I’m finished with my preservation of the criterion collection laserdisc. Myspleen hurts though. I gotta go lay down.
Weird. Sounds like you are doing everything right. I’ll consider reposting it someday if you don’t have any success.
Thanks! The latest 720p jsc posts should be good. I know that at least a few people have nabbed it successfully. Are you using the par files to repair missing chunks? If you use sabnzbd it will pretty much do all of that automatically for you.
The US97SE trilogy boxset is going up now. I was going to mux everything into one single mkv but it was way complicated to make everything work right and in the end it still wouldn’t work the way I wanted it to so…single MKV’s. They include the cinema DTS and Laserdisc 5.1 audio files thanks to Capablemetal and his project. This and all other special edition laserdiscs will be synced to his material as reference.
I believe that all of Harmy’s projects are gout synced so it should be fine to mux audio from his projects with my laserdisc preservations.
The French set that I just posted this weekend is the THX pressing which doesn’t seem to be from a theatrical print. I probably don’t know what I’m talking about but I’ve heard it’s from the same master as the DC/gout with less severe DVNR. The next french sets I post will be from the pre-THX era and those should all be theatrical prints because I’ve seen cue marks on all three sets that I have. That includes SW,ESB, and ROTJ.
Thanks for the praise! Happycube is still updating his software and has begun to implement gpu acceleration which should provide real-time decoding. This will bring the 5 capture average into the realm of the feasible once I get a card that supports his gpu stuff.
Atm, I’m trying to finish the 2001 criterion LD so that I can upload it to the spleen. I basically decided to take a break from SW LD’s by doing this “little” project. It’s turned out to be a nightmare to sync to the blu-ray lol. It has what I can only describe as rolling cadence changes that are custom for each scene. I’ve no idea why they’ve done this but it’s insane to try and sort it out. I’m on side 5 of 6 and hopefully it’ll be finished around christmas time. After this, I’ll be back to the SW stuff 😃.
I am finished with the US97SE trilogy though. I’m just trying to sort out some of the audio stuff before making it available.
All of my 720p files are gout-synced so you shouldn’t have to trim much. It might be better to conform the audio files to the video frames instead of the other way around.
At the end of the scene where the green pig guards shove chewbacca into his cell and then it cuts to the frog outside jabba’s palace there is a couple duplicate frames there. Also, one of the scenes as the shuttle tydirium approaches endor has the same thing.
Thanks for explaining your process. I’ve never used neat video. I guess I’ll have to check it out.
EDIT: Neat Video is pretty neat actually. A little pricey but neat nonetheless.
I like the way it turned out. I see that you replaced the black missing frames with duplicates though. I guess neither way is wrong but it made me do a double take before I realized what was happening. What kind of denoising did you use if you don’t mind me asking btw? The results are pretty acceptable imo. If I think about the fact that this is a 30 year old laserdisc, it is pretty surprising how much you can pull out of it actually.
Sweet. I look forward to checking it out!
Yeah. I think specifying the frame rate is enough but I’ve heard that for some reason a lot of vc1 encodes are vfr and the convertfps just makes sure that instead of variable you get normal frame rate instead. Simply specifying the fps might be enough to make directshowsource work right with this file. I’ll definitely play with it a bit to see what works before beginning to sync again. I’ll probably start working on it again this weekend.
Links are good for 7 days.
The audio is from the laserdisc and the ld video is on the left with bd on the right.
Can anybody make heads or tails of what’s happened here? The opening credits also don’t match up very well although they are very close. The criterion ld also has cue marks here and there btw.
EDIT: I can’t tell for sure yet but I popped in the blu-ray and it isn’t sped up so it seems dgavcindex did some funky stuff while parsing? I believe I had honor pulldown flags checked. Maybe I should just use force film or something?
EDIT2: Well, it’s got to be directshowsource or dgavcindex messing it up or something. I can get it to slow down to the right speed but the frames are duplicates instead of progressive like on the LD. I guess the BD is maybe encoded with vfr for some scenes. In any case those frames should be progressive and not duplicates so I still haven’t quite figured out what I’m doing wrong. Or there is still a chance that the BD is messed up and missing those frames. Is there a bulletproof way to take the vc-1 file from the bd and import it into avisynth with the correct frame structure? Right now I ripped the BD with makemkv then converted the vc-1 to huffyuv/avi with avidemux and then load it with avisource in avisynth. I first tried directshowsource with convertfps=true but this was the method used in the video posted up above. I’m tired of messing with it today so I’ll do some googling and try it again tomorrow perhaps.
EDIT 3: Bam! I think I solved my problem. Sorry for all of the updates. I don’t often rip BD’s and this is the first vfr/vc-1 disc I’ve had to deal with. The problem is that you have to tell directshowsource how many fps and to convertfps so the line to load the file would be like this:
DirectShowSource(“X:\Movies\2001 ASO-BD\2001-ASO-BD.mkv”,fps=23.976,audio=false,convertfps=true)
Everything seems to be progressive now although the criterion LD does still have a few extra frames here and there. It should be a bit easier to sync things up now. I’ll post the sync’ed LD to the spleen when I’m finished.
