Photoshop = Good for frame by frame restoration?

[Cyrax9]

I need your help guys and gals. I've got some old VHS tapes which I've (thankfully) been able to salvage/remaster and convert to DVD. Unfortunately VHS is VHS and source "flaws" are still slightly visible.

I'd like to go through roughly 40 episodes of a half hour TV show (about 22 minutes now that I've removed the commercials) that'll never air again and touch it up frame by frame.

I am a video production student but I've had no courses which deal with remastering and most of what I know that I haven't learned through college courses (that's anything related to DV) I've learned on my own time. I was wondering if Photoshop would be a good tool to use for frame by frame restoration? I know it has a "DV Preset" for graphics but I don't know if it's a good restoration tool.

Also, once I create new still images in photoshop (save as TIFF files?) what program should I use to render them as a video? I know they have to be shortened to one frame each (NTSC) but I'm not quite sure what to use. I'd prefer a Mac program if possible as my copy of photoshop is on my Mac, and although I usually edit on a PC my specs are less than spectacular. Any suggestions?

Finally, when I decide to mux the audio back with the video, and create a new file, what output format should I use? DV, DVD, or something else?

My main goal is to take these standard definition remastered tapes that are on DVD and remaster them "frame by frame" so that they'll look decent when upconverted to an HD format at a later date. I need an archival format to store them on prior to an HD Upconvert as well as one that would be useful after the HD Upconvert. Thankfully the conversions have come out very well but I think they'd need more touching up for an HD format. I'd appreciate any help you can give me for this project. Thank you in advance.

~Cyrax9~

P.S. I know frame by frame is tedious, so any tips to make it less so without reducing the video quality in the proccess are much appreciated as well.

[harley2ride]

Numerous video editing softwares offer filters which can be applied over multiple or all frames at once. There are also filters for VirtualDub, and AVISYNTH scripts which can clean up video. They won't do miracles, but they will clean up some minor stuff. If you are serious about it, then you need to look into a TBC device, or a S-VHS VCR with TBC and noise reduction to transfer your VHS tapes to PC.

[SLK001]

You've got to be kidding... That's 1,267,200 separate frames that you'll be hand editing (and THAT's assuming that the framerate is 24fps). Let's say that you can do one frame a minute. It will take you 21,120 hours, or 880 days, or 2.4 years to complete the task. Valuing your time at only $0.10 per hour, that means you will have wasted $2,112 to do this. I'd wait until the series is out on DVD before I'd even think of doing it by hand.

Better yet, do what harley2ride suggested and learn to use AviSynth and filters to accomplish this task. In my opinion, AviSynth has better filters than Virtualdub.

[redwudz]

I would have transferred them to the computer in DV format. VHS isn't that high of quality, so you are not likely to lose anything. Then, as mentioned above, you could work with then in VirtualDub and use the many available filters. A good site for VD filters is :http://neuron2.net/

Audio can be PCM if you want something universal for archiving. It would take a lot of HD space to store them that way, though. 13+Gb an hour for DV. DV is much easier to filter and edit than MPEG format. You can frameserve the edits and some filters out the encoder. AVISYNTH is well worth learning for that.

At 30fps you would need years to process all frames in PS.

I don't use my Mac for video, but I would still use DV with it. Look around the Mac forum a little and you should find some good tips.

[pijetro]

Like everybody says...
Capture with a decent VCR, use a TBC (Time base corrector), and perhaps throw in an Procamp, and then capture to DV...

If results are still not up to snuff, then most NLE's should be able to enhance your clips..

Forget about reusing any DVD footage..It's etched in stone. Any rerendering of that footage will result in quality loss...
I know it all seems like much, but see if a local video shop would bother renting you out any of these units...

Good luck!!!

[edDV]

... I've got some old VHS tapes ... I was wondering if Adobe Photoshop would be a good tool to use for frame by frame restoration?

I am a video production student ... My main goal is to take these standard definition remastered tapes that are on DVD and remaster them "frame by frame" so that they'll look decent when upconverted to an HD format
~Cyrax9~

Don't let them put you off on productivity. If you can figure a way to convert one minute of VHS to acceptable HDTV quality using frame by frame Photoshop techniques, you could be worth millions.

