MediaPortal Audio renderer - better video playback quality (2 Viewers)

[tourettes]

Re: AW: MediaPortal Audio renderer - better video playback quality

Ok, I can understand that, but when I read all these posts, it can took really long to get that fixed by ati :s
And the realtek drivers are only good without usning WASAPI... really nice ^^

I know it takes time, but I really wont feel happy about spending my free time for a free to do some work that HW manufacturers should have been doing in the first - after all it would be just a direct money transfer to their pockets, since they are saving money on QA and development costs.

...I wonder if NVIDIA / AMD people are going to fix ever those issues if they happen to stumble on this thread since it is quite much nagging that there is... but in any case I have spend already weeks debugging & troubleshooting the driver related issues.

Damn drivers, all crap. Ati, nvidia, creative *argh*

WASAPI drivers seems to be in a really bad shape... the working ones are the ones that ship with the OS itself.

 

[Truri]

AW: MediaPortal Audio renderer - better video playback quality

Yes you're absolutly right. It's a shame what these firms are doing with the driver support.
The other problem is, at the moment its ok as it is, because the new audiorenderer is only for testing.
But later its a big disadvantage if they didn't fix that issue.
And in addition to that, what will you do, if nvdida didn't fix their drivers that wasapi will work in any way?
Don't get me wrong, but as i read they didn't respond and all these bugs aren't new ones...

 

[arion_p]

In fact 24bit padded to 32bit is a valid format, and MPAudioRenderer supports it. But since MPAudioRenderer will output the exact same format as supplied for input, it is up the the codec to supply 24bit padded to 32bit samples and we will happily process and output that.

Now when we implement the adaptive bit-depth conversion, this could be one of the formats considered for output and if the driver supports it, it will be used automatically.

One observation, and possible explanation (but no excuse) for ATI supporting this format, is that 24bit packed samples are really slow to process. (Modern) CPUs work on 16,32,64 or 128 bits at a time. To work on 24bit samples you have to get 16 bits, then another 8 bits, pack them as 24 bits in a 32bit container, process them as 32 bits, then take out 16 bits, store them, take another 8 bits, store them, go on to the next sample.

 

[tourettes]

In fact 24bit padded to 32bit is a valid format, and MPAudioRenderer supports it. But since MPAudioRenderer will output the exact same format as supplied for input, it is up the the codec to supply 24bit padded to 32bit samples and we will happily process and output that.

Indeed, so most likely using 32 bit from ffdshow would be better (if the ffdshow just pads the 24 to 32 bits). Just disable 24 bit from ffdshow and it most likely is clever enough to use 16 bit for 16 and 32 for 24 & 32 bit streams.

Now when we implement the adaptive bit-depth conversion, this could be one of the formats considered for output and if the driver supports it, it will be used automatically.

yep, if 24 fails we should next try 32 in any case.

 

[tourettes]

Re: AW: MediaPortal Audio renderer - better video playback quality

Yes you're absolutly right. It's a shame what these firms are doing with the driver support.
The other problem is, at the moment its ok as it is, because the new audiorenderer is only for testing.

For testing it would be good to have working combination

And in addition to that, what will you do, if nvdida didn't fix their drivers that wasapi will work in any way?
Don't get me wrong, but as i read they didn't respond and all these bugs aren't new ones...

Wont probably buy ATI since it has the flickering bug, so most likely use old NVIDIA GT9500 and then SPDIF from MB.

But in any case these speculations start to go off topic. WASAPI is a standard that MS has developed for a reason. I really hope HW manufacturers notice that there is quite big issues with their drivers and they fix those. After all HTPC market is constantly growing.

Summary about those bugs and workaround:

1) ATI - quite easy to work around, no quality issues
2) NVIDIA - requires more work (extra code for the polling based buffer filling), reduces glitch resistency

 

[mironicus]

AW: MediaPortal Audio renderer - better video playback quality

Wont probably buy ATI since it has the flickering bug

ATI don't flicker with the right decoders - Cyberlink PowerDVD 9 MPEG2/H264/VC-1 combined decoder for example.

 

[tourettes]

Re: AW: MediaPortal Audio renderer - better video playback quality

Wont probably buy ATI since it has the flickering bug

ATI don't flicker with the right decoders - Cyberlink PowerDVD 9 MPEG2/H264/VC-1 combined decoder for example.

Some people report it to flicker it as well with those. Also some report that all HW decoding codecs cause the flickering. But in any case this is completely off topic.

 

[Truri]

AW: MediaPortal Audio renderer - better video playback quality

I have just installed the new Cat 10.8
There is a new HDMI driver included that should fix the "silent stream bug".

I played some video files and sometimes after stopping there remains a few "soundsamples" repeating all the time...
The 24 Bit issue is still present.
And I don't know why I can't play the 5.1 24 Bit Track that I had played already... Sometimes it makes me crazy

Besides of that the Cat seems to have overscan issues. Forgets the settings sometimes

 

[robyf]

Re: AW: MediaPortal Audio renderer - better video playback quality

Besides of that the Cat seems to have overscan issues. Forgets the settings sometimes

OT: be aware that you have to reboot after overscan setting to make it remember it, at least with 10.7.

 

[tourettes]

Re: AW: MediaPortal Audio renderer - better video playback quality

I have just installed the new Cat 10.8
There is a new HDMI driver included that should fix the "silent stream bug".

Good that at least one bug was fixed.

I played some video files and sometimes after stopping there remains a few "soundsamples" repeating all the time...

Is this something that is happening only with 10.8? At least I cannot remember you reporting such earlier. If it is I'll just list it as known issue with ATI.

Logs at leas didn't show any oddities.

The 24 Bit issue is still present.And I don't know why I can't play the 5.1 24 Bit Track that I had played already... Sometimes it makes me crazy

I doubt that 24 bit issue is going to be fixed by ATI/AMD, so I would wait for the resampling code to be in place and in meanwhile use ffdshow to convert to 16 or 32 bit (if 32 bit works ).