Spaced Ranger

msycamore said:

Scene order in English home video versions and 2004 Cut:

Drug evasion arrest --> Trial & "pain prods" (electro-shock batons) scene --> THX gets medical examinations --> THX is being monitored & researched --> LUH visits THX --> THX put in prison

Yes, I remember that. But first, a theory. My thinking was/is:

Lucas' original cut [1st version, 1970] was never seen when Warner Bros. took it away from him. WB then re-edited it (only removed some minutes of material -- by Lucas' own account) for theatrical release [2nd version, 1971]. Lucas' Star Wars success in 1977 gave him new clout to demand that THX 1138 be restored back to his original cut. So, it was this originally unreleased 1st version that was finally released in WB's 1978 theatrical "re-release" and, thereafter, on television.

Note: If your preservation of the locally(?) re-edited Italian TV (with flashback edits, voice-over, and "rush-job" snipping) has the missing SEN-talking-to-children shots, then it was recorded before (or shortly after) 1983. Otherwise, it was a "next version" recording.

The next version (so goes my theory) was with the home-video releases. Lucas himself removed some of the SEN-talking-to-children shots and re-ordered the THX-caught sequence. This was Lucas' second cut [3rd version, 1983].

The next version was the new home-video releases (notably DVD), which contained old outtakes and extensive CG & digital manipulation. This was Lucas' third cut [4th version, 2004].

(I haven't reviewed the 2010 Blu-ray release to determine if it is yet another cut or just color-timing cleanup.)

This makes sense, I believe, because it fits with Lucas' incessant tinkering with his already released works. (Or ... maybe it's just his father's business sense, in him, to resell the same product ... for more enjoyment and greater efficiency ... to make another buck?)

Anyway, I reviewed the 1997 Bravo letterbox broadcast recording (the US laserdisc vintage, I expect) and came up with your same "English home video versions and 2004 Cut" sequence of THX-caught:

• [dissolve] 2 roboPolice escort THX up hallway [fadeout]
  readout: "Drug Evasion Arrest" "Place In Research Cell Pending Trial"
  beating by roboPolice w human supervisor
  readout: number counting 5 to 6
• trial & computed verdict w readout: "Conditioned" "Detention"
  [dissolve] electro-prod live-practice by roboPolice in white limbo [fadeout]
medical exam & drugging w readout: "Chemical Imbalance" "Reusable Parts"
[dissolve] 1 roboPolice escort THX down hallway [fadeout]
• technicians neuro-conditioning THX
• THX & LUH join in white limbo & separation by roboPolice w human supervisor [fadeout]
• [fadein] delivery to detention in white limbo

Thanks.

I was planning to set up a parallel-stream playback to check what was juggled around and if any further edits were hidden in those swaps. (Frames-gone-off-sync would be regarded more as possible frame-drops rather than edits, until otherwise verified.)

metaleonid said:

I tested LD Video Essentials with bunch of cards including ATI 750 HD PCIe. ... Unfortunately ATI 750 HD PCIe triggered AGC problem. This same problem was triggered also by ATI 750 HD USB when I used S-Video out from LD-S2.

As far as comb filter goes 750 performed the best from DVD test captures. However, on LD VE capture it didn't do as good job perhaps due to the LD noise.

Is this AGC problem mostly a torture-test result that doesn't noticeably show in real-world captures? althor1138's Star Wars test capture back on page 1 seemed to be (shall I say?) impressive, but he mentioned small inaccuracies in the pattern tests.

@ SilverWook

I'm guessing that you've watched both the U.S. laserdisc and the 16mm enough times to definitively know their differences.
Are the newer WB logo and the removed SEN cuts/shots (5) when talking with the children, the only editorial changes between the 2 versions?

AntcuFaalb said:

Part 2: The Comb Filter

... I find 3D comb filter artifacts ... more objectionable than the composite video artifacts ... that a 3D comb filter is designed to remove from stationary frames. ... OTers who love 3D comb filters can always use a Leitch DPS-475/575 or tritical's TComb on the raws I'll be posting to Usenet.

So, for the present, only 2D comb filtering works without undesirable side-effects on moving picture areas that 3D filters generate. What about your research into the idea of a "counter-phase overlay", as a better comb filter technique for moving picture areas?

