Hi all,
I posted earlier this week about trying to capture using a Matrox X100 capture card with Gem 0.90.1 and Windows 2000. I've figured out that it sort of works using the analog input to the capture card.
However, when I open the gemwin (I don't even have to start rendering) PD slows down to the point where it takes as much as a minute to repond to my input. The CPU load meter jumps up to above 600, and in about 5 minutes PD crashes without a message.
I've used GEM (.87, w/ pd 0.36) with a video input on this computer before with no problems, so this is odd. It has dual P4 3.0 GHz processors, so it shouldn't have a problem there--this patch runs smoothly on a single P4 3.2G. The video card is a Radeon 9800, so /that/ shouldn't be a problem..
Any ideas what could be causing this, or what I could do to fix it?
Thanks, Ian
Hi Ian,
I've been working with pix_video on Win2k, and it worked quite well in general. However, I found that capturing from a DV source in full resolution (720x576) didn't leave a lot of CPU cycles for other purposes on my box (P4, 2.6gig) - lowering the capturing resolution might help in your case, too. If you don't have an adequate utility for your matrox capture card, you might want to try amcap.exe or graphedit to change the settings of the WDM video capture filter. It seems like the mac code for grabbing pixes is a lot more optimized: someone stated recently that handling hd video (1080i) posed no problem on recent macs ...
Another hint would be upgrading DirectX and/or the capture drivers on that machine. I couldn't manage to get GEM to render any textures with the version of directx that shipped with win2k on my laptop.
Hope this helps (although I'm merely stating the obvious), Thoralf.
___________________________________________________________ Gesendet von Yahoo! Mail - Jetzt mit 250MB Speicher kostenlos - Hier anmelden: http://mail.yahoo.de
Thanks for the response.
I've upgraded every driver under the sun and DirectX. It's good to hear you recommend that because I was considering downgrading it. Upgrading the Radeon driver helped a touch.
Unfortunately, the Matrox's new drivers only run on Windows XP; it looks like Matrox /withdrew/ support for Win 2k or something!?!
I also wasn't aware that it was possible to reduce the capture resolution before it got to Pd, I'll look into that.
For the time being I'm going to try to use a G4 laptop. What you say about the optimization is interesting. I was amazed at how easily the video software Isadora plays with full res DV, despite its rather heavy GUI. Isadora's author basically said he did Altivec optimizations, but I didn't think that that would account for the impressive performance. A Mac pecularity would help to explain it. Does anyone have info on this?
Thanks again, Ian
Thoralf Schulze wrote:
On Mar 29, 2005, at 11:06 PM, Ian Smith-Heisters wrote:
I think you will find that GEM on OSX will perform equal to or most likely better than Isadora on the same hardware. I gave the author of Isadora some tips on Quicktime and OpenGL code, but that software still does a lot of RGB processing. Plus GEM only has a hardware accelerated OpenGL graphics rather than QuickDraw.
There's nothing really peculiar about the Mac and video except that a few people (on the order of a few dozen in the world) have sat down and actually done the required work to make things efficient. Given that number, there should be a few hundred doing the same on Windows, but I'll be damned if I've ever seen their output.
What I find amazing is that you can't actually tell people about how poorly all this commercial software performs. Once you show them the alternative then they realize, but usually resort back to the stuff they (over)paid for. I also really enjoy reading press reviews that laud the latest and greatest software that has the same features Pd/GEM had years ago. I guess someone should hire a PD PR person.
chris clepper wrote:
I think the other thing is that the ones using commercial software usually aren't too interested in performance, control, or modifiability; anyone who is quickly gravitates to OSS. I spend hours and hours troubleshooting, compiling, installing, etc and I get software that will do whatever I want efficiently, smoothly, and for free. But sometimes I wish it "just worked", and that's when I could justify paying $500 for a sub par piece of software that, rather than doing what I want, will force me to want to do what it's capable of.
But programs like Final Cut Pro with its "Realtime Video" coming out these days pisses me off; capitalizing on ideas and even code that's been around for years for free. I don't mind FUD, but I do mind them claiming they invented the wheel--at the least there should be credit where credit is due.
But oops. Sorry for the rant. The only excuse I have is long rehearsals, tech week, and too much wine ;)
-Ian
a bit off topic...but this seems like a group that would know:
What is out there in terms of a cross-platform open-source MIDI sequencer along the lines of Cakewalk/SONAR or Digital Performer? Barring cross-platform, what are the recommended open-source sequencers for win32 and OS X?