Is there somebody that knows the differences between the uncut criterion cav LD and the blu-ray? I’m currently trying to sync the LD to the blu-ray but when I get to the point where floyd is sleeping on the shuttle with the pen floating by his hand the 2 no longer sync. The pen is maybe a different composite and the scene is cut differently. The LD contains several extra seconds of footage around here too. I think I can re-arrange the scenes and attain sync without losing too much. I’ll give it a further try tomorrow when I’m not so tired.
EDIT: Wow! The criterion LD contains so much extra footage/different cuts it’s going to be nearly impossible to sync them with the same exact footage. The blu-ray for example, has sections where they have doubled the speed by dropping every other frame whereas the criterion LD has the normal speed with all of the frames. I haven’t even gotten haywood to the orbiting space station yet either. Should I perservere and sync up everything that is possible, or just ivtc the criterion laserdisc as it is and put it on the spleen?
I’m thinking I will start by ivtc’ing and putting it on the spleen and then make another attempt at syncing it later.
EDIT2: I’m pretty pissed off about the blu-ray now actually lol. I used to hold it in high regard as one of the better looking titles on the format but what could the excuse be for dropping frames for entire sequences like that? That couldn’t possibly have been part of any print release. It seems like that would have been done post-telecine for whatever reason. Why did they need to speed up those portions?
These 2 frames are missing from the ntsc gout as far as I can tell:
(http://s25.postimg.cc/c1eba5zyn/esb_missingframe_1.png)
(http://s25.postimg.cc/l7wlxg573/esb_missingframe_2.png)
Taken from French THX pressing.
It is an average of 5 captures using temporalsoften. I use a function that motion compensates the captures to help with alignment issues. The pass3 clip will be the center/output clip. This uses avisynth 2.6 and the latest MVTools2 which has support for multi-threading, etc.
Function FPA(clip pass1,clip pass2,clip pass3, clip pass4, clip pass5)
{
interleave=interleave(pass1,pass2,pass3,pass4,pass5)
super=msuper(interleave,pel=2,mt=true)
multivectors=manalyse(super,multi=true,delta=2,mt=true)
multiclip=mcompensate(interleave,super,multivectors,mt=true,tr=2,center=true)
average=temporalsoften(multiclip,2,255,255,50,2).selectevery(25,12)
return(average)
}
I’m coming across these on the French pre-THX transfers:
[(http://s25.postimg.cc/3uqgcyibz/94fr_cuemark1.png)]
Does this mean they are from theatrical prints?
I’ve only bought 2 seagate hdd’s in my life and they both failed in under 2 years with normal usage. I’ve never had a WD fail despite years of operation.
That password doesn't work for me.
I'm working on the US 97SE boxset atm. Mostly because I've already got the 5 capture average/IVTC finished from awhile back. I've decided to use CapableMetal's project to integrate the cinema dts/laserdisc 5.1 mixes since he has posted everything I need including audio sync'ed video sources so that I can compare against my IVTC and insert/delete frames where needed.
I'd like to include all three films with the extra material into one mkv file. I'll have to figure out if it's possible to integrate all 3 films with chapters so that they can be separately demuxed. Also, I think some of the extra material is 29.97i video so I have to figure out if that stuff can be appended to 23.976fps material.
ROTJ pal is definitely a better source.
The only time I really notice it is around red lights and a little bit around the blue ones. As you say though, there is perhaps a way to eliminate that with the right comb filter.
The binary sunset is pretty bad though. That must be in the source, right?
Yeah. I contemplated breaking up the sections with different sizes and cropping them, etc., but as you say it varies wildly and takes up a bit of time. I figure as long as the existing black bars are set to black then it shouldn't really matter and makes it easier for me to just use one consistent sizing scheme.
So, I guess in the future I'll be resizing the P&S releases to 984x720 then adding 148 rows of black pixels on the left and right sides to reach a final resolution of 1280x720.
I don't know why but header checking has stopped working when I post to usenet. This means that the last couple of posts have a part missing from each file. This should be repairable with the par files but I don't know how to fix the header check problem. It's always worked flawlessly since I switched to jbinup. I've contacted my usenet provider to see what's up since I believe it is something on their side. Hopefully it resolves itself by the next time I need to post something.
Huh. I looked at the script and it seems I resized it to 860x720 before adding borders and it probably should have been 960x720. You were also right about the SWE-ESB-mitsubishi. It should be more around 2.59 AR instead of 3.12.
At least I wasn't that guy that crashed the probe into mars because of one too many zeros :). It looks like the other releases are sized properly but if anybody notices anything let me know. I will definitely spend some extra time checking that next time.
I will make note of and fix this. When all of these are finished I'm probably going to put together a mega-release sort of like schorman's laserdisc audio archive.
The fixes will all go into that release. I still have all of the raw files and scripts so making changes is possible without recapturing thankfully.
Haha. Atm, I have no plans of doing an encode without subtitles since I am knee deep in about 3-5 ld's at a time right now. When I'm done with the jsc trilogy I might put them on the spleen all at once without subs.
As far as AR I always use the same formula for widescreen LD's. I crop to the picture height (y). Then I do this to get the new picture height (z):
1280 x Y / 720 = Z
I noticed that about ESB as well but I figured I'd represent it like it was on the laserdisc.
The Fox LD is imo a massive achievement thanks to Darth Mallwalker's IVTC/gout-sync script. I rather like the way it looks as well considering it had some rot on side 1.
