I'd start by running SFSDoctor on your partitions. Certainly not an error I've seen before - especially since nothing in Wayfarer uses clang.

Quote:
I knew I shouldn't have used that word "clang" - it's just the best way to describe the sound effect I have for a system fail noise, it's not a bell sound it's more of a "clang" sound - so you know something went wrong, but not a glass-smashing sound as some users use I suppose.

Anyway, I just deleted the old(er) folder and replaced it with a new installation and all is OK again, and the video files are playing once more, and the video is quite fluid/continuous, so that's quite good for a 1.5GHz MacMini G4 - and/or your video implementation!

[Update] - just had to re-install Wayfarer v1.12 yet again with a fresh installation as I got an "error" sound on startup yet again, and videos/audio don't work in Wayfarer v1.12 anymore with that installation (fixed with re-installing) but playing/attempting to play another video also generated a Wayfarer Crash Log -> sent.

Should Youtube videos work also in X5000, or is running them one way or another Altivec dependent in Wayfarer...? In my machine Wf crashes as soon as I try to open a video ("Wayfarer Crash Handler / It looks like Wayfarer has crashed :( ..."). I have chosen the items recommended here from preferences.

Quote:

I have the same problem with my sam460ex (with radeon x1550).

@jacadcaps:
Spoofing to Safari/Ipad solved the problem in my case! Works now fine (even at 1080p)!! Cpu-load mostly below 50%.
Only german catalogfile has an issue: Media-Menue is'nt displayed correct.
Thank you for hints!


Btw: Why does'nt it play this embedded video (https://orf.at/stories/3209563/)? Shows "still-pictures" only, and sound is played)

[ Editiert durch Amigaharry2 18.04.2021 - 10:12 ]

Ok, I tried 720p Youtube on a Mac mini 1,25 GHz. Of course with some stutter here and there, but very usable at 70% cpu load. And no problem on a G5, too. Very impressive.

[ Edited by analogkid 18.04.2021 - 12:02 ]

I'm impressed! Runs on PEG2/1GHz (360p) flawless with about 60% CPU-load........

Works fine here! Only thing I've noticed, if screen saver runs and returns to desktop, screen is black and reloading won't help to get video back, it needs to restart wayfarer.

Quote:

Yes, it looks like ffmpeg does not work without Altivec right now. I'll see what I can do in future versions.

Quote:

Does it fail to start after a reboot? Do run SFSDoctor. Also, try to get a Snoopium log from starting Wayfarer.

Logging to "Protonmail" (old.prrotonmail.com) does not seem to work with Wayfarer... After giving password the 'wait' indicator on the login screen rolls a few seconds, then it stops and Wayfarer freezes, as well as the whole system. Only way out is to press reset button.

I also tested the different spoofing options but without success. Has anyone else tested this...? Any suggestions? It is possble to create a free mail account on that site.

Looking forward to try the next version with Youtube-support for non-altivec machines (X5000, Sam460 etc.).

Quote:
Not sure what you mean exactly. If I reboot and then launch Wayfarer, after I've re-installed it completely then it works OK.

I should also mention that I do not run/launch Wayfarer off/from a SFS partition, it runs from a USB FAT32L partition, as it has done since the early releases of Wayfarer, say from about v1.4 - without any issues, other than those reported for the issues than those interim versions exhibited during their release date(s).
Quote:
I cannot run it on the Wayfarer partition as it's a FAT partition, but the SYS: partition reports all OK - which is obviously where MOSSYS and the other SYS components are installed/run from.
Quote:
I did, but when I run Snoopium, and then immediately afterwards start Wayfarer, the Wayfarer Ambient Splash loads, then it stalls, as does Snoopium, and eventually the system stalls/becomes inactive so I have to reboot, and I never get to save a Snoopium log or get Wayfarer to load - that also applies to v1.11 as well as v1.12, probably other earlier releases as well, but I haven't tried those, as yet, unless you think there is some benefit to be obtained by doing so.

I have actually got a Snnopium and the SFS log files, so I will send you those by email, for you to see if there is anything in there to diagnose what the issue is, if any trouble is shown in that file.

Currently Wayfarer v1.12 is running as it should and will play videos/audio as before, but sometimes it's unreliable, loading audio but not video, sometimes video but no audio, and if I load one video then a subsequent one on the same page won't load either at all, or one or the other audio or video components, as though the video/audio loader is not being released from the first file requested to play and allowed/forced to stop the first one and play the 2nd one requested by me, the user.

