:12.278933,
3030995_1320.ts
#EXTINF:12.312300,
3030995_1321.ts
#EXTINF:10.243567,
3030995_1322.ts

[mvallevand]

I wonder if this is why another user is reporting that NextPVR is using extra connections https://forums.nextpvr.com/showthread.ph...#pid573868 ? This would be two not one.

Martin

[nextpvrnoob]

Realistically, without having access to a stream that is behaving this way, there is too much guess work involved to try making changes. If you want to temporarily provide access, I could see if there is anything I can do about it.

There is a slight chance that the next version already handles this scenario, since there have been improvements around this logic.

Not sure how I feel about providing credentials... I would prefer to discuss in private.

If the next version may have some better logic around this, I would be happy to run a test for you.

[nextpvrnoob]

I wonder if this is why another user is reporting that NextPVR is using extra connections https://forums.nextpvr.com/showthread.ph...#pid573868 ?  This would be two not one.

Martin

Hi Martin,

I'm not sure if this is related. My provider gives 5 connections. Through my testing, at one point I was playing from my iPhone, iPad, and NPVR on my PC.
I'm sure if it is the same issue, it would max out the connections for sure.

[mvallevand]

We see in yours that there are two base URL's being used for the m3u8 segments that come after redirects so it depends how long the provider keeps a connection open. If you want to PM me your main URL I can test which URL is being send with the unreleased new unreleased version to see if this is fixed.

Martin

[nextpvrnoob]

Hi Martin,

I PM'd you. We can discuss this there.

Hopefully, we can figure this out.

I'm still having trouble understanding why it's so NPVR isolated while all my other players handle this situation without issue.

[mvallevand]

Sub already mentioned there were changes related to redirects coming in v6 which might address the issue.

Martin

[nextpvrnoob]

Sub already mentioned there were changes related to redirects coming in v6 which might address the issue.

Martin

Hi Martin,

New log as requested. It's not uploading here for some reason. I don't know why.

UPDATE: Uploaded here instead of google drive

[mvallevand]

Attachments need to be zipped

As long as sub is using the second m3u8 I don't see how it could be a redirect to the other server unless cloudfare cdn is doing something.

Martin

[nextpvrnoob]

Thanks for reminding me to zip.
I updated the previous post.

I'm really crossing my fingers that the new version cleans up the issue.

I guess we're really at the point where someone will have to test with the new version