[BTW, to get up to speed on 2D/3D comb filters, see Extron Electronics' single page write-up NTSC Decoding Basics (Part 4) -- http://www.extron.com/company/article.aspx?id=ntscdb4 .]

Oh, BTW, good job, Harmy (here and on the others, too)!

As GL would quietly gush on-set ... "great".

armetilou said:

Verbs burn beautifully but don't have a good shelf life. ... Here is a Club Myce thread with some relevant information, including a link to a French degradation study showing that Panasonic and Sony are better for long term storage. ...
http://club.myce.com/f33/best-bd-r-archival-330175/

Read'em and weep (Verbatim doesn't fare well as either LTH or HTL):

http://www.myce.com/news/french-research-avoid-blu-ray-lth-discs-for-data-archival-64265/

deanwitty [Moderator]: "There seems to be clear agreement with other studies I've seen that Panasonic and Sony are the longevity leaders. Verbatim with not even 1/4 the lasting power of those two."
http://club.myce.com/f33/french-study-bd-r-archival-329441/
Their further discussion of other brands is certainly worth it for the one page read. Off-topic (even there) certainly is valuable!  ;)

AntcuFaalb said:

Part 1: The LD Player

... The only LD player that ... meets these requirements is the Pioneer CLD-1010.

Aside from "LD Player Requirements: ... 3. Cheap!", would your Panasonic LX-900 be an equivalent or better choice for best captures?
(Just an intellectual-exercise question for those with possible-future-purchase-options in mind.)

I was thinking ... if anyone (maybe the OP?) wanted the completeness of the Image Entertainment release but the better color of the Warner Bros. release, color correcting the IE might be the easiest way to go. So I tried it with the previously posted frames from The Mad Scientist episode.

 
                               color correct this                                                                           to this

In a paint program, use the Histogram Adjustment to alter the individual Red-Green-Blue components of the IE snapshot, while comparing it to the WB snapshot. Use an eye-dropper tool (one built in or an external program) to balance the whites & blacks and the R-G-B's of main colors in between. This is what I came up with after some back-and-forth (it takes some experimenting & practice to see what the settings do to the picture; the eye-dropper helps by showing the R-G-B numbers and which direction they move):

HISTOGRAM
RED        Gamma 0.7      Midtones +15 [compress]
GREEN    Gamma 0.75    Midtones +10 [compress]
BLUE       Gamma 0.7      Midtones +5   [compress]

This produced a good balance but the colors were weak. So I followed up with the Hue/Saturation/Lightness to increase the color strength:

H/S/L
SAT       Saturation +55

The resulting color correction ..

.. is not perfect, but it's pretty close.

                     IE original                                          IE color corrected                                       WB original

The releases seem picture-stable enough that this single correction should work for the full episode and maybe the entire series.

SilverWook said:

I think the experienced veteran Cinematographers George worked with probably helped too. ;)

Not to take anything away from GL, he was just not good enough to make THX 1138 as well as it came out. He knows it (he knew it even with the Star Wars films) and lamented it in one of the many George Lucas / Star Wars documentaries.

Don Pedro Colley explains, in this LucasFan.com interview ...
http://www.lucasfan.com/thx1138/interv.html :

At first, it felt like we were doing another student film. However, when the final cast was assembled, we worked very hard at letting George know that we were all responsible .... Every day, before we shot a scene, the four of us would brainstorm to get to the subtextral content in the script , on film. We kept telling George this film will not work unless we are able to breath life into these characters. Unless an audience can connect with these life forms, who live in that real world, we would have nothing!! George wasn't too happy to have to deal with our ideas, but he came around.

ww12345 said:

... "old school" cinematographers shot like that. ... I always over-frame shots, knowing that I can crop it in the darkroom to the image that I really wanted. ... a lot of cinematographers who came over from still photography (a la Stanley Kubrick) knew and practiced these methods.

George Lucas never had that background -- he began on motion pictures. What he framed was what he wanted (or the best he could get). Interesting that you mentioned Kubrick. I read where he was on the set of 2001: A Space Odyssey with a square camera viewfinder composing the shots for his cinematographer! How's that technique for a Cinerama extravaganza.  :)

THX11384EB said:

... the Japanese LD? I have it, and would like to help out any way I can.

