Star Wars Laserdisc Preservations. See 1st Post for Updates. — Page 20

I think there are so many settings (to be a laserdisc player) to tweak, that is possible to find out a good way to capture with no (or few) chroma errors; yes, some capture tests are needed!

What about Pioneer CLD-D925, 600, and maybe the Philips LDP600WS?

I had to scratch everything last night due to some bad settings but I've got it under control now. Here's a test encode of side 1 of 1130-84 using ld-decode. This has been decoded and run through the software comb filter but no other processing. The audio was also processed with ld-decode's cx decoder.

This should give an idea of how I plan on handling the gout sync'ing.  I just added in black frames with silent audio where frames are missing.  Since these are gout sync'ed it will be easy for anyone to mux in an audio file that doesn't have this.  It sort of gives a visual and audio indication so you know when gout frames are missing :).  This test file has ac3 audio but the final release will have pcm audio.

Side1 <--Good for 7 days.

althor, finished to download the test clip - if you can call it a "test clip" a 1GB file... (^^,) - and it's quite good; a bit too bright, but maybe it's the disc itself...

Meanwhile, thanks to the capture fever that hit me, I managed to grab the same frame you posted before, this:

US HLD-X0 Vs US CLD-R7G
http://screenshotcomparison.com/comparison/129269

and this is the results:

US HLD-X0 Vs US HLD-X9
http://screenshotcomparison.com/comparison/129450

US HLD-X0 Vs JP HLD-X9
http://screenshotcomparison.com/comparison/129452

US CLD-R7G Vs US HLD-X9
http://screenshotcomparison.com/comparison/129451

US CLD-R7G Vs JP HLD-X9
http://screenshotcomparison.com/comparison/129453

I used both the US Definitive Collection (0693-84) Japanese Definitive Collection (second edition, the one without book but with documentaries -> PIL-2021) because this is the latest version and is supposed to have a better video.

Hardware used: Pioneer HLD-X9 via S-Video Monster Cable to SweetSpot (a.k.a. PDI Deluxe)

Software used: VirtualDub with HuffYUV 2.2 codec; settings for JP: http://s11.postimg.org/d834rlmlv/SW_DC_JP_settings.jpg (for US, contrast was lowered to 95 and of course NTSC was used instead NTSC-J)

Well the x9 looks pretty good based on this screenshot imo.  Thanks for posting some comparisons. I think that the colors look a bit muted compared to the x0 and r7g though.

Also, I don't know if the site went down but I'm now getting 404's on all of the above links?

I've left the color setting as is, because I was afraid that if I raised it more, some rainbow artifacts could happen...

About screenshotcomparison: here it worked always, lately... some weeks ago was down for several days, though!

Star Wars P&S 1130-84 Laserdisc is now available at a.b.sw.  This was decoded with ld-decode. I think it looks alright for a 30 year old laserdisc.

There's a bit of rot here and there and the audio might crackle once in awhile. This is a preservation though and not a restoration so I made no attempts to clean up anything. During the gout-syncing I just added black frames if frames were missing. This laserdisc apparently had some sort of analog video noise reduction applied so the scene change artifacts are in the source.

EDIT: I'm going to start on ESB (1425-84) next but first I'm doing some comb filter testing and then I'm on vacation until August so it'll be awhile before it's ready.

Althor, you have a PM. Also... good work, carry on. :)

Does anybody know of any efforts to preserve the Laserdiscs (US and Japan) of the documentaries? (Making of SW, SP FX: ESB, Classic Creatures: ROTJ, and FSTTJ)

ESB P&S 1425-84 is now available at a.b.sw.

At this point, it seems it may be awhile before ld-decode can handle digital audio although Happycube is still working on it and seems to be more and more hopeful with each update.

For the time being, I believe I will capture the video with ld-decode and use other sources for the audio.  Thanks to people like Schorman and Hairy_Hen there are many sources available that are gout sync'ed already. Schorman's laserdisc audio archive contains bit perfect audio for many of the exact ld's I'm trying preserve so most of the time the exact audio I need would be available.