What makes you think this is possible and can you show us one restored frame?

[Cyrax9]

edDV, woah man -- let me finish getting evreything onto DVD first LOL.

The DVDs have been created by using a JVC S-VHS VCR --> JVC DVD Set-Top Recorder, Rip to PC, edit audio in SoundFORGE, and clip any extra commercials in MPEG II VCR. THen it's into Ulead and Burn to DVD.

This looks signifigantly better than VHS and the tapes lack any overscan thanks to the JVC S-VHS VCR's built-in TBC.

However, as I said VHS is VHS and that's not a format easily upconverted to HD.

My "solution" (using NTSC here) would be to repaint every frame in Photoshop CS; the problem I hit with "normal" filters is severe blurring whenever they're applied unless they're hardware. I coud effectively "re-draw" images in Photoshop and render each one as a frame of video. (No Audio) -- Then take the Audio which was fixed in SoundFORGE and drop it onto any ol' NLE and save it.

For the HD Upconvert I would simply use my 720x480 "hand-remastered" frames and run them through a proc amp into some form of HD Storage medium. Perhaps HDV, perhaps Blu-Ray or HD-DVD if they catch on. Since I'm redrawing the frames by hand (and perhaps re-rotoscoping effects) I could effectively deinterlace the video without quality loss, use the eyedropper to make sure the colors are accurate in P-Shop and "draw" progressive frames. Filters could be applied where needed and since this is unfortunately 4:3 video I would need to pillarbox it to create a 1080p image that's acceptable quality wise. In essence I'm "redrawing" each frame with P-Shop and removing the flaws using P-Shop -- perhaps at the pixel level in some areas.


Basically I've already got this footage in DVD Quality, or what's as close as possible from VHS Source. (Think of how "The Honeymooners" was remastered from DuMont Kinescopes for DVD.) Now I'm just redrawing the images and correcting the color frame by frame. It's by no means a small project and yes, I know it would probably be a good 5-10 year job if I did it at a fairly fast pace, but the chances of a DVD Release are zero (A certain company has all but flat out stated they won't release this series on DVD and if they did it would be signifigantly edited due to editing made after 9/11 thus making my effort to save this footage even more important to me as I despise the type of editing that it's been subjected too.) Basically I'd be touching up what's pretty much DVD quality; I'll attempt a few screen caps of a 1080i (1080p would require time for an experiment) Photoshop upconvert once I get my Macintosh fixed. Some bug in the BSD undercode has locked me out of all my admin functions (despite being the only system admin) and won't let me reinstall my printer. It doesn't help that it needs a new HDD and HD Video would effectively kill the machine in it's current state. I can't start until that laptop is fixed and that could be a month or two. In truth I'd really like to re-draw the problem areas first.

I realized that to get this into an HD Format would require frame by frame remastering, to get it into a DVD Format required a lot of nights in this forum and learning how a TBC works. I'd be much happier if the company that's overzealously edited a syndicated version of this show released unedited, or even lightly edited DVDs, but they haven't released one DVD of this seasonal series to date. They've released shows from 2003 and onward but never a full season. Anything prior to 2002 has all but been ignored and I'm finally about to snap as it's one of two shows I watch and I loathe the butchering it's been put through. I also have some friends willing to help me who are P-Shop experts and could probably teach me a few more tricks. Although I doubt I'd realistically do one frame a minute, if I could get this done in ten years I doubt a commercial release would be done by then. The company that owns the show doesn't seem interested in releasing it on DVD... even when petitioned into the ground. (PM me if you want to know who owns the show.)

My basic goal is to truely have a high quality standard def. copy to upconvert to HD -- It's not going to look like it came off of HDV tape, but it'll be better than just upconverting remastered SD signal that was run through a TBC and some other filters along the way. (Detailer/Proc amp etc.,etc.)

Sorry if I confused anyone. The source footage would be from remastered footage on a DVD -- I doubt the audio is going to get any better than it is and the video isn't too bad for standard def. The downside is that some falloff is still noticable as are analog artifacts and that's what I want to try and clean up.

[edDV]

While frame by frame hand restoration may be possible given enough time, the other approach is automated enhancement through computation. The colorization process commissioned by Ted Turner had 35mm film masters to work from and SD as the target.

Unfortunately VHS has only about 300x480 resolution and is very noisy. S-VHS can be as good as 500x480 but is equally noisy. Up scaling this to 1280x720 or 1440x1080 is asking alot.

"(Think of how "The Honeymooners" was remastered from DuMont Kinescopes for DVD.)"

Well those Kinescopes were 16mm black and white film which has resolution above HDTV. The Kinescope camera was however shooting a studio monitor which probably generated close to 5-6MHz or about 600x480. DVD is only 720x480 so the match wasn't as extreme.

[mattso]

Hand painting each FRAME? My God, your great great grandchildren will inherit this eternal project of yours!!!

Even if you ran a batch process with Photoshop 9 (say, levels, noise reduction, smart sharp, whatever) it'll still take forever.

Please look at avisynth, it has a tremendous amount of excellent filters.

[Snakebyte1]

Ya, I think you are underestimating the amount of time it takes to process each frame in Photoshop.

I did some rotoscoping work with Photoshop on a clip that ran about 30 secs, where I imported the AVI as a Filmstrip into Photoshop and then worked on each frame and it took several weeks!!!!

You might consider identifying any parts that are truely bad and clean those parts up, but unless you are in prison with time to kill you'll go insane long before you are finshed.....

[Cornucopia]

Coupla points:

1. The idea of Upconverting from (ultimately) Broacast/VHS source to (ultimately) HD, with multiple compression/resize stages thrown in the middle, is--sorry to be blunt--a joke. You may have more, finer pixels, but the content of each will still be of compressed VHS quality. Upconverting isn't magic, there are basic rules to resizing.

2. I'm not going to discount your patience or your skill--It's quite possible to do P-shop restoration given enough time. However, if you work on a frame-by-frame basis without special tools to do multi-frame comparisons (which is needed in restoration and rotoscope techniques--think onionskinning), you will very likely end up with that jumpy/wiggly/noisy animation/coloring that is common these days with AfterEffects, but was originally done with all cel animation before they had good interframe registration. BTW, this wreaks havoc with MPEG compression!

3. Not to sidetrack the issue, but there are IP (intellectual property) considerations here, too. It makes sense to have clean--maybe even processed and commercial-edited--backups. And it is worthwhile as a scientific endeavor to attempt restoration of excerpts of important footage, but "re-mastering" a series is really quite stepping on the toes of the true owners of this footage. That makes it difficult for me (and probably others here) to assist with tips, etc.

Scott

[tluxon]

What tools allow individual frames to be cut out of a video, edited and saved (I presume), then put back in as a replacement for the original frame? I didn't think such a thing was possible.

Tim

[SLK001]

What tools allow individual frames to be cut out of a video, edited and saved (I presume), then put back in as a replacement for the original frame? I didn't think such a thing was possible.

Tim

It's really easy. Use VirtualDub and save an image sequence to files.

Once you have spent the years required to clean them up one-by-one, then use Premiere or AviSynth to import the sequence as a video file. I find AviSynth's IMAGEREADER command to be quite easy to use.

[edDV]

Premiere and Photoshop can work on the same frame when Photoshop is accessed from within Premiere but this isn't optimized for rotoscoping.

Read this
http://www.digitalproducer.com/articles/viewarticle.jsp?id=33231

[Cyrax9]

Scott--

Let me reply to each of your three points first as they seem to address the key issues I'm facing.

1. I'm aware that I'm attempt to "draw blood from a stone" so to speak. The sad fact is that it's the only way this footage will ever make it into an HD Format. I do know that resizing SD to HD is usually ugly when taken from VHS Source. With that said the reason I wanted to use Photoshop was to go in and filter certain seagments piece by piece until they were fixed. Since some footage is recycled throughout the series several times I would only have to fix it once or twice; factor in one episode being a glorified clipshow and that reduces even more frames that would otherwise need signifigant remastering. There's a lot of "recycled" footage in this series so it's not as bad as it seems. I do need to figure out how to import AVIs as filmstrips using P-Shop CS on my Mac while I use my PC for audio. I've never been able to write scripts for AVISynth and when I tried to use VDub filters I got minor quality enhancement and excess blurring. (Something I'd hope to correct using P-Shop.) I know this isn't going to look like it came off of what are probably Panavision masters in some cases, but it's better than losing this footage, especially since it seems like every time the show is aired they cut more and more of it for their post-9/11 editing. (The audio has been modified several times as well.)

2. I was wondering if there was a good way to analyze one or two scenes at a time and compare the footage when re-rotoscoping and in some cases, to apply the restoration technique from one frame to a group of frames because it would be helpful. In some cases I really only want to filter a section of the video where as in others I'm going to want to go frame by frame through the entire scenes.

I'm familar with AfterEffects and it's annoying animation issues as I used it for a Chromakey last semester before I had (temporary) access to Avid for a day. I wish I had Avid sitting here because it was so much better than AfterEffects for keying.


3. I really have thought about the IP issues involved in this, in fact the intellectual property issues are the reason I didn't do this sooner. Although I am doing this only for myself, I've seen people who aren't as ethical on fansites and eBay while searching for legitimate memorabilia. The executive producer even posts on one of the fansites and is full aware of there lack of ethics. I do not condone piracy, but I do condone being able to keep a well-remastered back-up of footage that will never see a DVD Release let alone an HD Release. I'm working with my private collection, and really want to do this as an educational/personal experiment while I continue my TV Studies. I also want to preserve the original footage so that my children can enjoy it one day. I know there are heavily edited versions of the show occasionally broadcast on TV but no DVD Releases are planned, nor have any been done despite numerous requests from fans. Unfortunately this is the only way I'll be able to preserve the original footage, if only for my personal enjoyment. I would much rather the studio release DVDs, but the chances of that happening seem to be nonexistant.


I do appreciate evryone else's advice. I do think I might scale down the ammount of frame by frame work I decide to do and leave that to only the really disasterous frames but I do want to attempt a decent upconvert. I know this is drawing blood from a stone so to speak and that this will seem more like "enhanced definition" than "high-definition" in the long run, but I would like to try at least.

[edDV]

Why are you doing this? As an intellectual exercise?

Better you learn realworld video production techniques while you are in school and try for the Nobel Prize after you have mastered the basics. There isn't enough detail in a vhs tape to create an acceptable 13x pixel (720p) or 30x pixel (1080p) expansion using existing technology. It may well be possible in the future but the technique will require detail knowledge of complex digital filtering.

You need to decide whether you are a video production major or should transfer to a math or engineering career. I'm not trying to put you off. Many discoveries are made when attempting the "impossible".

Video production leverages other's engineering work to enable production of cutting art films or videos. Restoration falls into a different category and is a technical specialty.

[Barnabas]

Cyrax9,

Have you tried the new VirtualDub filter "neat video"? It's $100.00 but has a demo version you can try. I tried it on some really dark "Christmas Lights" stuff and it was able to REALLY clean up the picture.

[Cornucopia]

I've done a number of "restorations" before including creating a few trade secret techniques(Wilson Markle told me that the only reason he didn't hire me was because I wasn't a Canadian citizen), so I know that you're going about this the wrong way.

#1 Have you asked the company that produces this series if you can do this? If they say yes (hopefully in writing), then you have alot more people on your side. If they say no, then maybe you should leave well enough alone. Work on your own material...