Thanks! I don't have the project resources yet but I'll keep in mind anyone who would like to help -- which I imagine would be mostly in the extras now (that still would include comparisons to previous & contemporary versions).

AntcuFaalb said:

[Part 1: The LD Player]

[Reserved for Part 2: The Comb Filter]

[Reserved for Part 3: The Capture Card]

[Reserved for Part 4: Cables, etc.]

Awesome -- lookin' forward to this!
Got all my monitors on ...

SilverWook said:

http://www.hometheaterforum.com/topic/319803-tv-shows-and-tv-movies-gone-w-i-d-e/

Whoa. That was some link. Another evening lost to in-depth research.  :D

Very interesting to learn -- all those early TV shows (60's+) were shot over-framed (like the above Star Trek demonstration), only with "4:3 protected" for broadcast on their day's TV. Unexpectedly forward-thinking. (Thread OP HDvision is my new hero!)

SilverWook said:

There's a whole thread debating this contentious issue at the Home Theater Forum though.

Come on, SilverWook! Where's the link?  :)

HTF is a nice place to talk about the weather.  [Stimulating rhetoric? The theater of noise is proof of our potential.]  But here at OTF we actually do something about it!

SilverWook said:

I've only found one that broadcasts in 16:9 ... but some of the old shows like Charlie's Angels are horribly cropped 16:9 remasters.

I'm sure many of those old TV programs wouldn't be horrible in widescreen if they did them right ... by going back to the negatives.

The Star Trek (the original series) remastering for HD was a herculean effort and a missed opportunity for an awesome widescreen version. In those debates, the "integrity of the director's framing" is the oft-made argument but that is a phony defense (it's always a money issue of tight-wads who roll in the dough).

Here's a Star Trek you have never seen unless you have a film strip. This is was the way the cameras photographed it:

The reason you never saw this much was because it was framed for television -- the so-called "Director's framing" of the shots:

But you never saw even that much (at least, when it aired on the TV's of it's day) because the imprecise analogue technology forced a tighter "safe area" to guarantee what could actually be seen. This is the true "Director's framing":

So making a widescreen Star Trek means merely extending the edges of the safe area to the edges of the remaining image on film. Sure it might need the inevitable fix-up, but as this sample demonstrates, the TV show was cleanly shot:

SilverWook said:

The WB Tv logo at the end dates the print to pre 1984...

So, any release with the WB Kinney logo would have been mastered sometime 1970 through 1972.

However, with Lucas' Star Wars success (post 1977) and WB subsequently re-releasing THX 1138 (with the WB cuts restored), did WB simply put the excised footage back into the master and make new distribution prints (Lucas' restored original/1st edit) ... with the original logo also unchanged? Lucas probably wasn't even involved in that one (beyond demanding that it be done).

From the WB re-release forward, a TV broadcast would be of Lucas' restored original/1st edit (excluding mangling by individual TV stations or networks). So any broadcast that included those missing SEN shots would identify it as Lucas' original edit. -- Does this sound right?

AntcuFaalb said:

Also, the answer is: yes! :-)

Excellent.

You are a true believer.

Once I get the project resources I'll send you an inventory ... if you can use anything already accumulated (other releases, etc.).

msycamore said:

SilverWook said:

... Pan and scan can be less obvious to the eye when it's done on film versus a video master.

Yeah, it's basically down to the telecine operator how to handle it ...

So, the difference of my post-1990's TV recordings is just the original (1st) Lucas edit (the 16mm period) with it's P&S release doing snap-scans and the next (2nd) Lucas edit (the laserdisc period) with it's P&S release not doing snap-scans. That would make the laserdisc-period release Lucas' 2nd edit of THX 1138.

Keep in mind that stations can have unlimited play rights -- meaning they can broadcast an "older print/version" that they have after a newer print/version is released. This often happens with new & improved home video releases (like "restorations") that don't broadcast until some significant time afterwards. So, the 1st version could be broadcasting while the 2nd version is being released on home video. -- Is that too weird an extrapolation?

ww12345 said:

For what it's worth, Spaced Ranger, I think luminance stabilizing has to be done on a shot by shot basis.

I was afraid of that.

