Baslinux mailing list

["David Moberg" <davidjmoberg AT lycos.com>]

----- Original Message -----
From: "sindi keesan" <keesan AT sdf.lonestar.org>

> Would the Nvidia server be faster than SVGA or Xvesa?

Yes.

> > vesa writes directly to the hardware. No fooling around with X
> > and networks. icewm and swm do not affect the speed of MPlayer.
> I thought they used a bit of cpu time.

A very miniscule amount. Especially when they aren't doing anything.
KDE and Gnome don't really take much CPU time when they're idle,
but if you have applets or other eye candy they will choke out
MPlayer.

> >> I am starting to lose track of what works where and how and may not
> >> test #3 on the PIII since #1 works there.
> >
> > #3 works everywhere. It is the version to use. Sometime I will fix
> > the bzip2 bug and then I can kill #1 and #2 for good.
> The bzip2 bug is causing the executable to get corrupted?

Well, I like to blame the current corruption on bzip2, because it's
what packed the file in question. I will use a different compression
method, which should work better. It will also remove the need for
bzip2 to install MPlayer.

> >> I think you also said SVGA would be more efficient than Xvesa even
> >> with more colors.
> >
> > I don't recall that. A giant X configuration with real XVideo drivers
> > and an appropriate XFree86 configuration file will much better than
> > the SVGA X server. Usually plain old vesa (in CLI) works well enough,
> > however, and is much smaller.
> 
> Is there any other reason I would want the Nvidia server instead of 
> Xvesa, which works fine with this card?  I don't spend much time 
> with X - just to use Opera, or xv, or xpdf.

No.

> >> What is -hardframedrop?
> >
> > It is an option to MPlayer which tells it to drop as many frames as
> > it can. Whereas -framedrop is designed to give you jerky but
> > watchable playback, -hardframedrop will break decoding and just
> > draw whatever it can. -hardframedrop video can sometimes look like
> > a bunch of blocks moving around the display.
> 
> Do you know how framedrop decides what to drop? Is there a 
> default quality (say 50%, or 75%)?  Does it drop more frames when 
> there is more action?

It's fairly arbitrary. When it falls behind the audio, it just drops
frames until it catches up with the audio. hardframedrop doesn't
even try to make sure that whole frames are drawn; it will actually
"drop" pixels. When there is more action, MPlayer is more likely to
fall behind, so you will see more dropped frames in high-action
scenes.

> >>> Not likely. Sound takes very little CPU time, unless you're using
> >>> SDL (which is not the default)
> >>
> >>
> >> Is this why someone requested SDL?
> >
> > I don't think so. Stephen with a ph (who requested it) didn't state
> > a reason, IIRC. I've read that it's useful for when your OSS driver
> > is buggy.
> 
> Which is not the case for sb?

sb probably has thousands if not millions of eyes looking at it for
bugs. It is the most popular sound card driver, since there are
many cards which are compatible with it.

> I also got es1371 and esssolo1 
> working (on three PCI cards, for cases where I cannot figure out 
> the settings or where all the IRQs are taken).  Trident appears to 
> play, but no sound is heard, same for cs4232.

Not all sound cards have the volume turned up by default. Try
aumix?

>   Good chance to learn isapnptools.

I never did learn isapnptools. I just gave up and compiled a new kernel.
I found it much harder to learn isapnptools than compile kernels. I
may be interested in descriptions of how to use isapnptools.

> Which cards have buggy drivers?

Not sure. The ALSA emulation of OSS is perfect. I've also managed
to test the following OSS drivers which worked great with MPlayer:

cs4232 - back when I had a laptop with this chip. Great sound.
sb - My current Vibra16
i810_audio (is that the name of it?) - my current laptop
opl3sa2 - my old desktop sound card

> >
> >> Does 2.2.16 support it?
> >
> > I thought that you tried it with 2.2.16 and it failed. I'm not sure
> > whether it works or not.
> What were we talking about here?

SDL audio with 2.2.16.

> I don't know how to try SDL sound.

Use -ao sdl

> >>> PCs are fickle.
> >>
> >> This one seems a lot happier with the DVD drive as master not slave.
> >> I have noticed that hard drives transfer slower as slave.
> 
> Dos 'Syschk' gives us speed ratings. The same drive gets a higher 
> rating as master.  We put in a matched pair and measured once.

hdparm does IDE drive benchmarks in Linux. It is also the de facto way
to turn DMA on/off to improve speed/reliability. Busybox has a version,
but Steven with a v didn't include it in BL3 (hey, it rhymes)

> My DVD drive works badly with burned CDs - it gave me some reset 
> message but eventually did find the CD.

cdrecord or cdrdao (one of the two) told me that mine cannot read
CD-Rs. I can stick in the disk and hear the drive crank away. I
don't get any error messages, but I just hear the drive try to
access it. Eventually it gives up and mount acts as though nothing
is in the drive.

> > CPU type doesn't affect the OSS sound bug.
> But #2 would not even convert files on the PIII (-ao pcm).

OK, CPU type doesn't affect the OSS sound bug, barring a buggy CPU.

> >> #2 plays Realaudio - does #3?  #1 was said not to, I think.
> >
> > All 3 have the exact same codecs. #1 cannot play certain RA files
> > over the RTSP protocol. HTTP and FTP are fine.
> 
> 
> Do you mean that #1 and #3 both cannot play certain RA files 
> fetched via RTSP?

#1 only. #3 has proper RTSP support, as does #2. It's not the filetype,
it's the RTSP server type which MPlayer #1 doesn't like. (because I
didn't include the Live.COM library which handles most RTSP servers)

> Will #3 be a little less efficient on the PIII?

Not by much. I cannot tell any difference in speed between the
three versions on my PII.

> If so, I should probably keep #1 and use #3 elsewhere.  Does #3 
> differ only in that it supports all models of pentium-generation 
> computer?

I believe so.

> I will try it tomorrow.  Along with a list of other things if I can 
> remember them all.  There was a way to smooth the video, and #3 on 
> the PIII, and what else?

That's it, I guess.

> >> Does #3 do something better than #1 on a PIII?
> >
> > No. They are roughly the same except for CPU optimization. However,
> > #1 has been completely replaced by #3.
> 
> Does #1 work a bit faster on a PII or later?  If so, I will keep it 
> for playing DVDs.  How would I measure exact speeds, by looking at 
> top while playing the same DVD with x11 and -framedrop?

Do "-ao null" and set "-fps 30", 30 being some ridiculous frame rate
which your PC cannot handle. Do not use any frame dropping. This is
best done with short video clips if you're in a hurry.  Time how
long it takes to play. This will tell you exactly how long it takes
to decode the clip and display it.

David

-- 
_______________________________________________
NEW! Lycos Dating Search. The only place to search multiple dating sites at 
once.
http://datingsearch.lycos.com