Video Processing Question

[RLReady]

In the "DVD Movie, Concerts, and Music Discussion and Reviews" forum, there is always great discussions about what kind of a picture a particular setup gives. There has been a few threads over there recently that have brought up video processing, and whether a DVD player has 4-bit or 8-bit processing. Sorry to sound dumb (not too hard for me:D), but I am not sure specifically what they are all referring to. I just got the Oppo OPDV-971H player. The specs on the website state: "Video D/A Converter 6 x 108Mhz/12-bit". Is this what we are talking about? If I use the DVI connection to my display, does this discussion even have any impact anymore?

Thanks in advance for the education!

[Bob Pariseau]

There's lots of mysterious stuff that goes on inside a DVD player. For a basic discussion of some of it, just to give you the flavor, you might find this primer thread interesting:

http://www.avsforum.com/avs-vb/showthread.php?threadid=477740

To simplify quite a bit, you should think of the data coming off the DVD as three channels of 8 bit data for video and additional 8 bit data for audio, The internal processing inside the player can massage that data in various ways. This is usually done using digital math circuits and the resulting signal can be sent out in digital form or it can be converted to analog form. The Digital to Analog converters or DACs do that last bit. If you use the digital output(s) of the player then you bypass its DACs and their quality or lack of it is irrelevant.

Any digital math circuit is subject to what are most easily thought of as rounding errors. There are two common ways of dealing with this. First you do the math by pretending the data is coming in at a faster rate than it really is. For example you pretend there are twice as many pixels in the image than there really are. When you are finished, as the last step, you extract the proper pixel resolution from the results so far. The other scheme is to do the math using more bits to carry the result -- and round back down to the proper number of bits at the end. For example 8 bit math can only deal with 256 values. 9 bit math can deal with twice as many. 10 bit math twice as many again, and so on.

Modern video processing circuits use 10 bit math which is a whole lot better than 8 bit math. Some circuits use 12 bit math which is better still but not quite so dramatically better. Some go beyond that.

There are different ways to do the same thing in a given circuit as well -- different algorithms. Some are known to produce better results than others, but they take longer to accomplish. The digital audio and video circuits have only so much time to get their job done before the next frame of the movie comes along. So to do a more complex algorithm you also need FASTER circuits that can get the job done in time.

So the digital circuits are measured by their speed, their sampling rate, their bit depth, and the quality of the algorithms they execute, including whether there are any "bugs" in those algorithms that cause the unit to generate visible or audible mistakes.
--Bob

[RLReady]

Thanks Bob. Good reading.