AntcuFaalb said:

The easiest way to do it is to lock the brightness+contrast of a shot to that of its first frame. ... A certain program that ends in "clean" does this rather well.

Mister Clean?

I know he does Windows. But does he do Mountain Lion?

BTW, your PM inbox is full. So ...
... if you are doing the THX 1138 LD capture(s) any time soon -- which I expect will be the best ever -- could that one, with our extras, be the combined project? And no problem if you want to do the final construction.
If not, I'll continue with SilverWook's excellent captures. But I had to ask, for the obvious reason of why we're here.  :)

Puggo - Jar Jar's Yoda said:

... the light/dark cycle ... I have seen before, but I don't recall if this movie had it.

BTW, thanks for weighing in on this. I was afraid it might've been like picking up the beat when filming under fluorescent light.

SilverWook said:

It's unlikely it's a broadcast print, (except for the third reel) as even partial nudity was not going to be shown on broadcast tv.

I don't know about that. Those 2 TV broadcasts, from which I posted snapshots during our early color correction tests, dealt with it:

• The fullscreen cablecast on TNT (1995) cut the entire nude-hologram dance sequences and had cuts & close-ups in the THX & LUH white-prison intimacy sequence.

• The widescreen cablecast on Bravo (1997) left those sequences mostly intact but did whiteout-hidden-identity circles with head & shoulders close-ups in the nude-hologram dance sequence ..

<-- unretouched capture
.. and zoom-ins in the white-prison intimacy sequence.

TV broadcast content was shipped out to the stations on 16mm. Thankfully, brave someones held onto them ... for posterity sake! GL revisionism foiled again!

You are going to restore the original WB/Kinney logo myscamore has worked on to match the LD?

I don't have most of the project sources yet. Of course, a preservation will have what's actually on the laserdisc to be preserved. My rough mock-ups were just to compare it to a positive ID'ed previous version (whatever version number that might be, LOL) in the 16mm -- sorry if that made for any confusion.

All the extras msycamore has already (plus whatever else comes up) will be bundled with the preservation. I even thought of trying DVD's alternate view function to put some of it in the movie stream. But that's new territory to me. I don't want to get stuck on it if it won't show smoothly, or if it will make for a problematic source-extraction for future restoration projects. Such additions will be in the extras part anyway.

msycamore said:

Don't forget the '78 re-release trailer (it's low quality but kind of neat) and the 4 original 1971 Radio Spots ...

Definitely, and congrats on tracking down those radio spots!
Did I mention I hate hard drive crashes?  :)

All English home video versions of the film from 1982 - 1995 lack this extended footage ... SilverWook's 16mm print also shares the same weird editing differences in the middle of the film (reordering of scenes) as those two broadcast versions.

I just found something else while writing this up and referencing those (U.S.) TV broadcasts. Just before the SEN-missing-scene, Silverwook's fullscreen print (19??) has the children playing rats-in-the-maze with snap-pans to show them all running around. The TNT fullscreen cablecast (1995, without the SEN-missing-scene) does not pan during that shot -- it's locked at a single position.
Yet another version? (At least, a different fullscreen printing for the version-mix.)  :O

I just tested AntiFlicker and DeFlicker plugins for Avisynth. Neither can properly handle a shot with anything moving in and/or out of frame (like the opening, scrolling titles). Even mostly static shots (like the long-shot of the boy walking toward SEN) continued to fluctuate. I would expect the commercial product DaVinci Resolve Lite to be "smarter", but I can't test it to discover how much better. If this problem was positional film-damage, then it wouldn't be cyclical for very long with ever decreasing circumference while unreeling. Fortunately, a semi-manual solution wouldn't be too burdensome.

Only these small sections will be included as "interesting extras" for this preservation. A short video presentation of the differences would be nice, something like this (rough mock-up):

and this:

If we have a better guess on when the "missing scene" disappeared, the above date will be corrected. Feel free to chime in.

Depending on extra disc space, the full-size 16mm clips would be included, sprocket holes and all.

I haven't looked into those yet. If they can properly handle what looks like a 1½ second peak-to-peak (dim-to-bright or bright-to-dim) variation, that would be great.

Another thought would be to try the free DaVinci Resolve Lite program (in another system -- the one I'm using can't handle it), noted back in the Song Of The South thread, for it's luminance stabilization.

