sane-project-website/old-archive/1997-07/0085.html

81 wiersze
4.0 KiB
HTML

<!-- received="Fri Jul 18 09:08:04 1997 MST" -->
<!-- sent="Fri, 18 Jul 1997 08:47:46 -0700" -->
<!-- name="David Mosberger-Tang" -->
<!-- email="davidm@azstarnet.com" -->
<!-- subject="Re: Mustek 3-pass scanner test code" -->
<!-- id="199707181547.IAA26600@panda.mosberger" -->
<!-- inreplyto="davidm@AZStarNet.COM" -->
<title>sane-devel: Re: Mustek 3-pass scanner test code</title>
<h1>Re: Mustek 3-pass scanner test code</h1>
<b>David Mosberger-Tang</b> (<a href="mailto:davidm@azstarnet.com"><i>davidm@azstarnet.com</i></a>)<br>
<i>Fri, 18 Jul 1997 08:47:46 -0700</i>
<p>
<ul>
<li> <b>Messages sorted by:</b> <a href="date.html#85">[ date ]</a><a href="index.html#85">[ thread ]</a><a href="subject.html#85">[ subject ]</a><a href="author.html#85">[ author ]</a>
<!-- next="start" -->
<li> <b>Next message:</b> <a href="0086.html">Jeremy Hansen: "Re: Mustek 3-pass scanner test code"</a>
<li> <b>Previous message:</b> <a href="0084.html">David Mosberger-Tang: "mustek 3-pass fix"</a>
<!-- nextthread="start" -->
<!-- reply="end" -->
</ul>
<!-- body="start" -->
<i>&gt;&gt;&gt;&gt;&gt; On Fri, 18 Jul 1997 16:56:53 -0600, "Mathias Weigt (Stud. 94)" &lt;<a href="mailto:p5iba@medchem2.pharmazie.uni-halle.de">p5iba@medchem2.pharmazie.uni-halle.de</a>&gt; said:</i><br>
<p>
David&gt; Could somebody with Linux and a Mustek 3-pass scanner try the<br>
David&gt; code below and tell me how/if it works?<br>
<p>
Mathias&gt; I tested it very shortly and I believe that scanning has<br>
Mathias&gt; become faster :-) -- but it seems to be memory dependend<br>
Mathias&gt; :-( I have only 40MB)... scanimage --mode color<br>
Mathias&gt; --resolution 300 dpi &gt;/dev/null needs 3:45 min.<br>
<p>
Wow, it used to take 15 minutes, right? That would be a nice<br>
improvement. How does it compare to the scan times with other<br>
programs?<br>
<p>
Mathias&gt; The system crashes sometimes (so on a 600 dpi scan in the<br>
Mathias&gt; middle of the third pass) but this may be kernel instabilities ?<br>
<p>
I certainly haven't seen any crashes due to SANE. Do you mean the<br>
_entire_ system crashes or just some tasks get killed? The latter<br>
happens under Linux when running out of swap space.<br>
<p>
Mathias&gt; The first pass is done really fast, the second and third pass<br>
Mathias&gt; becomes slower. It seems to depend on how fast the system<br>
Mathias&gt; is able to swap, because during one pass the scanner<br>
Mathias&gt; sometimes gets fast for some seconds and becomes slow<br>
Mathias&gt; again. I assume that a 128 MB System won`t have<br>
Mathias&gt; performance problems unless it is driving StarOffice :-).<br>
<p>
This makes sense with the current buffering scheme: the first pass<br>
simply writes out data. The second and third passes in contrast have<br>
to read the existing data, add info for one channel, then write it<br>
back out. There was some discussion to implement a better buffering<br>
scheme that would fix such problems for scanimage, but AFAIK, nobody<br>
has done it so far. If you scan with xscanimage instead of scanimage,<br>
the swap problem _ought_ to go away. I'd be interested to hear how<br>
well it works under GIMP as well.<br>
<p>
Mathias&gt; Sorry but I can`t provide debug output or look deeper into<br>
Mathias&gt; it because I`m going on holiday tomorrow... Nice to see<br>
Mathias&gt; 3-pass patches :-)...<br>
<p>
Well, enjoy your vacation then! Thanks for testing things!<br>
<p>
--david<br>
<p>
<pre>
--
Source code, list archive, and docs: <a href="http://www.azstarnet.com/~axplinux/sane/">http://www.azstarnet.com/~axplinux/sane/</a>
To unsubscribe: mail -s unsubscribe <a href="mailto:sane-devel-request@listserv.azstarnet.com">sane-devel-request@listserv.azstarnet.com</a>
</pre>
<!-- body="end" -->
<p>
<ul>
<!-- next="start" -->
<li> <b>Next message:</b> <a href="0086.html">Jeremy Hansen: "Re: Mustek 3-pass scanner test code"</a>
<li> <b>Previous message:</b> <a href="0084.html">David Mosberger-Tang: "mustek 3-pass fix"</a>
<!-- nextthread="start" -->
<!-- reply="end" -->
</ul>