What's new

Danger, Will Robinson, Danger! (1 Viewer)

Michel_Hafner

Screenwriter
Joined
Feb 28, 2002
Messages
1,350
http://www.companynewsgroup.com/comm...p?co_id=110490

I sincerely hope this technology is ignored by all studios dedicated to a quality product. It's as bad as overenhanced edges. How typical of claiming that this is to improve image quality when in reality it's all about lowering bitrates to levels that can't handle grain (so they can squeeze more extras on the same one disc) and avoid the compression uglyness by removing the grain (and image detail), adding grain removal artifacts, compress, decompress and add fake grain on top. Bon appetit.
 

RobertR

Senior HTF Member
Joined
Dec 19, 1998
Messages
10,675
I feel the same way. The emphasis should be on preserving the original grain, not eliminating it, then simulating it.

Not a good development.
 

Ed St. Clair

Senior HTF Member
Joined
May 7, 2001
Messages
3,320
Is everyone one here sure that the "extracted" film grain, is NOT *"faithfully re-created"?
(*as appossed to what has been posted; "simulated")
Are we sure that "faithfully re-created" does NOT mean re-introduced?
Thanks.
Also the article say's this is "mandatory" for HD DVD. So, does that mean every HD DVD title has this "processing"?
 

Michel_Hafner

Screenwriter
Joined
Feb 28, 2002
Messages
1,350
If the system would actually remove the grain and put it later back so the images look the same as the original compressed with a higher bitrate (no visible compression artifacts) then there is no reason to complain. But it does not work like that as I understand it. The grain/noise removed is only statistically described with some parameters and put back as output of a random process driven by these parameters. So there is no 1:1 correspondence. But even that might visually look practically the same. The real problem is that the pictures without grain don't look like the original if it were shot on grainless film (film with no visible grain). The removal always introduces artifacts of all kinds from subtle to in your face depending on the sophistication of the system used, the skill of the operator, the processing time and the characteristics of the footage. Any real time solution is bound to have artifacts plenty which are obvious to trained and sometimes visible to untrained eyes as well. This is not high fidelity in any sense of the word. It does not belong on a high quality product. It's a compromise to save bits and exchange compression artifacts with other artifacts.
 

Ed St. Clair

Senior HTF Member
Joined
May 7, 2001
Messages
3,320
I'm with you on the "always adds artifacts" argument!
I've be trying unsuccessfully to convince DaViD that 1080p to 1080i to 1080p will add artifacts all year! No such luck. :-( ;-)
Thanks for taking the time to "clue" me in. Still this is mandatory, do you think that just the equipment is a must or every title "must" go though this process?

Keep HD Clean!*

*not "clean" of grain, as in clean of any added processing in the transfer.
The time to "clean" is in restoration!
 

RobertR

Senior HTF Member
Joined
Dec 19, 1998
Messages
10,675
From what I just read, this isn't that big of a deal after all. It's just a marketing gimmick that Thompson persuaded Toshiba to use in exchange for RCA putting its name on rebadged players. HDDVD likely will never use it. BD has already rejected it. Here's the complete quote from a decoder supplier:

 

Ed St. Clair

Senior HTF Member
Joined
May 7, 2001
Messages
3,320
Timely find, Robert.
That's bad when the chip maker that produces the decoder slams the 'thing'!
 

RobertR

Senior HTF Member
Joined
Dec 19, 1998
Messages
10,675
Apparently, there's only a spec for using it with H.264 compression. MPEG2 and VC-1 wouldn't even recognize it.
 

Michel_Hafner

Screenwriter
Joined
Feb 28, 2002
Messages
1,350
That depends on the type of 1080i. If it's video based footage where all 60 fields are shot at different times there will be deinterlacing artifacts because it's very difficult to estimate all the motion of all the objects in the whole frame so precisely that a full frame matches up 60 times a second. Half the frame was taken at another time and must be interpolated to the current time (or ignored which means massive loss of resolution). With film based 1080i which shows a deterministic 2:3 pulldown the reverse operation is straightforward as long as the pulldown has no cadence breaks. There need be no artifacts when things are properly implemented.
 

Chris S

Senior HTF Member
Joined
Apr 9, 2000
Messages
2,546
Real Name
Chris S

Why would Thompson care about adding this "feature"? Just to have another bullet point on their marketing flyers?
 

JeremyErwin

Senior HTF Member
Joined
Feb 11, 2001
Messages
3,218
Hmm. While trying to find some psnr statistics for high-bandwidth video (no, google, I couldn't care less about cell phone video), I came away with the strong impression that the newer video compressors separate the film into discrete elements and compress those elements individually. Later, the de-compressor re-composites those separate elements into a frame. So it's not that VC-1 is somehow immune to macroblocking, it's just that those macroblocks are smoothed over (in a controlled, detailed manner). So, if you could separate the film grain from a image, and compress the grain and the despeckled image, it might be possible to get better results.
 

Users who are viewing this thread

Latest Articles

Forum statistics

Threads
356,815
Messages
5,123,855
Members
144,184
Latest member
H-508
Recent bookmarks
0
Top