So, It works!!
Indeed it does!
I saw your performance log in the readme, and I guess it's quite good, for a 206 MHZ proccesor.
It's fairly impressive, but it probably could be even better. I suppose a lot more ARM optimizations could be done in some of the libraries that VLC uses for decoding audio and video -- possibly even by using routines from Intel's IPP for the SA1100 instead of existing routines (IDCT and the like).
I just have one question, do you use VLC as your main media player?
As a matter of fact, yes -- while I haven't gotten rid of the original player yet, I will eventually remove it from my ROMs' CRAMFS. So far, I've just been too plain lazy
How does it stand up against Mplayer (In terms of features, performance, possible power)?
Actually, I wouldn't know. I
never got a fairly recent release of MPlayer up and running on the Z. The binaries available that I found wouldn't run for a variety of reasons (either containing XScale instructions or needing a non-SharpROM environment), and building it myself, well -- I tried, but ran into problems, looked at the code and decided it was just too plain ugly to lay hands on. Sorry if that offends some people, but that's the way I feel.
To make a long story short, the only version of MPlayer I ever got running was an ancient 0.6-something, and its performance was more or less a joke. So I can't give you any comparison that would stand up close scrutiny.
If VLC is good enough, then mabye you could team up with other programmers (The original Videolan team?) to optimise the player and mabye even make a functional UI!
It's good enough for
me, and, as mentioned before, I am giving feedback to the VideoLAN developers to help improve VLC for the Z. As for an UI, well --
I'm probably not gonna do it, since, as I mentioned in the other thread already, I'm not a UI kinda guy
Anyways, I think you should post your project down in ELSIX (I don't know if this is yours), you know, make it known.Mabye then, people might actually get interested.
Uhm, no, that's not 'mine'. That's the (by now way outdated) original port by the VideoLAN team. As for a
project, well, there's not really one. All
I do is take the sources needed and run them through the (old) compiler, make sure everything is statically linked so I won't end up in
.so hell and that's more or less it.
Time- and effortwise I don't see myself able to do much more than that, especially not maintaining a binary release that makes everybody in the Z community happy and giving support for it as well. Anyone who wants to try
my build (thanks once again to
porovaara for providing the space) is welcome to do so, and I will certainly try to help out as much as is feasible with my schedule, but other than that, people will be on their own.
So I won't do anything like you suggested, since I could never live up to the standards I expect from myself when handling such a project, simply due to the constraints I would have to operate within -- so I prefer not doing it at all rather than disappointing myself and, even worse, others.
Best regards,
Chris.
