logo Sign In

Cowclops

User Group
Members
Join date
27-Sep-2004
Last activity
28-Aug-2011
Posts
39

Post History

Post
#94377
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/94377/action/topic#94377
Time
I'm the guy that made BOTH versions. TR47 is just the one who sold a crapload of my first set. He was in fact one of the last people on ebay to buy it from me, so ironically the old set will probably be continued to be called the "TR47" set but he only has anything to do with the newer set. He bought a CLD-97 so I could re capture it with much better picture quality.
Post
#94368
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/94368/action/topic#94368
Time
Well I know I'm never doing it again after this one. I don't think its really possible to treat the LD material better than I have (with the version that is "sort of done" for now but I still need to make menus). All of this excitement over miraculously picture quality improving filters really mean they miracuously improve one aspect of the video at a complete loss of some other aspect. I say, the more you try to filter it, the worse its gonig to look compared to the laserdisc. All I did was tweak the LD's weird color balance, did some very very weak noise reduction to make it more compressible, and compressed it to mpeg2. Done and done.
Post
#94177
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/94177/action/topic#94177
Time
Well, if he was complaining why it doesn't look as good as a real DVD, then GIGO would be an apt suggestion. In this case, to claim DVD9 is somehow "better" you'd have to assume that DVD5 wasn't "good enough." Since the source is laserdisc and not film, it means its not as sharp. Since its not as sharp as film, its easier to compress. And since its easier to compress, it doesn't need DVD9. Bitrate viewer agrees.
Post
#93998
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/93998/action/topic#93998
Time
The 4 seconds of black is at the side change between Han saying he has to replace the negative power couplings and the inexplicably infamous Leia welding scene. My original version has all of the welding footage. The black is simply footage from the LD player that I forgot to delete. The new version, of course, does not have this issue.

As far as fffffff being the 47th person to ask whether it will be DVD5 or DVD9, I already explained in the most simple terms capable of using to illustrate the problem why additional video bitrate has no affect on it whatsoever. If you don't want to understand why it doesn't make a difference then at the very least you have to believe me when I say it DOES. NOT. MATTER.
Post
#93785
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/93785/action/topic#93785
Time
I've never heard of huffy. What sort of codec is it? It seems the general workflow on these discs could greatly benifit from some high end gear - hardware translate your analog video to 4:2:2 SDI and capture at 10 bit Uncompressed resolutions. Obviously you're going to have to have a large amount of storage, but you'd quickly factor out artifacts from these other formats.


Well, the "what is huffyuv" bit has been answered (lossless compression) but as I've stated already, there is no improvement between the current 10 bit phillips chip as you can find in the $45 Asus TVFM and supposed "professional" quality hardware. I work for Avid and I know what goes into this "professional hardware." What you're suggesting is what I'm already doing and it doesn't take a $20,000 Media Composer Adrenaline setup to do it. The main benefits of professional hardware are usually A) more inputs and outputs and B) easier to use "all in one" software packages and C) real time video effects. Since I only need one input, and using a number of freeware programs actually gives me more flexibility than working entirely within media composer, and i'm not adding any video effects, that kind of rules out any reason to spend more than $50 on a capture card.
Post
#93723
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/93723/action/topic#93723
Time
Yes. In DV (like what consumer camcorders like what I used to capture it the first time) compression, the color is sampled only once per every four luminance pixels. When one well saturated color sits on top of another (like the red of a lightsaber vs the blue of cloud city) then you can see the effect of only having one quarter of the color resolution. This is called 4:1:1 sampling, as the luminance is samples 4 times as often each color component. Mpeg2 on a DVD handles two color samples per every four luminance pixels horizontally, but it also only records color for every other line. Since it doesn't even have a color sample on every line, this is called 4:2:0. Due to the original capture being DV, it ends up being the worst of both wolrds on DVD (4:1:0) sampling. This is why the color seems to blur out.

This is also in fact a reason why I want to slap everyone bragging about their Canopus AVC-100 DV capture boxes. DV compression absolutely blows. Once I figured out how to work around the quirks of this $50 capture card, it ended up that it records as good quality as any professional equipment, and I can say this safely because I work at Avid. No matter how good the analog stage of the Canopus stuff is, it is compressing to a format that will destroy the quality. I use huffyuv lossless compression so there are no digital artifacts whatsoever until I run it through CCE.
Post
#93620
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/93620/action/topic#93620
Time
I have a dual-layer version of the video on my hard drive, but I'm not sure where everbody keeps getting this whole "its way too compressed" business. The Q Level (Which isn't an opinion of quality but an absolute statement of deviation from the source frame) is less than 3 in both versions. Most Hollywood DVDs aren't even compressed this perfectly (but thats because they have detail that makes it harder to compress anyway).