#2 If your ultimate goal is HD, then the sooner you get it at HD rez the better. Assuming VHS is truly the best that is available, you path should be more like this:

VHS tapes (hopefully coupla same episodes on different recordings to be able to "average" out the noise) -->Hi quality SVHS deck with TBC --> 601-compliant capture card --> Uncompressed or losslessly compressed SD master AVI (YUV 4:2:2 or 4:4:4 or RGB 720x480/486 @ 29.97, maybe >8bit/colorprimary) --> Highest Quality multipass Motion-compensated DeInterlacing or ITVC, if applicable (also saved uncompressed) --> Software Up-rez (precise Lanczos or better algorithm) to 1080 or 720 (also uncompressed).
THEN, use P-shop or AfterEffects or Vdub etc to clean/paint/restore the pixels. Again, save uncompressed.
Filesizes--HUGE! Yup, you bet. That's the way the industry really works when they are doing something right and want quality.

#3 re:AfterEffects. If you can't get a razorsharp and smoothly clean key/matte then you don't really know what you're doing with AE. Take a class or two. I work on AVID's almost daily, and while their new MC/Adrenaline keys/mattes are much better than they used to be (and are way better than most), AE still has them beat when done correctly.

#4 If you just want to "preserve" the footage. Digitize it (similar to above) and save the uncompressed AVI's to DL DVD±R (or BluRay, etc when available). Preserved.

Sorry, just trying to take a more pragmatic approach...

Scott

[edDV]

#3 re:AfterEffects. If you can't get a razorsharp and smoothly clean key/matte then you don't really know what you're doing with AE. Take a class or two. I work on AVID's almost daily, and while their new MC/Adrenaline keys/mattes are much better than they used to be (and are way better than most), AE still has them beat when done correctly.

Is there any history of this being successful from a VHS source?

[edDV]

#4 If you just want to "preserve" the footage. Digitize it (similar to above) and save the uncompressed AVI's to DL DVD±R (or BluRay, etc when available). Preserved.

Sorry, just trying to take a more pragmatic approach...

Scott

I agree with that. Preserve what is there and fight another day. Technology will improve but the VHS copy will degrade.

[vhelp]

@ Cyrax9

From a restoration point of view, your dvd recorder method
is short-changing your efforts.

I realize that this method may be your best bet, and I will
understand if this is your only choice

But a few words, if I may..

MPEG, as you know, is YUV color space. But, it is aprox one
layer under sub-sampling. That is, when your dvd recorder
gets your VCR's output, (which is YUV 4:2:0 sub-sampling) at
best, your dvd recorder will UPsample the source to YUV 4:2:2
(this is my belief)

Now, assuming that is true, and your dvd recorder has a YUV 4:2:2
UPsampled source, then internally, through its processes to MPEG-2,
it re-sub-samples it back down to YUV 4:2:0 color space, to a
final MPEG-2 source video file.

So, your setup situation seems to flow like this:

** VCR[out yuv420] -> dvd_recorder[UPsample yuv422 >> sub-sample yuv420]

My point that I am trying to make here, is that if quality is your
main objective, then shouldn't you be using a setup for a finished
YUV 4:2:2 color space ??

With a YUV 4:2:2 color space source, [ie, Huffy set to YUV 4:2:2] you have
a better chance of a final archival format, (for further process)

Now, considering the fact that you are going to be working with some
form of filter(s) for these videos, and assuming that your goal is to
stay inside your mentioned tools (ie, P.Shop) then you have a better
chance of consistancy with working with an YUV422 -> RGB for P.Shop
filtering around with, than you would, if your final source videos are
in YUV 4:2:0 color space (converted by) YUV420 -> RGB for P.shop.

Thus.. what I'm saying here, is this:

bad or not-so-good:
YUV420 -> RGB

GOOD:
YUV422 -> RGB

Then..

FINAL:
YUV422 -> RGB -> [rgb]filter/process -> (RGB->YUV422) -> [yuv422]New_Container

The above is under the assumption that your tools are RGB color space,
as in the example, P.Shop software -- your preference.

But, on the other hand, AVIsynth and it's arson of filter plugins
can keep the color space conversion(s) to a minimum, if any.
And, assuming that you are capable of re-configuring your Video Transfer
setup, you could configure a YUV 4:2:2 color space format capture,
with the proper codec. One popular codec is the HuffyUV or something.
This one is a Windows version. But, since you work inside MAC pc's,
you may have a problem, because I don't think there is support for
this codec on a MAC enviornment. But, surely, there must be a codec
similar for this type of machine

In any case..

