HD Encoding Face-Off: WMV-HD vs. DivX-HD : Introduction
Source: Tom's Guide US | Keywords: hd, encoding, face, off
- 1. Introduction
- 2. The Test
- 3. WMV-HD
1. Introduction
I recently reviewed SnapStream's Beyond TV 4.3, which among its large feature set includes the ability to save (or re-encode) recordings in a wide variety of formats. The review contained just a few before-and-after screenshot examples, but I actually ran a number of other High Definition encoding tests. I thought the results would be useful to those of you trying to decide which format is best for HD recording.
Why is it important to worry about compression options for the HD video file format? The answer boils down to how much money you want to spend on hard drive space and how much content you want to have available without having to resort to archiving to DVD. OTA HD tuner cards generally output MPEG 2 in Transport Stream format that requires an average of 10 gigabytes of hard drive space per hour of recorded programming. At 10 GB per hour, a 400 GB hard drive is going to fill up with HD recordings in about 40 hours. Considering that the average TV viewer watches 17 hours of TV per week, you would fill your hard drive with HD content in a little over 2 weeks!
The easiest way to deal with the space restrictions is to allow your PVR software to manage the recordings by deleting the old content as new shows are recorded. Then again, if you're like me and tend to collect shows that you enjoy watching, any amount of space is eventually going to become cramped at 10 GB/hour. Compression, or file encoding doesn't cost anything but time, and allows you to maximize your hard drive space.
Another reason for compression is to increase the odds for being able to successfully stream video wirelessly around your home. Even gear based on the upcoming 802.11n standard will be hard-pressed to support a 20 Mbps HD stream at all locations around a moderate-sized home. But compression can lower the bit rate to 8 - 10 Mbps and in some cases closer to 5 Mbps, which increases the chances of a successful streaming experience.
Performance Considerations
The system used for my encoding experiments is shown in Table 1. All encoding was done with SnapStream's Beyond TV 4.3. The system isn't the fastest on the block, but it's not an econo-box either. Two items of note are the 1 GB of memory and dual-drive RAID 0 array. While you could get by with 512 MB of memory, there's nothing like more RAM to ease performance bottlenecks. Also, considering the massive amounts of data involved in capturing high definition video, it's important to make sure that your hard drive is fast enough to handle large file sizes. If you decide to use several TV tuners, you will have an even greater need for hard drive performance.
| Processor | AMD Athlon 64 3800+ |
| Motherboard | MSI Platinum K8N Neo 4 |
| Memory | 2X 512 Corsair PC3200 (1024 MB) |
| HDD | 2X 250 GB SATA RAID 0 Array |
| Graphics Card | Geforce 7900 GTX |
| HD Tuner | Dvico Fusion HDTV5 Lite PCI TV Tuner |
| Remote | Snapstream Firefly RF remote |
Table 1: The system used for encoding testing.
To give my hard drives a boost, I decided to use a two drive RAID 0 array. Raid 0 offers increased data transfer speeds that help to ensure a smooth transition from multiple tuners to the hard drives. However, before you consider RAID 0, keep in mind that even though it's fast and efficient, it's also a bit risky. It does not allow for data redundancy; meaning that if one drive fails in a RAID 0 array, the data is lost on all of the drives within the array. Click here for a more detailed explanation of RAID 0 .
Another way to boost drive performance is to adjust the cluster size of your hard drive. The default NTFS cluster size is 4 kB on drives larger than 2 GB. However, the clusters can be manually set as high as 64 kB. Larger clusters will boost performance by allowing larger chunks of information to be sent from the hard drive to the RAM.
A larger cluster size also helps keep the drive from becoming fragmented. Using 64 kB clusters, a 1 MB file would be split into about 16 sections. The same file would be split into 250 sections using 4 kB clusters which leaves 234 more opportunities for the data to become fragmented.
The reason why large cluster sizes are not usually recommended for normal desktop computer use is because it can be wasteful. The remainder of the last cluster in a split-up file takes up the same space as a whole cluster. For example: if a 66 kB file were saved on hard drive using a 64 bB cluster system, it would fill one entire 64 kB cluster and dump the remaining 2 kB in another cluster. The pesky 2 kB remainder wastes the remaining 62 kB in the cluster. After saving a lot of small files, the truncated ends will begin to add up quickly. On the other hand, the huge high definition files that are managed in an HTPC reduce the waste problem because they are so large that only waste one cluster every 10 GB or so.
- Next page The Test
Related albums
Google Ads
Comments