2009/2/9 Stephan Assmus <superstippi@xxxxxx>: > > David McPaul <dlmcpaul@xxxxxxxxx> wrote: >> On 2009-02-09, Stephan Assmus <superstippi@xxxxxx> wrote: >> > Hi, >> > >> > > Audio is supposed to be a Azalia ALC888 Audio Chip. Any thoughts on >> > > a driver for this? >> > >> > You have not mentioned if you tried it with the OpenSound optional >> > package. have you done that? >> >> Not yet. I will try that next. Ok things are looking better on sound. OpenSound works. >> > > Video is a Intel GMA 950. Has anyone else got that one working? >> > >> > IIRC, this is a problem in the app_server when it tries to use double >> > buffered drawing with acceleration (you can disable it in >> > src/servers/app/drawing/AccelerantHWInterface.cpp::SetMode()). >> >> So you get a blank screen? > > No, it should simply disable double buffering, if you disable a #define > there at the top of the function: Hmm, I get a blank screen when booting with the Intel 950. No errors in syslog but the monitor goes off so I suspect a display out of range type error. Something similar happens when I boot on my other machine. It's a Nvidia 7800GT so I suspect that the calculations to choose a resolution for the display are wrong for my monitor or something. > [...] > bool tryOffscreenBackBuffer = false; > fOffscreenBackBuffer = false; > #if 1 > if (fVGADevice < 0 && (color_space)newMode.space == B_RGB32) { > // we should have an accelerated graphics driver, try > // to allocate a frame buffer large enough to contain > // the back buffer for double buffered drawing > newMode.virtual_height *= 2; > tryOffscreenBackBuffer = true; > } > #endif > > status_t status = fAccSetDisplayMode(&newMode); > if (status != B_OK) { > [...] > > Just make that "#if 0" and it will not try to allocate a screen double the > height for double buffering. From what I recall, this prevents certain GMA > based setups from working correctly. If everything goes well, the effect > will be to have more flickering again when GUIs update, but an otherwise > working setup. ;-) I will try this once I get a compile environment working. > [...] >> I am not expecting miracles for performance under Haiku (The CPU is >> unlikely to be up to the task of 1080p) but a machine where I can fix >> decoder issues does appeal to me. >> >> >And if you really, really want >> > smooth scaling in VLC as well, you can hack app_server to check the app >> > signature in ServerWindow::DispatchGraphicsMessage() when handling >> > DrawBitmap() and add the smooth scaling option for calls comming from >> > VLC... ;-) >> >> Not interested in VLC. Trying to improve all the decoders so that >> MediaPlayer works the best. >> >> But first step is getting this machine to a state where I can develop on >> it. Reduce that code->compile->transfer to other machine->test cycle. > > I see what you mean. Hopefully either OpenSound or if you enable the native > HDA driver will give you sound to work with. And the GMA issue should > definitely be solvable so that you get video overlay. Well, I dd'ed an image onto a partition on the machine, used the usb haiku to makebootable and then setup ntldr to load the partition and now I have haiku booting off the hard disk. Unfortunately DriveSetup will not initialise any partitions so I need to figure out a different way to get a full size partition setup. Then I can bring down the haiku source and work on the same machine which would be nice. > One thing I really, really want to have is an M2TS extractor. The 0.9 VLC > series has that. At least it has everything that's needed to play back the > AVCHD clips of my video camera. Don't know if that is due to ffmpeg (codecs > including extractor) or if the extractor is separate from ffmpeg and only > included in VLC. One of these days I want to investigate that if no one > beats me to it, but for now, I need to finish what I am working on first... I started on a Mpeg2 transport stream extractor but it is such a screwy format that I gave up in disgust. I might get back to it once I have audio video sync/seeking issues sorted. -- Cheers David