• Welcome to Milwaukee HDTV User Group.
 

News:

If your having any issues logging in, please email admin@milwaukeehdtv.org with your user name, and we'll get you fixed up!

Main Menu

I promised Jim Hall...

Started by Tom Snyder, Thursday Mar 21, 2002, 08:02:00 PM

Previous topic - Next topic

Tom Snyder

that if we saw motion artifacts on the HiDef NCAA broadcast he'd read about it hear.

I'm seeing them BIG time tonight. When the red NCAA logo scrolls... when a I see a guy in a red jersey rolling around on the floor... when they do a sweeping closeup of a player moving across the stationary crowd in the background... even in a couple of the slow motion replays.

Anybody else getting 'em?


 
Tom Snyder
Administrator and Webmaster for milwaukeehdtv.org
tsnyder@milwaukeehdtv.org

Joseph S

When they go to a replay they flash the CBS logo show replay then flash the logo again as they go back to live.

The flashing replay logo is nothing but pixels. The side cameras are definitely not HD and they show it. Other than that and the dropouts it looks fine.

Duke choked big time.   I missed the second half in HD because I'm recording the Kurosawa special on PBS for later viewing.

Pat

I noticed motion artifacts as well a couple of times, and for the first time.  My son has always seen them, but I hadn't.  Also mostly red's.  Maybe its our eyes being more sensitive to red or something like that.

But, although noticeable and mildly distracting, it isn't bad enough to my eyes to be concerned about.  I'll just consider myself lucky.

With the increased sharpness and wide view, the directors can now change the camera positions or lenses to avoid those close-ups of a guy running down the floor after having made a basket.  The longer shots are able to carry all the action in detail without them.

kjnorman

I have never thought about this before, but many DVD players are prone to a chroma upsampling bug that manivests itself on reds.  I have heard that digital cable and satellite boxes can also have this chroma bug.

Now this is a fault of the hardware and not the MPEG stream, but I wonder if HDTV receiver are also prone to this?  This could be why you see more problems with reds and pixellation.

Just my two cents worth...

[This message has been edited by kjnorman (edited 03-22-2002).]

PS.  I do not like these editted messages.  They always show how many times I've needed to edit for spelling mistakes...  


[This message has been edited by kjnorman (edited 03-22-2002).]

GS kid

I noticed the motion artifacts big time. They were the worst when the camera follows a player as he runs back up the court to defend after a basket is made. The winter games on the HD OTA broadcast seemed alot better when it came to this problem. I also noticed that HD seems to love the outdoors. I take that to be because of lighting issues. On the winter games, outside looked better then inside. Ski jumping looked best followed by speed skating, figure skating, and hockey looking the worst. It looked good, but ski jumping outside seemed the best in over all look. Only when there is attention to good lighting, does HD look good indoors. Jay Leno is a great example of it.
----- GS kid

Pat

I've seen the chroma-upsampling bug as displayed by some (most) DVD players, and this is not exacctly the same, but it sure is suspicious that it's the reds.

To me, the motion artifact seems like enlarged pixels.  This corresponds with what little I know of how MPEG works.  As I understand it, MPEG, and in a similar way, JPEG, sends only the differences from one frame to the next.  So when nearly every pixel is different from one frame to the next, it would ideally re-send every pixel.  But there isn't enough time at the data rate available, so it sends less than it should.

There also is a "successive refinement" process going on, so the blocks you see are a sort of average of the pixels the block covers.  As time permits, the blocks are refined by re-sending them as smaller and smaller blocks until the level of the pixel is reached.

There is also a "motion sensing" process that goes on, but I'm not clear on how it works.

So, with a fully static picture, no data is sent at all.  When the change from frame to frame is small, only a little data is sent, and so on.

It would be an artifical intelligence issue if the software or hardware could concentrate on the most important features of the image, and treat them better.  But I don't think that's possible at this time -- maybe next year.