--- a/gigapixel.htm
+++ b/gigapixel.htm
@@ -1,14 +1,14 @@
 <h1>1.2 Gigapixel Image Test</h1>
 <p class="author">By Andrew Mihal. Last update: 17 October 2004</p>
-<p>Enblend 2.0 features a memory/disk balancing system that lets you control how 
-much memory Enblend can use before it swaps to disk. If you are doing a small 
-blending project, all of the data should fit in memory and the calculations 
-should be fast. If you are doing a big project, Enblend will put frequently-used 
-data in memory and swap the rest out to temporary files on disk. This is a 
-little slower, but it enables you to blend very large projects. You are no 
-longer limited by the amount of memory in your computer, or by address space 
-limits on a 32-bit machine. The only limit should be the amount of disk space 
-you have for temporary files, the length of time you are willing to let Enblend 
+<p>Enblend 2.0 features a memory/disk balancing system that lets you control how
+much memory Enblend can use before it swaps to disk. If you are doing a small
+blending project, all of the data should fit in memory and the calculations
+should be fast. If you are doing a big project, Enblend will put frequently-used
+data in memory and swap the rest out to temporary files on disk. This is a
+little slower, but it enables you to blend very large projects. You are no
+longer limited by the amount of memory in your computer, or by address space
+limits on a 32-bit machine. The only limit should be the amount of disk space
+you have for temporary files, the length of time you are willing to let Enblend
 work, and the 4 gigabyte limit of the TIFF format.</p>
 <p>To test this out, I ran Enblend on a 1.2 gigapixel image.</p>
 <div align="center">
@@ -52,23 +52,23 @@
 </table>
     </center>
 </div>
-<p>The working assumption is that each horizontal row of images in the panorama 
-has been previously blended. This is an easier job for Enblend because the 
-overlap regions are smaller. In this test I am simulating blending two such 
-horizontal rows together vertically. This is a good stress test because the 
-overlap region is large. Each "row" is 10,000 pixels in height. The overlap 
-region is 5,000 pixels in height. The total size of the overlap region alone is 
+<p>The working assumption is that each horizontal row of images in the panorama
+has been previously blended. This is an easier job for Enblend because the
+overlap regions are smaller. In this test I am simulating blending two such
+horizontal rows together vertically. This is a good stress test because the
+overlap region is large. Each "row" is 10,000 pixels in height. The overlap
+region is 5,000 pixels in height. The total size of the overlap region alone is
 200 megapixels!</p>
 <p>Just to make things more interesting, these images are <b>16-bits per channel</b>.</p>
-<p>I performed this test on a 2.8 GHz Pentium IV, with 2 GB of memory, running 
+<p>I performed this test on a 2.8 GHz Pentium IV, with 2 GB of memory, running
 Linux. I used the following options:</p>
 <p><code>enblend -g -m 1500 -o gigapixel_out.tif gigapixel1.tif gigapixel2.tif</code></p>
-<p>I limited Enblend's memory allocation to 1.5 GB to prevent the Linux kernel 
-from swapping. Enblend used 10 blending levels. From start to finish, the job 
-took <b>1 hour 8 minutes</b>. About 27 GB of disk space was used for temporary 
+<p>I limited Enblend's memory allocation to 1.5 GB to prevent the Linux kernel
+from swapping. Enblend used 10 blending levels. From start to finish, the job
+took <b>1 hour 8 minutes</b>. About 27 GB of disk space was used for temporary
 files.</p>
-<p>Note that without compression, each of these images would be 9.6 GB, which 
-exceeds the limit of the TIFF format. However, these results show that Enblend 
+<p>Note that without compression, each of these images would be 9.6 GB, which
+exceeds the limit of the TIFF format. However, these results show that Enblend
 is perfectly capable of blending projects right up to that limit.</p>
-      
 
+