schorman13

Thought it was Austrian.

Basically, the German HDTV rip has the best picture quality, however, there are many instances of missing frames, damaged frames, encoding/decoding errors.  In v.1, I patched these using the "hV" versions found on torrent sites.  The "hV" versions are re-encoded from different UK Sky TV Broadcasts.   In some cases, the Blu Ray was utilized to help fix errors as well, though very rarely was it used as is.  Due to the re-encoding and different bit-rates, the German and hV versions didn't exactly match-up perfectly, and some sharpening /color, contrast, or brightness tweaking/grain addition was necessary to match the different sources.

Errors in the German rip were initially spotted by resizing, inverting, and merging the HDTV rip with the 2004 DVD versions.  This made most errors in the HDTV rip relatively obvious, and the patches were put in to fix these errors.  AOTC is by far the most complicated film in this regard, as there are encoding errors roughly every 10-15 seconds during the film (This might be an exaggeration, but not by much!).  Other films, like ROTS, had only a couple of errors (I think it only has three or four patches in total). 

Anyway, v.1 looks pretty great as it is, and it's frame accurate with the 2004 DVD's, however looking closely at the video reveals frequent encoding errors on the stream.   These errors are not readily apparent using the procedure I used to look for errors in v.1.  I can post pictures later, but they appear as blocky frames and frequently display horizontal banding in the image.  

I have noticed a few patterns to the errors.  Mostly, blockiness is apparent in dark/black portions of the frame and the encoders used seem to have a problem with the color green.  Basically, dark scenes and scenes with lots of green/vegetation have the most errors.   As such, AOTC and ROTJ have many errors.

About a year ago, I started to work on v.2, and completed new versions of AOTC and ROTJ using the same sources for patching.  However, they were extremely tedious and hard to slog through.  In addition, the patches were not matching as well as I would have liked, and I just got a little discouraged with the process.  

After another thread mentioned some alternate sources for HDTV rips, I was able to download some French-language versions based on the same UK SkyTV  rips.  However, these were much higher bit-rate, and not re-encodes.  As such, they are not quite the quality of the German rips, but close.  The colors match almost perfectly and they generally look much better than the "hV" versions.

The snag is that ANH is missing the last 20 minutes or so of the film, and ROTJ is unavailable in this form.  I've found torrents for the original SkyTV rips, but they're from 2008 and have no seeders now (If only I could get on the HDBits tracker, I might find them there).  I've continued to search for these, but with no success.  As an alternative, I've located some alternate rips from Cinemax for ANH and ROTJ, again, in slightly lower quality, but still better than the hV.

The plan is to redo all patches and fix encoding errors using these new sources.

As of now, I'm about 25% done with ANH, 10% done with ROTS, and have the completed versions of ROTJ and AOTC from last year, which need to be redone using the new source material.  That leaves TESB and TPM with the most work remaining.  I hope to finish ROTS first, and publish both the DVD and Theatrical versions around the end of this month.   

Sure, I will post the theatrical version, eventually.  I am working on v.2 of these using some improved sources as backup to the German HDTV rips.  I'm doing a second pass looking for encoding errors in the stream, which is extremely tedious I might add.

I've moved ROTS to the front of the queue so it will be completed first, but I can only go through about 20,000 frames per sitting. That means it will take around 10 sessions to complete.  This should take between 2 weeks and a month for me to finish.

I also need to double check some things with the DTS mix, as some changes have been made to my decoding procedure recently.  That might take another 2-3 days.

That's really strange.  I'm seeing the same result in AVIsynth.  It works fine in every other program I've tried (Word, Photoshop, Aegisub).  Would it be possible to apply them as an .ass file in avisynth?

I did this in aegisub quickly, using the font I linked above.  No change in font width or height was necessary.  Kerning issues on the word "Your" aside, I'd say it's as close to a perfect match as you could get.

I know it's not Demi, but it looks like a closer match.

This looks like it might be a better match.

http://www.myfonts.com/fonts/adobe/franklin-gothic/condensed-37102/

Obviously, you might need to play around with the font width.

edit:

Found it free here:

http://ufonts.com/fonts/franklin-gothic-condensed.html

That thread Mentions "Franklin Gothic Medium Condensed"

I think FG "Demi Condensed" looks like a closer match.  (It also looks like what RU.08 is using, maybe ??)

edit:

Check that. :)  I can't decide whether it's demi condensed or medium condensed bold.  The "d" looks closer to the later.

What was your final determination on the font?

I don't think that link mentions any HD source for the missing shots from the theatrical version.  He was just using a 720p version of the 2004 DVD master.

Also, StarmanSkywalker has never made his recording available.  I wish we could help him liberate that footage :)

There has to be an HD version of this broadcast.  I hope it surfaces at some point.

Is the ORF1 rip 16:9?

Well, the iTunes extras cannot be downloaded, so I don't think the drm strippers would work.  Btw it does has the clapping shot included as well.  Send me a pm if you have any other insights.  I'll try WM Capture to see how it works.

I have the digital version already, but would need some advice for capturing the video.

Check your PMs. :)

It's definitely not in the deleted scenes.  I think it's one of the web docs.

From memory, I remember seeing the different mechanical hand shot in the digital release at the time.  I believe this was a known difference.

Very true, but one method will take much more effort. Obviously, it's the difference between Harmy's approach and Team Negative1's. I would appreciate either outcome.

Here's an untested, but probably working CP55 for only $129.99, buy it now.

http://www.ebay.com/itm/Dolby-Laboratories-CP55-Cinema-Sound-Processor-Receiver-Amplifier-/191640237424?pt=LH_DefaultDomain_0&hash=item2c9ea5f570

seems like a good price.

I don't think it was a wipe.  I'm pretty sure it was a regular hard cut.  It's just missing 30 or 40 frames around the transition.

Yep

Don't forget that the Tattooine garage scene had some added shots.  The shots before and after were trimmed as well, so they would need to be replaced as well.

Those two shots are available on the second disc of the DVD release (I don't remember exactly which video), but they were cropped to 16:9.  Those are the only shots that should affect the run time of the theatrical version.

btw, I would love to work on this if Team Negative1 has a scan lying around :)

I've already done sync's for ROTS and TPM.  I'm not sure why I never shared them.  I'll be able to post them to myspleen when I get home from vacation in the middle of August.

I have a full-on theatrical version of ROTS that I did based on my HDTV reconstruction and the Blu Ray which I can post as well.

clutchins, check your pm's. :)

I think we came to the same conclusion regarding the foobar plugin.  It's a little half-baked.  I never noticed any errors other than the surround channel clipping, but the output was definitely different than the winamp plugin, even for the L,C, and R channels.

I don't think the DTS specs are completely useless though.  They give a clue to the analogue processing in the system, which has to be accounted for when decoding.

Setting the surround speaker output 3dB lower, at 82dB, in the mixing studio, hypothetically means that they're encoded 3dB hot.  This also suggets that setting the mixing studio LFE at 88dB leads to the encoded signal being 3dB less than what should be heard when decoding.  

Unfortunately, the DTS Encoder Suite won't let you apply that 3dB cut at the encoding stage for DTS HD-MA.

Anyway, boosting by 3dB always sounded weird to me which is why I left it alone before.