In AVIsynth, there are a number of filter plugins that work under the
YUV422 color space. You only need to have a little bit more patiance
and strive harder to learn and understand them, and use them to your
advantage in your source.

An AVIsynth 'For Example'..

if you have a group of pictures (images) in your video source that:

A - [Group A] - need to deal with 'background' noise..
B - [Group B] - need to sharpen blurry edges..
C - [Group C] - need to brighten bad drop outs..

EDIT: - revised..

then you could write an AVIsynth (AVS) script, call it,
"vhs_transfer_ep27.avs"

 # vhs_transfer_ep27.avs
 # EP 27 requires three segments to filterize.
 # sharpen; blur; and levels;

 clip1=TRIM(0,1200).sharpen(1)
 clip2=TRIM(1201,4500).blur(1.58)
 clip3=TRIM(4501,7980).Levels(16, 1, 255, 0, 255)

 return clip1++clip2++clip3

(Note, the above will work as it is, but realize that it is ment to
illustrate how you can string several filters inside one video source
and process those segments accordingly)

The sharpen() function does what it says.
blur() will blur (lets say, denoise here) at its max value, 1.58
levels() in this case will brighten this area of video

The above are all Internal filters of AVIsynth.

My suggestion would be to do some research on all the various AVIsynth
filter plugins that are available, and see what each one does. You
may find various ones that your source(s) will benefit from.

About the only tricky part of filter plugin usage is knowing which
filter to use inside a color space.. ie, YUV422.YUY2 vs. YUV420.YV12

But, just imagine how you could apply this to your library of video
archives and process (filterize) them to your needs.


RESOURCES:

** Internal filters - listing of all (by catagory)
** AVIsynthManual - listing of all keywords/commands/filters etc.
** External filters - listing (maintained by warpenterprises doom9 member)

-vhelp 3750

[Cyrax9]

I've done a number of "restorations" before including creating a few trade secret techniques(Wilson Markle told me that the only reason he didn't hire me was because I wasn't a Canadian citizen), so I know that you're going about this the wrong way.

#1 Have you asked the company that produces this series if you can do this? If they say yes (hopefully in writing), then you have alot more people on your side. If they say no, then maybe you should leave well enough alone. Work on your own material...

#2 If your ultimate goal is HD, then the sooner you get it at HD rez the better. Assuming VHS is truly the best that is available, you path should be more like this:

VHS tapes (hopefully coupla same episodes on different recordings to be able to "average" out the noise) -->Hi quality SVHS deck with TBC --> 601-compliant capture card --> Uncompressed or losslessly compressed SD master AVI (YUV 4:2:2 or 4:4:4 or RGB 720x480/486 @ 29.97, maybe >8bit/colorprimary) --> Highest Quality multipass Motion-compensated DeInterlacing or ITVC, if applicable (also saved uncompressed) --> Software Up-rez (precise Lanczos or better algorithm) to 1080 or 720 (also uncompressed).
THEN, use P-shop or AfterEffects or Vdub etc to clean/paint/restore the pixels. Again, save uncompressed.
Filesizes--HUGE! Yup, you bet. That's the way the industry really works when they are doing something right and want quality.

#3 re:AfterEffects. If you can't get a razorsharp and smoothly clean key/matte then you don't really know what you're doing with AE. Take a class or two. I work on AVID's almost daily, and while their new MC/Adrenaline keys/mattes are much better than they used to be (and are way better than most), AE still has them beat when done correctly.

#4 If you just want to "preserve" the footage. Digitize it (similar to above) and save the uncompressed AVI's to DL DVD±R (or BluRay, etc when available). Preserved.

Sorry, just trying to take a more pragmatic approach...

Scott

Scott--

I'm going to try and answer each of these as best I can. Again, I appreciate your help.

