|ASUS MY-CINEMA PHC3-150 TV-Tuner Card|
|Reviews - Featured Reviews: Accessories|
|Written by Dan Ferguson|
|Tuesday, 23 February 2010|
Page 5 of 6
Testing & Results
Two aspects of TV card performance were of interest; resource consumption and recording quality. To monitor resource performance the ASUS My Cinema PHC3-150 TV Combo Card was run simultaneously on the same machine with the ATI TV Wonder HD 650 Dual Tuner. Percent CPU usage was recorded while streaming over-the-air channels at resolutions of 4801, 720p, and 1080i. Next, CPU usage was recorded while recording over-the-air streams at the same resolutions. In both tests the same signal was simultaneously sent to both the PHC3-150 and 650 HD. The recorded videos were then analyzed for recording quality. Two objective measurements were used to assess the quality of the recordings. First, the file size was divided by the total number of video frames to determine the kilobytes per frame. For identical file types, more kilobytes means a better image. Second, the recording bit-rate was divided by the combination of frame rate and resolution to determine the bits per pixel. This helps compare recording quality between different resolutions. For both metrics, higher values mean more information was captured and saved from the original stream.
As a subjective test, the videos were reviewed to see if there were any noticeable recording defects. Also, since the PHC3-150 reports automatic noise filtering, an old VHS video was recorded to see if there was an obvious improvement in quality.
To start, there were major differences in recording settings between the ATI and ASUS applications. ATI has a wide range of selectable recording qualities including the ability to define a custom video. The maximum recording resolution for ATI was the same as the broadcast resolution. For TotalMedia, four options are available; 'High Quality', 'Standard Play', 'Long Play', and 'Extended Play'. Repeated tests indicate that the exact recording format for each option are automatically chosen by the software based on the broadcast resolution. For example, a high quality recording of both 1080i and 720p sources resulted in a 1920x1080 video, and a 480i broadcast resulted in 720x480 recording. The extended play option resulted in recordings with resolutions matching the broadcast resolution. There did not seem to be a way to capture really low quality video.
Based on those results, the 650 HD was set to capture the maximum resolution via Catalyst Media Center, and the PHC3-150 was set to capture the lowest quality via TotalMedia. This resulted in identical recording formats. For reference, the file details are included at the end of this section.
The CPU usage obviously doesn't stay at one value during the entire test. Reporting only one value thus oversimplifies the actual performance. For example, if the CPU were at 100% for half the test, and at 0% for the second half, the average CPU reported would be 50%. The average of 50% hardly describes what really happened. One way to show CPU performance over a length of time is to use a chart called a box plot. A box plot shows how frequently a range of values occur, or in this case how long the CPU stayed in a specific range.
The box plot is a box with a line in the middle and "whiskers" or lines that extend from the top and bottom of the box. The range of the box covers the range in the middle where the CPU stayed for half the test. The line in the middle of the box represents the middle most value experienced in the test. The ends of the whiskers show the maximum and minimum values encountered during the test. Finally, the length of the top whisker covers the range where the CPU stayed for 25% of the time. Likewise for the bottom whisker.
Interpreting the box plot above, the 650HD and PHC3-150 had similar performance. As the resolution increased the system demands also increased. Since no recording was taking place, this may be simply be the resources required by the application to display video. As a reference, Windows Media Player used an average 33% CPU to play back the 1080i video that was recorded. For both cards the averages, minimums and maximums are all comparable. For all resolutions the PHC3-150 took slightly more average CPU and had higher spikes.
The CPU required to record and stream simultaneously was only a few percent higher than the requirements to stream only. This indicates that the on-card mpeg encoder really does the majority of the work. There are some notable differences. The CPU spikes were higher and the 720p recordings were significantly higher. This is likely an indicator of the encoding performance since the progressive scan format of 720p uses twice the frame rate as 480i and 1080i.
For all resolutions, the kilobytes per frame for both cards are within a half kilobyte. This indicates the quality between tuner cards is well matched. As a reminder, the ATI is recording at the best setting while the ASUS is recording at it's lowest setting. If quality is important than the ASUS has more to offer in the higher resolutions.
The bits per pixel reveal slightly more information about the recordings. The bits per pixel decrease with increasing recording resolution. This means that the higher resolution frames are of lower relative quality. But the advantage of having more pixels outweighs the slight degradation. This metric more clearly shows the higher recording bitrate of the ASUS card. This means that on average the PHC3-150 captured more detail of the original stream than the 650 HD.
For the subjective analysis, videos from both cards shared common defects. At random times, especially during quickly moving scenes, horizontal bars of distortion could be seen. At some instants the image almost seemed fuzzy. The ASUS card at times would pause for a brief instant then snap to the present image. This is probably due to dropped frames during recording. Dropped frames are often a result of multi-tasking on the machine, and may have been affected by the simultaneous testing. It may also have been due to a flicker in signal strength as the recording was over-the-air. To determine if the dropped frames are due to the card or due to the signal a separate recording was performed using a satellite signal while running only the PHC3-150. The video recorded from this test did not suffer from dropped frames or horizontal bars. The only defects seen from this recording are the typical random blocks that is inherent in MPEG compression. With the over-the-air signal the ATI did not suffer from dropped frames, but there was a noticeable difference in the image quality. In many clips there were shaded, thatched lines that were not seen in the ASUS videos. While these may have been due to using different antennas, it was more likely due to the encoding process. These differences were not detected during the hours of video watched on both cards.
Recorded Video Details: