ATI TV Wonder Elite
 |
|
In an effort to evaluate PVR hardware (and software) with more understanding, we need to explain a few things first. So let's take a look at how the TV signal gets from your wall jack to your monitor for your viewing pleasure.
Analog To Digital Converting and Sample Rate:
Picture a sine wave, which represents the analog video TV tuner signal. Once the TV tuner on the PVR card brings the signal in, the video processor's analog-to-digital converter (ADC) converts the signal into digital information. Now here is where PVRs set themselves apart. Points on the sine wave need to be translated into numbers via the ADCs. The higher the bit the ADC, the more precise the digital information can be relative to the analog data. (Note that the audio and video processor is not the same as the Graphics Processor. The best example would be the Theater RXXX chips from ATI or Conexant chips on WinTV cards.)
Another way to think of this is a sine wave laid down on graph paper. Vertical lines would represent increments of time, and the horizontal lines would represent amplitude in volts (bits). The sampling frequency is basically fixed because there are only 60 interlaced fields per second on the NTSC standard. This is the maximum sampling frequency rate for video, because anything beyond this would be pointless, as video is broadcast at that spec. This makes the video sampling rate (vertical lines) for all NTSC standard A/V processors the same. However, specs for audio sampling may vary, because it is a much less uniform standard. Just think of it similarly to the different sampling rates for MP3 tracks for example.
If you aren't familiar with sine waves, you can think of sampling and amplitude in the context of the game "Battleship." You want more gridlines because you can hide your battleships and sub(s) better. This is the same as a sine wave, as more gridlines gives you more accuracy possibilities.
We are going to focus on the video side of the tuning equation, as its importance seems to be discussed a lot less than audio. Let's face it, poor audio sampling is much easier to pick out. You can think of the crackles and pops you hear on a weak radio signal or low bit rate MP3 file.
With a higher bit ADC, it would be like having more horizontal lines, where you can get a better approximation of the value of points on the sine wave. The by-product of having more horizontal lines is less quantization noise, which is the rounding that occurs when the amplitude gets really low (i.e. quiet parts of a movie). This rounding occurs because the ADC cannot represent imaginary numbers that lie between lines say in a 4-bit amplitude wave.
The Signal Processing Hardware Chain:
Once the signal is converted into digital information, the video processor will further process the video data via chroma, comb filter, and luminance processing. These three components: TV tuner, ADC, and video processing methods are the core to what sorts PVR cards into the "good, bad, and the ugly" categories. The TV tuner contributes by the quality in which it brings in the signal, i.e. some have lower quality shielding. The standard for TV tuners in PVR cards is Phillips, though we have seen some good ones that use Microtune or Sony components. And of course, the higher number of bit accuracy for the ADC the better.
Keep in mind that silicon chip based tuners like those on the All-in-Wonder 8500DV are not digital in nature. When we talk about bringing in the signal to the A/V processor, this is done by the work of the circuits in the TV tuner. The only difference in a silicon chip based tuner is that a single chip replaces many of those circuits. This is what NVIDIA has chosen for its more recent Personal Cinema reference designs. The only problem with this is that while there is less "noise" (interference) by discarding some of the circuit designs, manufacturers have to ante up the power consumption for the PVR card, which generally means that it is going to run a bit hotter than those PVR cards with circuit based TV tuners.
As for the filtering methods, we refer to native brightness, contrast, and line twitter. Brightness and contrast have the ability to be manually manipulated on the software side, but the manufacturer still needs to get the default setting down. Typically, a dynamic/adaptive comb filter is preferred over some preset number like 2/4, because line twitter can vary depending on the original quality of the broadcast signal.