ROTJ 1478-80 is next.

Would it be alright if you added info on the front post regarding the status of each LD? i.e. 1425-84 is COMPLETE or something like that.

Yeah. I had an "x" next to the ones that are done but it didn't look that great. I do plan on doing something similar to what you suggested sometime in the future.

So I'm kind of tinkering with manual ivtc of 1130-80(fox P&S/time compressed). Like Mallwalker says the first part can be done successfully with this:

side1.assumetff().separatefields().trim(1,0).selectevery(61,0,1,2,3,5,6,7,8,10,11,12,13,15,16,17,18,20,21,22,23,24,25,27,28,29,30,32,33,34,35,37,38,39,40,41,42,44,45,46,47,49,50,51,52,54,55,56,57,59,60).weave()

However, when I get to the point where the droids walk into the hut and it cuts to c-3p0's oil bath I can't get it back on track. I've counted up to about 120 fields and it seems to jive with the cadence up above but it's definitely not right for some reason. Is it possible that this cadence instead of being 61 fields long is maybe like 200 or 300 or more?  Is anybody willing to take a crack at it? I probably will work on it more after I get done with the swe for esb and rotj.  Something tells me this old P&S will be a bear to properly gout-sync though.

If I'm to join team_1130-80 then I should get my source(s) aligned with yours.

If you do side1.assumetff().separatefields()
then what's the field number when the title card first appears? (GOUT 689)

Likewise side2.assumetff().separatefields()
What's the first field of the 2nd shot? (GOUT 87380, Han disengaging hyperdrive)

1130-80 Side2 IVTC script is available for beta test.
Check the trim() command at the top, and adjust to match your source.
Starting from zero, we expect field 127 to be exterior shot of Falcon in hyperspace.
Field 128 should be Han disengaging hyperdrive.

If you find script errors please tell us.
We already know about this guy:

He's not a mistake-- he's a Deliberate Creative Decision (TM) for our coach, waiting to happen.
Script contains pseudocode to that effect.



If that script doesn't make your heads spin, then this preview of Side1 might do

We said:

cReel3B=\
cSide1.trim(175410,211072).doubleWeave().selectEvery(242\
,000,002,005,007, 9,012,014,017,019,022,024, 26,029,031,034,036, 38,041,043,046,048,051,053, 55,058,060,063,065,068,070, 72,075,077,080,082, 84,087,089,092,094,097,099, 101,104,106,109,111, 113,116,118\
,121,123,126,128, 130,133,135,138,140,143,145, 147,150,152,155,157, 159,162,164,167,169,172,174, 176,179,181,184,186,189,191, 193,196,198,201,203, 205,208,210,213,215,218,220, 222,225,227,230,232,235,237, 239\
).assumeFPS("ntsc_film")

cSync selects the first two fields, while
cSynch picks the 2nd & 3rd field from each '3' frame.




In the LD those are consecutive frames; however, they are one and the same film frame: GOUT 26252.

O the Deliberate Creative Decisions . . .

This is only a 2-field frame so cSynch can't help.
Instead we can use that method recycled from our JSC script.
(There used to be a picture of Red leader behind the dead link in that post)

Thanks for the update Mallwalker. It's always fascinating to see the quirky little things about each different release.  BTW, side2 seems to look good to me. I haven't gone through each frame of course but I did watch it through in real-time and didn't notice anything out of order.

It's a shame that side1 has some audio and chroma noise towards the end. I will probably step outside of my normal modus operandi and apply some chroma denoising to it just to reduce the flickering.  I wonder if the cbs/fox pressing has less noise than the plain old fox release that I have.

I can see one of them now.

They did a separate telecine session using spherical lens, then edited those shots in the video realm.
That's not too surprising itself, but it gets weirder.
Instead of time-compressed, the replacement clip(s) are time-expanded.
Rolling slower than ntsc_film: 50 frames comprise 127 fields.

That was the last of the unidentified rates, now solved. Almost there...