1. This one's tricky to answer. Officially the company that owns the footage won't say anything if/when they reply to fans requests; the seem to modify the same form letter for requests for Sub-Lisences, Official DVD Distribution, and airing the footage unedited. Unoffically however, two different executive producers have commented on the "excellent talent" of the fans and "how well they preserve the show."

We know that the EP knows that the show is freely traded on said fansite and in an interview he was quoted as saying "Everyone involved with the show, and I do mean everyone, reads -site name withheld for privacy reasons -- PM me for a link please." Thus, the company isn't in the dark about what some fans are doing and has been quite accepting of what the fans are doing. As long as nobody is profiteering they seem to be fine with fans "playing" with their work.

This is what's confusing, some of us would really like to get a green-light hand written on paper and not just typed by the EP onto a fan-run message board. One person who's given me some tips had ties to the cast of the season I'm concerned with (he was actually invited to the wrap party) and star talent is known to post at said message board as well. Several of us (fans) have tried to get a concrete answer in writing but it seems like all we've got is the Executive Producer's typed word. I'm not sure if we can take it or not but it's there. As long as the fans use their footage and do the legwork the company doesn't seem to care. For the few seasons that do have a few DVD Releases (nothing complete mind you) a few of the company's employees have said that as long as nobody rips off their official DVDs they don't really care what's done with the footage. I'd prefer a hand-written "contract" if you will, but I'll take what I can get. This is what confuses me on the legalities, the EP and basically everyone involved seems to only comment that they "don't care" (also commented at several comic conventions in 2005) but the EP runs a branch of a much larger company and nobody can get anything out of them.

Numerous people have asked if the website was ever sent any letters from a legal dept. over the years and the answer is always "no" so it seems like there's a sort of grey area here. The crew "doesn't care" what the fans do as long as the few shows offically released aren't pirated off of the offical DVDs but they haven't put anything in writing which is what's left me fairly reluctant to ask for help on this project.

2. I figured there had to be an easier way, although my Mac has a puny 60GB hard drive for data and video and while my PC could dedicate roughly 100GB it's ridiculously slow and might not be able to handle the load. (I had 100GB to work with in the school as well.) I do have a 200GB external drive that I would connect to the Mac to save works in progress but that'd be the best I could do.

Unfortunately I only have one set of Tapes that are truely decent, I have an official Video CD Set of the edited footage (Non-US and PAL mind you) and a few episodes recorded from DBS onto a Set-top DVD Writer so I can judge what scenes would really need extra work. The downside is that the show was produced in 2001 and edited after 9/11 so some scenes are either missing or altered and it shows. I can do a rough comparison though. (Had I known this was going to happen I would have bought a DVD Recorder in 2001 and made multiple copies of the show.)

My capture set-up was going to probably be JVC HR-S9911U --> Hardware Filters --> Pass through Sony PDX10 to "capture" into iMovie using the DV Standard (IIRC that's YUV 4:2:2 Lossless) --> 720x480 @29.97Fps Interlaced --> Copy (or Move) Audio to PC and remaster audio --> drop new audio into iMovie --> Render Highest possible quality DV Tape Output --> Output to Panasonic AY-DVM83MQ MiniDV Tapes via PDX10 to save new Standard Definition Masters w/PCM WAV Audio in DVCAM Mode. (55 Minutes to a tape or two episodes.)

After I had remastered in Standard Definition I intended to take a break and put some money aside to invest in a better/larger Desktop unit that would obviously have a much larger set of Hard Drives (300GB?) and more RAM than my Laptop does. Considering I'm merely using iMovie as a stop-gap and software filters have actually WORSENED the quality of the video in VirtualDub and AVISynth (the one time I got it to work) I'm trying to avoid them for the SD Remastered tapes.

Once I had a new machine I intended to subsequently go MiniDV --> FCP HD Uncompressed/lossless capture --> P-Shop to clean/paint/restore various pixels in truely bad scenes --> Output RAW video through FCP HD either as an AVI or onto HDV compliant DV Tapes through a HD-compliant NLE Editing Deck using HD Tapes. Save tapes until this format war is solved and transfer to whichever format wins.