The dual layer version exists but as I've stressed at least 5 times before and can't believe i'm saying it a 6th time... THE COMPRESSION IS NOT THE ISSUE.
Post
#83016
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/83016/action/topic#83016
Time
For some reason unbeknownst to me, everybody seems to "prefer" the audio of the original version I made to the audio on the discs everyone else has made. Why, I don't know. Perhaps other people are clipping the audio during their recordings, perhaps they're just using cheap crappy sound cards. I think the main reason nobody is talking about the audio is because audio is the easy part.

The very fact that you'd suggest "generic" cables would be somehow detrimental to the signal would lead me to believe that you wouldn't be satisfied even if Tom Holman (the TH in THX) personally showed up and approved the work. While I don't USUALLY like to throw around credentials as some kind of excuse to say whatever crap I feel like, as an electrical engineer that is into audio and video stuff, I can definitely say that only extremely poorly designed cables will be detrimental to the sound. I am using non-specific cables (well, for the audio anyway). http://www.roger-russell.com/wire/wire.htm has some more info on amplifier->speaker wiring, and while not all of that directly applies to line level connections, the basic theory is the same. Anything of sufficient guage for the signal you're putting through it will be ok. Just don't run the cables next to the power wires and oyu're all set.

If anyone cares, I built my own s-video "cable" by way of building a couple s-video to dual f-connector breakout cables, and then I ran that over separate RG6 cabling, so thats how I hooked the LD player up.

Since you did ask the otherwise fair question "how exactly was it done" i'll just say that I hooked the analog outputs of the CLD-97 to the line in on my Turtle Beach Santa Cruz, recorded at 48khz, and made sure it was close to full level without actually clipping.

Basically, if you didn't find a problem with the audio in the first set, you won't find a problem with the audio in this set either. The analog in on a TBSC has to be at least as good as the analog in on a sony TR7000 digital8 camcorder. Even if it is a more "technically accurate" manner of recording, syncing a digital rip of the audio to the video would be far more of a pain in the ass and imperfect sync will be more of a bother than the quality lost due to going analog to digital to analog (since it has to be resampled ANYWAY, why not just let the ADC on the sound card do it in the process of recording?)

Last bit of info... I checked the brightness, contrast, and color of the set on a crappy directview CRT, and it seems to be as good as I hoped. The color saturation will be FAR better this time around. If you look at the original set, the title scroll in each movie is very pale whitish/yellowish. This time, every color that should be saturated IS saturated, rather than over or under. The title scrolls are unmistakably yellow. Red is red, green is green, blue is blue. Skin tones aren't weird purplish crap, but actual human looking. Lightsabers look considerably better than they do even in the SE DVD release. What it comes down to is, it looks good on my monitor (which I'd like to say is well calibrated, but I don't want to assume its 100% perfect) and it looks good on a TV, and i'll test it on my Panasonic AE700 projector that should be arriving this thursday. If it looks good on all 3 devices, then the main movie video is definitely done. If I don't like it, I'll recapture the entire movie again and tweak it. I don't have enough space to store all ~7 hours of video on my hard drive in huffyuv format, but capturing it again isn't that much of a problem compared to hte length of time it takes to compress it properly. Nonetheless, I like what I've seen so far, so I BET it should be good.

Thats all for now.

Post
#82863
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/82863/action/topic#82863
Time
I did consider the possibility of doing something like starwarslegacy.com but since the person on there is already doing such an excellent job, i'm better off preserving the original than creating yet another edit (even if its an edit people agree with).

I can take some screen shots, i'll probably do so this weekend. I figured out an easy way to rip out a lot of screen shots in a short amount of time, so there will be stuff to see soon enough.

I'm aware of some of the techniques being used by people, but frankly i'm not convinced. Most of the reason the first versoin came out so good is because I tweaked the video as LITTLE as possible... if you try to sharpen it you might just bring out artifacts that you don't really want to see, and I've already got a good noise reduction method. The reality is, you can't polish a turd, and if you try to you're probably just going to break it up. I'm still split as to whether the noise reduction i did is an "improvement" but what it comes down to is, its necessary. If I just captured the video and compressed to mpeg2 with no processing in between, it would be too hard to compress for even dual layer, and thus it would suffer. The noise reduction I'm doing (which is similiar to what you saw on the last set) allows me to do unheard of things like compressing the video at "only" 3000kbps. Its hard to explain without illustrating the full process, but when I'm done you guys should like it.

The two major "hardware" differences between what I used this time and last time, is that last time I used a Pioneer CLD-59 and a DV camcorder. Since then, I've figured out how to keep the video in sync with a cheap consumer TV capture card (10 bit phillips based asus tv fm). So now i don't need to use the camera, and i got a CLD-97 (TR47 funded that purchase). I know those who talk more often than they actually do something will probably scoff at the fact that I didn't use an HLD-X0 to do the recording, but these are all but impossible to come by and I'd have to question what I'd really gain by using such a player. If it was a perfect world, i'd just pay somebody to break into the Lucasfilm archives, steal me a print, and steal me a telecine machine and I'd be all set. Unfortunately, that isn't going to happen, so I have to draw the line somewhere of what is possible and what is unfeasible.
Post
#82843
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/82843/action/topic#82843
Time
I believe you missed what I said about bitrate viewer. Every mpeg encoded video, be it mpeg1 or mpeg2, has a number associated with each small sequence of frames. This number is called the "Q level" and describes, basically, the difference in quality between the source frame and the mpeg2 compressed frame. It ranges from 1 to 100 with 1 being best and 100 being worst. There is nothing to be "gained" by forcing the bitrate higher because there is nothing LOST at the bitrate I used for the single layer discs (about 3500kbps).

An average of less than 10 is usually considered acceptable (at least by people who author SVCDs, who usually accept less than perfect quality anyway). Most pro hollywood DVDs have an average ranging between 2 and 8, and usually a max of not more than 10. The Q level on all 3 of the discs I made now is about 2-3 average, 5 max. This means that it is LESS COMPRESSED than most high-quality professional DVDs. Consider extreme examples of compression difficulty... on one hand you could have a "video" comprised of a static black background and nothing else for 5 minutes, and on the other hand you could have a white-noise video. It might take over 15 megabits per second to get anything resembling white noise on a DVD in mpeg2, but the static black background will be "perfect" with very very little bitrate. Because of the softness of the LD source, and the noise reduction i've applied (I used a slightly different method this time around, so the "Grainy" clouds on Cloud City in ESB no longer look as such. This was actually a side effect of too much noise reduction rather than "not enough) keeps the bitrate required to maintain near-uncompressed quality VERY low. So, with "just" 3,000 kbits per second, its less compressed than some movies that require 5000 or 6000 kbits per second to appear acceptable.

If you take the original TR47 set as proof that I can be trusted on matters of video quality, believe me when I say that "quality gain" by going dual layer or using DD audio is, for all intents and purposes, ZILCH. Nonetheless, it costs me nothing to leave a higher bitrate version on my hard drive, so I did encode a dual-layer friendly version, but mostly just to keep around to show people that the difference just isn't there.
Post
#82836
Topic
<strong>The Cowclops Transfers (a.k.a. the PCM audio DVD's, Row47 set) Info and Feedback Thread</strong> (Released)
Link
https://originaltrilogy.com/post/id/82836/action/topic#82836
Time
Hi. I'm the one that made "The TR47" set and I'm the one making the new one. So far, the in movie video is finalized for the new set, and this week I'm finishing up the extras. One thing that TR47 said (he's sort of the executive producer of this little project... I sold him a set back when I was selling them on ebay in March 2003 and he liked it, so thats how he ended up on this) is that I would use the actual Definitive Collection chapter stops. I may not be doing this because I think it has an unnecessary number of stops per movie... something like 60 i believe. I'd prefer to do less than 30, as this actually makes it easier to navigate the movie (and frankly takes less effort to author). I might just do every other stop, or "the most important ones" but anyway, I didn't use the original stops for the so-called "TR47" set and it works. As long as they're not arbitrary, I can't imagine anyone complaining.

As far as using PCM audio on a single layer disc, there has been talk of the video quality suffering far too much. This is not the case. Because the Laserdiscs are so much softer than a true film->dvd transfer, it actually makes it very easy to compress. The average bitrate I'm using on the new set (which is actually a little less than I used on the old set because I'm making room for menus) still produces a maximum Q level (as reported by bitrateviewer) of LESS THAN 5, and an average of less than 3. While this doesn't imply that the final video quality is perfect, it DOES imply that the quality lost due to compression is extremely minimal. As a comparison, I popped in the theatrical version of Fellowship of the Ring, and the average Q was around 4 and it peaked all the way up to 8 for a max. The point is, because of the nature of the source, it is very compressable... so you lose just about nothing when comparing single layer to dual layer versions. For those who don't believe me, I did actually compress it twice and I will keep the "dual layer" version archived for future purposes if necessary, but right now, it just isn't necessary.

They will be anamorphic, so that people with 16:9 sets can watch them without black bars on the sides. The one thing that might be an issue is that while the subtitles in the Jabba the Hutt scenes aren't cut off because of the resizing process, they well almost certainly be in the overscan area on typical 16:9 CRT sets. This shouldn't be an issue for LCD/DLP projection and it won't be an issue on a PC monitor, however.

For those that think the color saturation in the original set is lacking, I've picked up a lot more knowledge on color processing in the last 2 years, so hopefully people should agree that the color on the new set far exceeds the quality of the previous set.

If you have any more questions, you can get me on "Cowclops" on AIM.

For those wondering about the previous version, you can find my crappy leftover "faq" site on the original version at http://www.cowclops.net/sw.htm

Feel free to continue referring to it as the TR47 set, as he is going to do most of the "promotion" work anyway, as you've probably already noticed.

Enjoy.