YT: Is there any posibility to increase the stream-cache a bit? Playback paused sometimes to load (longer videos). This does'nt occur in OWB or Mplayer - downloadspeed is much higher than playbackspeed.

Quote:

No, it isn't something I can control. With Media Source, the buffers are supplied by the website's JavaScript code.

How does OWB prelod then the whole video?

Here you can see what I mean:
When stopping the video, OWB loads continuosly the whole video (Wayfarer stops loading). After that even a PEG2 is able to play it in window-mode....... (seems download sucks much cpu-load)

41014724lu.jpeg

Odyssey does not have Media Source support. I suggest you disable it if you're having issues with buffer underruns.

I was playing around with Wayfarer on the BBC website last night and I kind of got BBC iPlayer to work.

As Jacadcaps said before sound is absolutely fine, but when I dropped the picture quality to low it does play.

What I've found on all BBC videos (both on iPlayer and their main website) is that when they do play there is quite often a spinning circle in the middle of the screen as if it is loading, however the video is playing at the correct speed so not sure if it is just because it is buffering.

So impressed with how quickly this has been developed by Jacadcaps and the MorphOS Team. I feel that now its on the edge of making MorphOS a real modern multimedia OS. Just needs that final push and at least on a media side it will be an OS that can handle modern online media.

Donation to support this is on its way. Keep up the great work, what you are doing is really appreciated.

For a while now I cannot even get BBC News videos to play on my mac/iPad here so there's little I can do - and iPlayer doesn't even work in Canada, so don't expect any targeted fixes.

Quote:

Don't worry, I know you have to just be general in your approach as once you start targeting different broadcasters, international broadcast rights get in the way.

The point I was trying to make is that even in this not targeted approach things that I thought would never work have started to work. I mean the overlay YouTube is great and a feels nice and modern when compared to OWB.

Just keep on doing what you are doing on your plan as your achievements to date are truly brilliant.

Wayfarer 1.13 is now out.

• HLS support can now be disabled in settings


• Fixed a regression in cURL code


• Fixed a crash in media playback initialization


• Minor player improvements


• Disabled media playback support for CPUs other than the G4 and G5 family

@ jacadcaps - Most impressive these speedy new version increments, most impressive indeed!

Though I was just trying to load a Formula1.com video in their website and as I clicked the link it crashed. I had quit Wayfarer just before as the video I already tried had just been showing the busy loading pointer/cyclic-arrow with no video or audio - so no success in that regard as yet.

Later I got another failure/crash with v1.13 - Task Manager reported [FROZEN] WkWebView Acinerella Audio Decoder.

This happened as I was watching a video on BBC iPlayer, with issues, if I pause a video I have to reload the whole page to get the audio and video to play from the point where I paused it. However, the Formula1.com site seems a lot worse, more troublesome to play, or fail to play videos with v1.13 than was the case with v1.12, but it may need some choices changing in the HLS, VP9, and other options to "tweak it" to get it to play better, but it would certainly be probably better, if possible to implement, if the video engine could work out internally what type of video/audio data it was being sent and choose the correct options accordingly›.

I did try iPlayer on the BBC and to my surprise it loaded a couple of shows - not without problems as the audio "Load/Download" requester popped up, initially I tried Download which didn't seem to load the audio correctly, so I tried Load, and that seemed to manage better, but when I clicked the link to reduce the video resolution it paused the video, and would not restart the video/audio again, so it's still hit+miss with video/audio working together and being able to pause/restart a video or move from one video to another on the same page without issues, but I was amazed that iPlayer on the BBC was useable - great work indeed.

CPU load on my 1.5GHz MacMini G4 was about 50% for most of the time, but I did have the video option set to low, which is the worst setting possible, as the Medium and Best video options seemed to stall so only the audio continued without pausing, but the image/video went into pause mode all on its own, so either it cannot keep up with the video data throughput or it just loses its way in controlling it, or at least that is the way it seems to me.

I am aware of issues with video / audio start, especially when Youtube changes the quality automatically. Work on this is on-going. Still, if your CPU is too slow to decode the video, it is going to stall, nothing I can do about it.

Meanwhile, 1.14 is out to address a crash we're seeing in the bugreports coming in.

Quote:

I'm struggling to keep up here. I only updated my G5 to 1.13 this morning, my MacMini is still on 1.12 and my Powerbook is on 1.11.

Quote:
Altivec seems to have other problems too in current code (see here). But looks like at least for the 'yuv2rgb_altivec.c'-part this fix probably works.