Thanks in advance for suggestions.
-John
On Wed, Mar 30, 2005 at 12:23:14AM -0500, John Harrison wrote:
there are basically a lot of commercial closed/source apps that exist on Mac+Win, but not linux, and a lot of open source that exist only on linux with maybe a shabby mac port, but very few win+linux apps, besidse PD (which makes a great sequencer, using [pipe], data-structures or [xeq]) but to answer your question:
only one i have heard of is JAZZ -> http://www.jazzware.com
I am a PD newbie...I'm open to using pipe, xeq, or qlist for my sequencing. Because of my lack of experience with these, perhaps, it appears to me like I'd actually have to build the front end to the sequencer...and/or I would be without the editing features of a commercial sequencer (at least in any kind of convenient and time-efficient package.) It would make my day to be wrong about this. Are there any front end MIDI sequencers for PD already built on the engines you mention that I could take a look at?
I checked out Jazz before I posted. It doesn't look promising outside of Linux at this time, but I will look further. The current development (since 2000) is at http://jazzplusplus.sourceforge.net/
Thanks in advance for the guidance.
-John
ix@replic.net wrote:
You can also use Pd's graphical data structures to do sequencing.
Here's an example of sample manipulation done this way:
You can download the patch, but you need a fast machine to run it, like
2GHz+. I composed this using a dual G5 2.5GHz.
.hc
On Mar 30, 2005, at 9:36 AM, John Harrison wrote:
"Looking at things from a more basic level, you can come up with a more
direct solution... It may sound small in theory, but it in practice, it
can change entire economies."
- Amy Smith
Hans-Christoph Steiner wrote:
Csound is the obvious counterpart to Pd. It's available for Mac, Win and Lin. There's also openmusic from Ircam but this is Linux only (at present) and I'm told Ircam products tend to be on the buggy side - perhaps more academically useful than practical.
http://freesoftware.ircam.fr/ http://www.csounds.com/
I'm told Ircam products tend to be on the buggy side -
-Is this because of the rather 'experimental nature' of it, or just because of incompetent programming habits?
perhaps more academically useful than practical.
-This is quite a intriguing statement. Please elaborate.
AvS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................................................
` |Schreck Ensemble . . . . . . . . . . . . . . . . . . . . +
` |# -laboratory for live electro-acoustic music- # |
| http://www.schreck.nl/ |
| http://www.xs4all.nl/~schreck/ |
` *===========================================================++
` |Compositions http://www.xs4all.nl/~schreck/html/compo.html |
` |Samples http://www.xs4all.nl/~schreck/html/samp.html |
` |Patches http://www.xs4all.nl/~schreck/html/pat.html |
` |Videos http://www.xs4all.nl/~schreck/html/video.html |
` |Scores http://www.xs4all.nl/~schreck/html/scores.html |
*===========================================================++
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....................................................................
Arie van Schutterhoef wrote:
No idea. Having never used OM myself, I don't even know if what I said is true, it's just what I was told.
perhaps more academically useful than practical.
-This is quite a intriguing statement. Please elaborate.
Ircam is an academic institution trying to solve problems, not sell products. The problem here, presumably, is how to compose electronic music using a language designed specifically for that task instead of trying to adapt conventional notation to suit electronic musician's needs, like we do at present. The former is considered elegant and liberating while the latter is clunky and constraining. In theory I agree but in practice I still compose music with a staff or piano-roll matrix, not data structs and pointers.
not cross platform as far as i know, but seq24 looks decent if you need simple, stripped down sequencing:
http://www.filter24.org/seq24/
John Harrison wrote:
Thanks for all of these replies. I am anxious especially to look at Hans-Christoph Steiner's stuff, since he did the sequencing in PD.
Since I posted, I found for win32 Music Studio Producer
http://www.frieve.com/english/index.html
which is a bit buggy but has the features I was looking for. It isn't open source, but it is free.
-John
Josh Steiner wrote:
Yes!
I think a made with PD sticker/logo would be cute. I think it should be classy though, more-so than banners on pages and more-so than icons etc.. I'm thinking something that looks like an art-council seal, that would really lend legitimacy on a level that a web-banner can't. (we'll outside the web-world).
Other than spreading the work, improving PD, increasing its visiblity through the project data-base and teaching it to others what else can we do?
B.
chris clepper wrote: