Baslinux mailing list

[Stephen Clement <s.clement AT sympatico.ca>]

sindi keesan wrote:

> On Sun, 8 May 2005, David Moberg wrote:
>
> > ----- Original Message -----
> > From: "sindi keesan" <keesan AT sdf.lonestar.org>
> >
> > > > What does "jobs" say?
> > >
> > >
> > > Nothing.  which jobs  - no jobs in .......
> > > What is jobs?
>
> When I typed 'jobs' nothing seemed to happen.
>
> > jobs, like fg, is built into the shell. The ash manpage
> > should be able to tell you more about it than I can.
>
>
> Do some programs work okay in the background (perhaps wget?) but 
> mplayer and play do not?

Most programs should work okay in the background, although a few don't I 
think.

> > > If I remove the & and everything following it, your method works:
> > >
> > > mkfifo audiodump.wav
> > > mplayer -ao pcm http:  (or pnm: etc.)  WITHOUT & or fg
> > > switch vt's
> > > play audiodump.wav    - the file is 4096 not 0 now and stays that size.
> > >
> > >
> > > I probably misunderstood very clear directions for use and left in 
> > > the &.
> > > If I type Ctrl-Z mplayer stops doing anything, fg brings it back.
> >
> >
> > Normal.
> >
> > > Play audiodump.wav & does not play either, works without the &.
> > > If I start running out of terminals I will use 'screen' to make ten
> > > terminals out of one.
> >
> >
> > Keep reading for a script which will work with just one terminal,
> > and greatly simplify what you need to type.
>
> With bated breath!
>
> > > I am getting lovely sound quality from wdav now.  I have only a 4K
> > > .wav file but no skips.  It sounds far better than the radio.
> > > Mplayer said something about a cache - maybe that is acting to
> > > buffer the file before it gets into the .wav file.  Player is
> > > telling me
> > > A: 6:52.3 (time elapsed) 3% 15%
> > > The 3% went down to 2.9% and the 15 alternates with 16 and 17.
> >
> >
> > You can use the -cache option to change the size of the buffer.
> > e.g. mplayer -ao pcm -cache 8192
>
>
> The default cache seems to be 320K and it works perfectly.  I think I 
> can start playing before the cache actually fills up, at about 7%.

320K should be just fine. Up it to 512K if you have issues.

> > > On a previous occasion the middle figure was 92% and rising - is
> > > this something to do with buffering or lost clusters?  You did not
> > > include docs for mplayer.
> >
> >
> > I included the manpage in the package. Both the regular one (if you
> > have man on BL2) and a version that works on regular BL3.
>
>
> Thanks, I somehow missed it.  I have man2html in BL2 and a script 
> called mann that looks for the man page and converts and displays with 
> lynx.  I will look up what the numbers mean.
>
> > > ps now tells me there are five copies of mplayer running.
> > >
> > > I killed them all and started over, and after I started playing I
> > > again have five copies running.  Sizes about 8K, three are State T
> > > and two state S.  I have one copy of sox and none of mpg123 or
> > > play.  The same process numbers keep coming back except for the
> > > last two copies of mplayer.
> > >
> > > The second to last is taking up 3% of cpu and the others 0%, but
> > > they are all taking up about 1-1.5% of memory.
> > >
> > > If I kill the first three it keeps playing.  If it kill the fourth
> > > it stops.  The fourth is the one at the top of the list in 'top'.
> > > What are the others doing?
> >
> >
> > I think this is a scheme to improve performance. Certain CPU features
> > (for example, SMP, HyperThreading, MMX) can greatly improve performance
> > if a program is split into several threads or processes. (Mozilla does
> > the same thing.) Likely, one thread handles networking, another audio
> > output, another handles the status output. The fourth is probably the
> > master process (or "parent"), so killing it also kills the child
> > processes.
>
>
> But without the first three child processes it still works and I don't 
> notice any degradation in quality.

Well, then the main MPlayer binary probably just takes over the jobs 
that those child processes used to perform.

> > > Sox is using 1.3% of cpu and 0.4 % of memory.  What else besides
> > > compiling and gzipping uses more cpu?  Image manipulation?
> >
> >
> > Try playing a DVD on a framebuffer with anything slower than a P4
> > and watch the CPU usage climb to 100%. Or try mp3 decoding on a 386
> > or slow 486.
>
>
> A DVD involves image manipulation.  I have played mp3s in DOS on a 486.

DVD isn't really image manipulation, it is more of displaying lots of 
images really, really fast.

> > > For some reason this 200MHz computer (which I think came with this
> > > cpu in the board, set to its maximum speed of 200MHz), is
> > > identified by mplayer as AMD K6 324.0 MHz (Family: 5, Stepping; 2).
> > >   dmesg tells me the computer has a 200461 KHz processor.  I think
> > > the cpu was also labelled that way, or we looked up its number to
> > > get the speed.
> >
> >
> > The two CPU models may share the Family and Stepping codes.
>
>
> So is mplayer wrong here?

Yes.

> > > A friend says the 300 and 350 MHz AMD cpus, even in a super socket
> > > 7, are i686 and ought to work with your first mplayer and that any
> > > board which can go over 233 MHz is 686.
> >
> >
> > I think that the Pentium MMX 266 and lower are merely i586. I know
> > that all other Intels 233 or higher are 686s.
>
>
> But what is an AMD K6 350MHz?  It did not work with the first mplayer 
> you compiled.  In super socket 7.  Did your version actually require 
> Intel? That may determine where we put the DVD drive.  I am not at 
> that computer right now but how would I check if it is equivalent to 
> i686 intel?

The first version required i686, I think. That would've been the problem.

> > > Perhaps you live in Australia?
> >
> >
> > No, NW Oregon. Quite a ways from Australia.
>
> You are probably not getting our warm dry sunny weather.

Well it's nice and sunny here in Ontario (Canada).

> > I've figured out that MPlayer only needs the fg junk if it is sent
> > to the background. Sox doesn't have this problem. So, here is the
> > updated script:
> >
> > #!/bin/sh
> > mkfifo /tmp/audiodump.wav
> > play /tmp/audiodump.wav &
> > exec real_mplayer -ao pcm -aofile /tmp/audiodump.wav "$@"
> >
> > # End of script
>
>
> I have already renamed the second mplayer to mplayeralsa  (and the 
> first version to mplayerp2).  I will try this when next in linux (now 
> in DOS).
> The "$@" means I can put more than one argument after mplayer?

I'd guess so.

> Does the -aofile /tmp/audiodump.wav tell it where to make the .wav 
> file so I don't need to be in /tmp?

Yes

> > See the real_mplayer? I think that for convenience you may want to
> > rename /usr/local/bin/mplayer to /usr/local/bin/real_mplayer, then
> > put the above script into /usr/local/bin/mplayer. chmod +x. You
> > should then be able to merely run:
> >
> > mplayer http://www.foobar.com/radio/music.mp3
>
>
> And I could paste the URL in with gpm (or screen once I figure out 
> how). I had Realplayer8 in linux working with links on one computer, 
> but I can't think of a way to coordinate mplayer with links or lynx 
> since I first need to extract the URL from files of various formats 
> such as .pls or .m3u or whatever is used by realaudio and 
> windows-media.  I bet Steven or you could come up with some script 
> that looks for things starting in http:// or pnm:// or rstp:// (?) and 
> cuts off anything preceding, except that some of these little files 
> have TWO URLs in them with words in between.

(I'm guessing you meant me when you said Steven)

#!/bin/sh
wget -q $1 -O /tmp/urlfile && sed -e 's/http:\/\//WORKShttp:\/\//g' -e 
's/rstp:\/\//WORKSrstp:\/\//g' -e 's/pnm:\/\//^WORKSpnm:\/\//g' 
/tmp/urlfile |grep WORKS |sed -e 's/WORKS//g'

With this it will automatically fetch the file and search out the URL's. 
Getting it to automatically play is beyond the use of this script, however.

> > and get sound. You can also add options directly to the command
> > line, since I used the "$@" tag. All you need is a writable /tmp
> > directory to put the audiodump file in.
> >
> > > My question was, why someone would broadcast in MP3 and also OGG -
> > > does OGG have some advantage, better compression or something?
> >
> >
> > I answered your second question and completely neglected the first
> > one. Oggs sound better to my ears, all else being equal. Also,
> > don't forget the patent issues with MP3. Not everyone wants to
> > deal with those.
>
>
> I will try to hear the difference once I find a 20K OGG broadcast.
> All the formats I have tried sound infinitely better than listening to 
> radio broadcasts from 60-75 miles away, even at 20K, though this is 
> supposed to be less than FM quality I think (is that 64K?).

Yes, FM quality is supposedly 64K. Also, I will say too that I find OGG 
to be superior.

> > > So mplayer can't handle this realplayer format for some reason. What
> > > is afm?
> >
> >
> > Audio File (or Format) Manager?
> >
> > > I presume it looks for three variants by name and then by
> > > number.  Is this something I can add later or must it be compiled in?
> >
> >
> > It may need the binary codecs from the MPlayer website. The Linux ones
> > should work on a PC with my MPlayer binary. Setup may be a bit tricky.
>
>
> It said to compile them in.  I have no idea what to do with a codec, 
> whatever that is.  No big deal, now that I have 20 MP3 and some usable 
> RA and WM stations.  I will try your script soon.
>
> > David
Stephen Clement

> > --
> > _______________________________________________
> > NEW! Lycos Dating Search. The only place to search multiple dating 
> > sites at once.
> > http://datingsearch.lycos.com
> >
> > _______________________________________________
> > BasLinux mailing list
> > BasLinux AT lists.ibiblio.org
> > http://lists.ibiblio.org/mailman/listinfo/baslinux
>
> keesan AT sdf.lonestar.org
> SDF Public Access UNIX System - http://sdf.lonestar.org
> _______________________________________________
> BasLinux mailing list
> BasLinux AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/baslinux