AVS Forum banner
Status
Not open for further replies.
1 - 7 of 7 Posts

·
Registered
Joined
·
29 Posts
Discussion Starter · #1 ·
As a preamble, I hope this post doesn't come off as antagonistic towards Dilard. I truly think it is very professional in its implementation. It is very impressive and Mark, as its author, is equally impressive as a developer.


That said, I have two concerns about letting Dilard calibrate my projector.


FIRST

=====


It seems to me that the result can be only as accurate as the video card in the PC. If the card is non-linear or it has tracking errors between the channels, the results are necessarily going to be off the mark (no pun intended!).


Of course, it may be that our collective eyes won't even notice the slight errors. Also, for HTPC users, this may actually be a plus - the errors in the projector calibration cancel out the video card errors (although there may be some dynamic range that is sacrificed).


SECOND

======


The DILA has something like 16 points which can be adjusted for each (R,G+B) gamma curve. Does Dilard actually attempt true curve-fitting in that it will move each point along both the X and Y axis? Or is one axis fixed in that the values are arbitrarily chosen? How many shades of gray are actually sent to the projector as test patterns? With a human, you could only do so many, but with the computer at the helm, it seems to me that you could send a ton of brightness levels and check and adjust each one against its target result.


It seems to me that this could be the REAL strength of computer controlled calibration - the multitude of test patterns that can be generated and adjusted for.
 

·
Registered
Joined
·
842 Posts
quote from above


The DILA has something like 16 points which can be adjusted for each (R,G+B) gamma curve. Does Dilard actually attempt



WRONGOLA. The factory tool has 17 points 2 of which

are fixed. SO WHAT. The actual table is 1024 x 1024

points. As far as i know only william phelps software

and dilard use the entire table.


And yes you can see the difference between 17 points

with straight (or starling) lines and a full calibration.


As far as video cards, some are really good (<1%) and

some are pretty rotten (>5%). So if you have a rotten

one, go and spend $100 and get a decent one.


Someone needs to sell calibrated video cards.
 

·
Registered
Joined
·
29 Posts
Discussion Starter · #3 ·
Kevin -


Thanks for addressing my "curve-fitting" concern. So, can I assume that Dilard sends 1024 shades of gray and adjusts each individual point accordingly?


(I'm hoping that I'm RIGHTOLA in this case...)
 

·
Registered
Joined
·
4,525 Posts
I don't think that your post was antagonistic at all. Just questions, that's all...


Number 1.) - Pretty much what Kevin said. Results may vary based on the calibration "equipment", but it shouldn't be by very much.


Number 2.) - The 10 bit gamma (2^10) in the projector has a lot of range. The JVC software doesn't do it justice, although 16 points, if placed correctly can do fairly (that is the way that these calibrations used to be performed!). In the Calibration Wizard, you get to select the number of grayscale levels to characterize, and even specify the transfer function.


You are not limited to 16, 20 or even 32 levels. I use 32 myself, because adding more just makes the process take longer and I find no real difference in the end, but if you wish to use 64 (or more) levels, you certainly can. 1024 levels is going to be overkill, and will cause the process to unnecessarily take hours.


The default is 32 levels, and I have found in practice, that beyond that, little (if any) improvement is possible.

http://www.dilard.com/dilard/images/.../calwizard.jpg

(Note: This is an old screen shot. It is used to display the # of gray scale levels choice only. The rest has changed.)


Thanks!


PS - Just as a silly side-note, the Dilard look-up table is compatible with the JVC software. I just felt that it was important to remain compatible, if just to keep on JVC's good side.


[This message has been edited by milori (edited 09-10-2001).]
 

·
Registered
Joined
·
29 Posts
Discussion Starter · #5 ·
Mark -


Thanks for the clarification. It seems really well thought out.


I guess at this point, I have no more concerns... (except that I didn't take advantage of the pre-buy!!!)



[This message has been edited by mjws (edited 09-10-2001).]
 

·
Registered
Joined
·
611 Posts
This is related to the subject --- sorta.


I am reasonably happy with my 9015 image. But, I did order the DHU so that I could use Dilard 2.2 to calibrate. My question: If I let 2.2 do its thing after a full backup, can I confidently revert to my pre-calibration condition if I don't like the result? IOW, are there any values modified that the backup ignores? TIA

 

·
Registered
Joined
·
842 Posts
Yes, I've been backing up and restoring throughout my alpha and beta testing and so has Mark.


Also, we're going to extend the pre-buy another week since CEDIA kept everyone so busy and we've had tons of emails asking us to extend it. I'll make a post to that effect later tonight.


------------------

Tom Strade

Immersive, Inc.
www.immersiveinc.com
 
1 - 7 of 7 Posts
Status
Not open for further replies.
Top