HalfD1 Mode seems to use Non ITU compliant resizing

[scharfis_brain]

Thanks.
This Page describes it quick and correct!

I am very interested into your test results!
(At least because I don't have much equipment to test on)

[onesoul]

@ scharfis_brain

I tested the (tiny ;)) DVD you supplied with software PowerDVD 5.0 (I will test later with stand alone player).

The results:

Titles 1, 3 and 5 displayed exactly same wrong "roundness" in my 19" monitor of 9,05 centimeters wide and height of 8,85 cm.

Titles 2 and 4 displayed correct roundness with a diameter of 8,85 centimeters.

These results concur with what jdobbs described on his test with tv and stand alone player which seems that the 704 and the 352 are being stretched wider.

Now I am wandering if many or few stand alone players are behaving like these.

[Jeffster]

@scharfis_brain

I downloaded your circle test out of curiosity, although I wasn't sure what it would achieve at first (I guess it was only for SMD).

Anyway, from your two conclusions I am now able to confirm that both my standalone DVD players scale correctly (Philips & Panasonic)... also, while the circles appear to be round from a distance, measurements show they are fractionally off, as best as I could measure, which would indicate my TV is ever so slightly off.

Thanks, an interesting test :)



Edit: Removed part of my post about halfD1 resizing. In retrospect I don't think it was helpful to the debate.

[Kika]

@onesoul

Some Players are doing it the right way, some not. I've checked the ITU-rules on my two Players: a Denon DVD2910 and a Afreey LD2060 - both of them will show me the correct picture. So my results are different to jdobbs results.

But, and that's the problem: With the replacement TV i have since Monday, the results are very different - none of the testfiles will give me a correct AR.


@all
Just an idea: Maybe some encoders or authoring programs are writing wrong flags? Like generic PAR for Half-D1 and so on?

[scharfis_brain]

@Jeffster: this test is already included in my disc.
see VTS 2, 4 and 5.


also I cannot test this because my TV doesn't handle 60 Hz.

[Jeffster]

@scharfis_brain
Sorry, I got a bit carried away. :o
I do agree with your conclusions though.

[scharfis_brain]

SMD: any measuring results jet?

[TuRiSOft]

@ all : come here (http://forum.doom9.org/showthread.php?s=&threadid=92951) and vote for your favorite encoder in DVD-RB!!!!

Bye and Thanks!!!

[SeeMoreDigital]

Hi scharfis_brain,

Here are the results of the British jury.... Jeez.... that's very Eurovision Song Contest!

Anyway, using my Pioneer DV-575A DVD player connected via component video to my 42" Hitachi 42DP5000 plasma, I obtain the following: -
Encode 01: Generic PAR: 768->720 - 320w x 318h
Encode 02: ITU PAR plus Overscan: 768->8+704+8=720 - 312w x 318h
Encode 03: ITU PAR without Overscan: 768->704 - 312w x 318h
Encode 04: ITU PAR plus Overscan: 768->8+704+8=720
Reduced straight to 352 - 306w x 318h
Encode 05: ITU PAR plus Overscan: 768->8+704+8=720
Cropped then reduced to 352 - 312w x 318h
And using my Sigma Xcard connected via component video to my 42" Hitachi 42DP5000 plasma, I obtain the following: -
Encode 01: Generic PAR: 768->720 - 315w x 315h
Encode 02: ITU PAR plus Overscan: 768->8+704+8=720 - 308w x 317h
Encode 03: ITU PAR without Overscan: 768->704 - 315w x 315h
Encode 04: ITU PAR plus Overscan: 768->8+704+8=720
Reduced straight to 352 - 308w x 317h
Encode 05: ITU PAR plus Overscan: 768->8+704+8=720
Cropped then reduced to 352 - 315w x 315h

Cheers

[scharfis_brain]

hmmm. somehow missed your post.
sorry.

Encode 01: Generic PAR: 768->720 - 320w x 318h
Encode 02: ITU PAR plus Overscan: 768->8+704+8=720 - 312w x 318h
Encode 03: ITU PAR without Overscan: 768->704 - 312w x 318h
Encode 04: ITU PAR plus Overscan: 768->8+704+8=720
Reduced straight to 352 - 306w x 318h
Encode 05: ITU PAR plus Overscan: 768->8+704+8=720
Cropped then reduced to 352 - 312w x 318h

Here, it seems, that your player is doing 720 <--> 704 pretty fine, but one of both devices scales incorrecly!

I came to this conclusion, because 2, 3 and 5 do have the same width!

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


Encode 01: Generic PAR: 768->720 - 315w x 315h
Encode 02: ITU PAR plus Overscan: 768->8+704+8=720 - 308w x 317h
Encode 03: ITU PAR without Overscan: 768->704 - 315w x 315h
Encode 04: ITU PAR plus Overscan: 768->8+704+8=720
Reduced straight to 352 - 308w x 317h
Encode 05: ITU PAR plus Overscan: 768->8+704+8=720
Cropped then reduced to 352 - 315w x 315h

Wow! That's really odd!
height NEVER should change.
either it is a measuring error, or the sigma-card is really doing crap ALL the time...

but when I ignore the heights (assuming they were identical) I see that the card doesn't treat 720 pixels width in the ITU-way
1,3 and 5 are equal as well as 2 & 4 are!

but here, the AR of 3 & 5 seem to be correct. so I ask, how are both, the DVD-Player and the Sigma-Card connected to your display?

[SeeMoreDigital]

Originally posted by scharfis_brain
... so I ask, how are both, the DVD-Player and the Sigma-Card connected to your display? Both devices are connected using analogue Y,Pr,Pb (YUV) component video. But I also tried composite and s-video connections

The Sigma results surprised me as well... So, just in case there were any discrepancies in the .IFO file, I decided to test each one of the five encodes again by de-muxing the video streams out of the VOB container.... I obtained the same results.

Very odd indeed!


Cheers