Since Optimized LAS is new on the market people are obviously interested in different ways it compares to other solutions. Optimized LAS has a number of benefits in terms of providing compression, spatial indexing and statistics required for better user experiences as well as restructuring for use in cloud based storage and distribution. Other solutions may do things differently and it can be hard to make fair comparisons. In this post we’ll try to clarify how to do this from the LAS Optimizer’s perspective. First up, compression speed:
Apples-to-apples comparison of compression performance between LAS Optimizer and other software
Raw compression speed is one factor to look at when comparing the LAS Optimizer to other solutions. Most other software that compresses LAS does just that, compression and nothing else. Since the LAS Optimizer does a combination of things, to look at compression speed alone we need to exclude those other factors for a fair comparison. Here’s how you run the Optimizer at its fastest with it only performing the work of compression:
- Create a ‘*.lasx’ file for a LAS file. This sidecar auxiliary file contains statistics and indexes for the LAS file. You can get this made in one of two ways:
- Use the Optimizer itself to compress and decompress a file. The resulting LAS will have an additional *.lasx sitting next to it.
- Use the LAS dataset in ArcGIS 10.2. One of the options on the LAS Dataset Properties dialog is to Calculate Statistics. This will make a lasx for each referenced LAS file.
- Under the Optimizer’s Advanced Options make sure the CPU power is set to 100%.
- Turn off ‘Rearrange points’ and ‘Import PRJ file’ options.
- Run the Optimizer.
An alternative, if the other software has options to build indexes, sort point records, etc., is to include them in the process and the time it takes to execute them.
The LAS format was intended for data exchange, not direct use. One significant example of this is the lack of indexing. Because of this, any query on a file requires it be scanned, in full, to find the appropriate point records. As files get bigger and increasingly stored on networks, the cost of not having an index becomes prohibitive.
While compression can reduce file size, if indexing isn’t included those files need to be scanned, in full, just like regular LAS does. The files may be smaller, but even a 100MB compressed file is a lot to read over a network if you only want a small piece of it, especially if you’re doing something repeatedly or there’re multiple users accessing the data. Also, consider this is happening against a collection of files.
Optimized LAS not only reduces the size of LAS files via compression but also provides direct, indexed, access. The efficiency of indexing can also be maximized through the optional rearrangement, or clustering, of point records which places points in close spatial proximity next to one another in a file. The net result is significantly reduced I/O cost (i.e., bandwidth, time, money). Here are some steps you can take in ArcGIS to see how this works:
- Take a LAS file, the larger the better, and run it through the EzLAS Optimizer using the default options.
- Take note of the original LAS file size and the size of the optimized output.
- Make a LAS dataset from the optimized zLAS file and add it to ArcMap.
- Keep the LAS dataset in point view mode.
- Open Windows Task Manager. In Windows 7, switch to the Processes tab, and add the ‘I/O Read Bytes’ column, if you don’t already have it, via View>Select Columns. In Windows 8, switch to the Details tab and right click the header to get to the Select Columns choice.
- Keeping an eye on I/O Read Bytes for the ArcMap.exe process, zoom in on your data and pan around.
- For each redraw compare the I/O Read Bytes to the size of the optimized zLAS file. Only a portion of the file is read. The more you’re zoomed in the smaller the amount.