Optimizing Your PC Hardware for DSV Recording

By Bill Mollon, Digital Imaging Product Manager, Gatan Inc.

In our December 2005 issue of KnowHow we introduced the DigitalMicrograph plug-in called Digital Streaming Video (DSV). The digital video stream as it is delivered by the DVCapture module within DM is intended to be used by another software program. Digital Micrograph merely acts as the supplier of the digital stream. It is up to the end-user to utilize other 3rd party mastering or authoring programs to capture and create “movies” from the stream. Movie creation in itself has many controllable variables that will affect the final output. Some are software related and others are highly dependent on the hardware being used. Namely, the PC and its components should be carefully considered when movie recording and creation is part of your application needs. This article will go over the areas you should be aware of to optimize your hardware system for DSV recording.

Part I. Digital Streaming Video v.s. Analog Video

In order to understand the hardware components involved, let’s take a look at what is different between the “older” conventional video signal recording and the newer digital method. Table 1 shows you a comparison between the two types of video signal production. There are obvious advantages when using DSV compared to analog video. A major component was the cost of the equipment to record the analog signal. Usually this consisted of expensive tape machines, character generators, external signal information that was overlaid on the video signal, etc.. A lot of the same effects can now be done with the PC and a few bits of software.

Table 1

PC Specifications

Careful consideration must be paid to your choice of PC or computer if you are going to be dedicating your production work to creating quality digital imaging movies. We have found that there are certain bottlenecks that can affect your production time and also quality. They are:

• Speed of the digital acquisition
• RAM memory in the PC
• Hard drive space in the PC
• Hard drive read/write speed
• Application software for editing

In regards to the speed of the acquisition device, the camera’s frame rate is usually the specification to look at. Matching the dynamic event interval that you want to record with the frame rate capability of the CCD is very important. Using a slow frame rate CCD (< 12fps) and trying to capture a very fast interval in time can be done but you will be obviously missing data.

RAM memory in the PC is important to the application software that you will use for the production side of the movie. Some digital authoring software products demand more memory than others so please check the specifications of the software when configuring the PC memory. A certain overhead amount is always used by the operating system (Windows™ 2000/XP) so in general the rule of thumb is “512MB for the O.S.”. Our test PC had at least 1GB of RAM which proved to be sufficient but the application software would slow things down a bit when it came to memory intensive operations like rendering the output data to a certain file format. Another 1GB would have helped alleviate the slow down.

Figure 1 Hard drive read/write arm and 3 platters

The most critical component we found was the hard drive. The technology of hard drives today is rapidly changing and the price per megabyte has lowered significantly. Since storage space has become relatively cheap, larger drives in the 100’s or 1000’s of GB can be used without putting a stress on your budget. Technology advances give rise to performance increases in PC’s and in the case of the hard drive that involves the read and write access times to the hard drive platters (Fig. 1) Digital streams from a source camera can be acquired quite quickly but that speed can be quickly compromised by a hard drive that is slow to write the data down. A natural bottleneck or data “dam” will occur in which you will see

skipping of frames or sluggish response in the frame rate of the recorded movie. This can be seen very easily in your finished movie as hesitation or jumps in the movie frames instead of a natural smooth flow of data as it is displayed on the monitor. The best hard drives on the market usual carry the price based on this throughput specification of “access time”. The IDE format (ATA) of the past is giving way to the newer SATA drives which are faster and more capable of passing large amounts of data per unit time. In the computer server arena this hard drive and file access time was a big consideration when you had multiple users trying to access data simultaneously on a file server. This problem was solved with the introduction of RAID array systems. In order to understand how this can benefit you for DSV recording we need to understand the basics of a RAID array device.

RAID (Redundant Arrays of Inexpensive Disks)

A modern RAID is comprised of multiple, fast hard drives arranged in a certain configuration depending on the application need. There are different levels defined ranging from RAID 0 to RAID 10. Each level number has its own special purpose and advantages. Some of the features of a RAID design are: (1) Faster read/write access; (2) Data security and integrity; (3) Fault tolerance (protection against failure); (4) Increased capacity for storage, and (5) Performance. For the sake of this article we will look at the level that will give us the fastest “access times” for reading and writing data to the hard drives. This is referred to as RAID 0. Figure 2 is an illustration of a RAID 0 configuration. In order to create this type of RAID you will need at least 2 hard drives, 3 or 4 being preferred. RAID 0 files are broken into stripes which improves performance by splitting up files into small pieces and distributing them to multiple hard disks.

Below is an example (Fig. 2) of how data is written in a RAID 0 implementation. Each row in the chart represents a physical block on the drive and each column is the individual drive. The numbers in the table represent the data blocks (file).

Drive 1 Drive 2 Block 1 1 2 Block 2 3 4 Block 3 5 6

Figure 2 RAID 0

Thus, if the 6 blocks of data above constitute a single data file, it can be read and written to the drive much faster than if it were on a single drive. Each drive working in parallel could read only 3 physical blocks while it would take a single drive twice as long because it has to read 6 physical blocks. At least 2 drives are required for RAID 0 but usually 3-4 are used to improve the speed. The drawback of course is that if one drive fails, the data is no longer functional. All 6 data blocks are needed for the file, but only three are accessible. Table 2 shows a comparison of all the most common RAID levels and their characteristics.

Most PC’s offered today have the space and capability of multiple hard drives within the tower configuration used. One hard drive should be reserved for the operating system and programs with the additional drives being devoted to the RAID configuration for your data collection. Check to see if your PC model has built-in RAID capabilities in the motherboard. It is usually an option and knowing will help determine if you need to consider a separate RAID controller PCI-e card to be added. We have tested both scenarios and can make suggestions on what worked well.

Table 2 RAID Levels

Hard Drives and Performance

Let’s examine the details about the drive performance and its role in optimizing your PC. Some of the points we have introduced can be summarized as follows

All of these factors should be considered when choosing the hard drives that will be part of the RAID array in the PC. Try to look for fast spindle speeds (10k), short seek times (5-8ms) and high data transfer rates (150-300MB/sec).