Ah so test15 has possibly fixed it? I had a feeling it would. I was sending data incorrectly and the transcoding process had an easy chance to break. With test15, I fixed all transcoding related issues (that I could see), and my data is now being sent out correctly.
Bug Report Won't show stream thumbnail on Twitch
- Thread starter defunct
- Start date
- Status
Seems to be stuck for me again, if you're up right now Jim: http://www.twitch.tv/directory/game/Guild%20Wars%202 ctrl + f for: "vesu verra"
It's been on that thumbnail for a while now. The guy in there called "mindtrick" seems to have his stuck as well. I tried to ask if he was using OBS, but didn't respond.
It's been on that thumbnail for a while now. The guy in there called "mindtrick" seems to have his stuck as well. I tried to ask if he was using OBS, but didn't respond.
I did notice the thumbnail stopped for a prolonged period, but it ended up fixing itself again...
I also noticed that this version drops a lot of frames... compared to .452a. Not enough to be really worrysome, but around .08% I believe or around 2.5k before I auto reconnected after 3 1/2 hours.
I also noticed that this version drops a lot of frames... compared to .452a. Not enough to be really worrysome, but around .08% I believe or around 2.5k before I auto reconnected after 3 1/2 hours.
Gol D. Ace
Member
no problems so far with 0.454a (test15).
legal reasons or why did you removed aac?
legal reasons or why did you removed aac?
The only time I personally get a black thumbnail is from game source, but it does have the images I have covering up my game chat etc...sometimes the first one would not have the game capture but it would update after.
But yeah as Bensam said, it would miss a couple updates, then update later, then sticks on one for the rest of the stream.
I tried the new ffsplit to compare and it does indeed update the thumbnail more frequently, but I doubt that helps you, Jim.
Either way it seems to always degrade the thumbnail generation over time, perhaps meaning it's some kinda of frame/timestamp issue?
But yeah as Bensam said, it would miss a couple updates, then update later, then sticks on one for the rest of the stream.
I tried the new ffsplit to compare and it does indeed update the thumbnail more frequently, but I doubt that helps you, Jim.
Either way it seems to always degrade the thumbnail generation over time, perhaps meaning it's some kinda of frame/timestamp issue?
If you guys can help me do some tests in chat some time when I happen to be around, I could possibly try to pinpoint the issue if it is indeed on my end. We'll do some experiments and see what specifically causes breaking if it's possible. If you guys ever manage to get free time please drop by.
ffsplit just uses ffmpeg, and ffmpeg doesn't really ever generate any erroneous output, and twitch is also based off of ffmpeg internally, so it makes sense that it would work. I don't use ffmpeg so I am doing things "on my own" so to speak, but I should be able to pinpoint the issues if you guys ever have time for testing.
I just have no other choice than to do some in-depth testing with you guys. Would be greatly appreciated (when I'm around, which is almost always). Sorry I didn't suggest doing this sooner, I had been waiting for twitch to get back, but they can't do anything until after the new year.
Let's see if we can get this problem fixed, gentlemen.
ffsplit just uses ffmpeg, and ffmpeg doesn't really ever generate any erroneous output, and twitch is also based off of ffmpeg internally, so it makes sense that it would work. I don't use ffmpeg so I am doing things "on my own" so to speak, but I should be able to pinpoint the issues if you guys ever have time for testing.
I just have no other choice than to do some in-depth testing with you guys. Would be greatly appreciated (when I'm around, which is almost always). Sorry I didn't suggest doing this sooner, I had been waiting for twitch to get back, but they can't do anything until after the new year.
Let's see if we can get this problem fixed, gentlemen.
asmodee1337
New Member
I hope you will fix this soon, it's hard to get viewers when you got a black screen front page :?
It's good you're trail blazin Jim even if they don't work right off the bat. Definitely get kudos for that. XD
To clarify my black thumbnail, I believe this has to do with me not having any input initially. So the first frame is black and stays black, but really it's just a frozen thumbnail.
Keep in mind that my thumbnail sometimes randomly fixes itself even if it's been frozen for like two hours (happened last night). It seems rather random, but at least it updates from a black thumbnail so people think I'm doing something (really all that matters).
To clarify my black thumbnail, I believe this has to do with me not having any input initially. So the first frame is black and stays black, but really it's just a frozen thumbnail.
Keep in mind that my thumbnail sometimes randomly fixes itself even if it's been frozen for like two hours (happened last night). It seems rather random, but at least it updates from a black thumbnail so people think I'm doing something (really all that matters).
Fire - I had a twitch coder tell me that things checked out on his end in terms of what I was transmitting. Though I'm still not sure myself.
I had someone tell me that 0.447a works. That one was actually the update right before I started adding in the youtube uploading fixes. Could anyone else verify if this is the case, or another false-positive from someone? Not sure.
Also, I added an option in 0.458a pre-release 2's advanced section "Allow negative CTS" which should duplicate 0.447a's behavior, though somehow I doubt that this would actually fix it. Still, couldn't hurt to test.
I had someone tell me that 0.447a works. That one was actually the update right before I started adding in the youtube uploading fixes. Could anyone else verify if this is the case, or another false-positive from someone? Not sure.
Also, I added an option in 0.458a pre-release 2's advanced section "Allow negative CTS" which should duplicate 0.447a's behavior, though somehow I doubt that this would actually fix it. Still, couldn't hurt to test.
- Status