Since it is cyclical, I might try a 36-frame manual adjustment (and reversed) repeated across the timeline. Hmmm ...

Puggo - Jar Jar's Yoda said:

... could be a damaged sprocket hole.

I only noticed one of these when playing the clips (over & over), but hadn't considered the mechanical aspect of it. That's good, as it would mean the frames are all there and such a misaligned frame could be easily reconstructed.

Have you noticed the rhythmic, slight, bright/dim variation in your other captures like I observed here? It can be seen on any steady background that stays on screen long enough to pick up on it. (That's why my earlier sample-frame color correction was not stable across even individual shots.)

@ msycamore

PM sent to Puggo. Yes, interesting that the 2nd audio needed to stretch an extra ¾ second to stay in sync during clip-section-2 (1min:33sec). A mistaken assignment of numbers -- 23.98 / 29.97 fps, 48000 / 44100 Hz -- wouldn't account for that small an error.


@ SilverWook

I searched around, too. Even YouTube. Nothing ... not even once-there/now-deleted. (Who's been scrubbing this Internet anyway?)

Puggo - Jar Jar's Yoda said:

Impressive!

Hi, Puggo - Jar Jar's Yoda! I'm using your quote from a year ago in case the forum flags posters of replies (if apparently not, I'll just alert you with a PM).

I'm going over your sample 16mm-transfer clips made some time back for msycamore. He was having trouble with syncing the video/audio and posted it's re-mux, which also ran fast.

I initially noticed that the video portion was identified in Media Player Classic: Homecinema as "Video: DVSD 720x480 29.97fps 28771Kbps". Of course this video was a film capture that should've been running at 24 frames per second or "23.98fps" in the video. Setting it to film mode in an Avisynth script ...

vid001 = OpenDMLSource( "bits.avi" ).AssumeFPS( "ntsc_film" )

... brought it down to it's proper speed. As the .ZIP file I received was the original one from you (I'm guessing), I was wondering if you could describe your transfer procedure and what software you used to get the video into this final state.

msycamore mentioned that the corresponding audio clips were captured separately. These, in the MPC:H player, were identified as "Audio: PCM 48000Hz stereo 1536Kbps". However, when I tried to sync the longest running dialogue clip (about 1½ minutes), the video could not keep up with the audio -- starting in-sync and ending significantly out-of-sync. There wasn't any obvious double-frame-captures to explain it (and there would have to be quite a few for this bad of a problem). So I tried setting the audio at a lower rate ...

aud002 = WAVSource( "audioBit2.wav" ).AssumeSampleRate( 47650 )

... to stretch it out over the length of the video (but not exactly -- I adjusted it by trial-and-error). Then, of course, it remained in-sync over the entire the clip. So, again, please describe your procedure and software for preparing the audio.

One thing I did find (and I will go over the video, frame by frame, to look for more anomalies) was what appears to be an "impossible capture" of misaligned-fields in an interlaced frame (!):

This would indicate a bug in the transfer/processing software. (It may affect the frame count, though probably not enough for this sync problem.) You might want to check into it.

One other thing I noticed was the periodic lightening/darkening of the opening titles' blank background. Could it be some kind of AC hum varying the projector brightness, which is caught at various points by the camera capture? If so, that would be a major design flaw in the hardware -- which then might need a power regulator for the light. Or it could be just this print. You might want to test your equipment with a length of uniformly dark leader.

drngr said:

HDMI 1x2 splitter that strips HDCP + Blackmagic Design DeckLink Mini Recorder ...

What do you think about the Speed Demos Archive forum's thread I mentioned earlier? There's allot of activity there (regarding BlackMagic and other capture cards) and the mixed results with various splitters -- where some previously working boxes seemingly have been updated to be non-working, as regards taming HDCP.

SilverWook said:

I need to get on photgraphing that vintage newspaper supplement with the Lucas interview from '71.

Do you mean ... THX 1138 - Made in San Francisco (American Cinematographer, 1971) ... in the George Lucas Interviews book?
http://books.google.com/books?id=P2P7pwHeZSkC&pg=PA8

A good quote to keep in mind from manic-revisionist George Lucas (1971): "... with minor exceptions, THX came out pretty much as I had visualized it ..."