In essence this is a three step project.

Step 1: Remaster Standard Def. footage from VHS to MiniDV.
Step 2: Upconvert to Hi-Def and output to an HD Tape format.
Step 3: Copy to a Hi-Def Disc format once one has proven itself to be reliable.

3. I did take a class on AE. The problem (so says the Proffessor) was the age of the machines we were using and the version of AE we used. Don't get me wrong -- AE has an excellent key but I was using machines that weren't configured to let us use the NTSC Monitors next to the LCDs to double-check my keying. When I used AVID the machines were configured to work with the NTSC monitors. I'm sure AE would have been better if we had A) Newer Machines (I missed that by one year mind you -- these were no newer than 2001 and hadn't been upgraded; it showed) and B) The NTSC Monitor that was configured for AVID was configured for AE. Basically the machines had two boot settings. "GENERAL" for Premiere Pro 1.5 and AE and "AVID" for Avid. Since the two settings were to resolve compatibility issues and our IT dept. is a mess I could only use the NTSC Monitor when we finally were allowed to use Avid. At home I just use my regular/newest TV as my secondary monitor since I know the color is adjusted correctly.

4. LOL that was step one of three in my plan. Preserve the footage on MiniDV Tapes (HDD Space is at a premium here due to the ammount of data I have on this beast.) after remastering. I already know that using only the TBC in the JVC HR-S9911U -- no standalone -- totally rips out and restores the overscan on these and they're in very good condition for VHS. Upconvert to HD/HD Tapes once I have a new Desktop. Transfer to whichever Hi-Def Disc format takes off if any.

So basically my steps are:
-Preserve
-Upconvert
-Transfer

(And I just realized I have a ridiculous acronymn now for this project -- PUT HD. Preserve, Upconvert, Transfer to HD Discs.)

Thank you for your help Scott, it's something I really appreciate. In truth just reading these forums has helped me with my classes. My Professor still can't believe I was able to name the three Standard Definition Broadcasting formats. (NTSC/PAL/SECAM) on the first day of class last Spring.

Why are you doing this? As an intellectual exercise?

Better you learn realworld video production techniques while you are in school and try for the Nobel Prize after you have mastered the basics. There isn't enough detail in a vhs tape to create an acceptable 13x pixel (720p) or 30x pixel (1080p) expansion using existing technology. It may well be possible in the future but the technique will require detail knowledge of complex digital filtering.

You need to decide whether you are a video production major or should transfer to a math or engineering career. I'm not trying to put you off. Many discoveries are made when attempting the "impossible".

Video production leverages other's engineering work to enable production of cutting art films or videos. Restoration falls into a different category and is a technical specialty.

edDV--

I'm doing this as part intellectual exercise/personal challange, part stress relief (yes, this is stress relief for me,) and part preservation of unedited footage. Censorship editing; especially the overzealous editing this series was subjected too after 9/11 truely irks me and it hurts to watch what they've done to the series. Many of the most serious and intense lines were removed as were some of the visual effects departments best scenes. The audio editing hurt the plot and the lack of DVDs, edited or otherwise means that this is all I've got. The edited Video CDs I imported a few years ago (PAL format) are actually worse than my VHS Source footage on a PAL Monitor. That's saying something.

I'm also a bit of a techie so getting "under the hood" of my videos and taking on a hard project is something I enjoy.

I'm taking production courses precisely so that I can learn the basics. I'm one of the people at the top of my class so I do help a lot of other students out when I finish a project and my Professor needs someone to help him out. (I can think of at least two instances in the same day where he asked me to show some of my classmates how to work the capture decks we have and do some basic editing techniques. It's fun though because I also get to learn some extra information that most students don't since I'm a bit ahead of my classmates at times.)

As I said before I know this is a "drawing blood from a stone" project and it's really a spare time thing, but it's something I'd like to at least attempt.

The truth is I was looking into computer science majors when I first started college, but I just don't have a mind for math which is admittedly the only field I've ever truely struggled in. I can do several physics projects and explain why they work in detail, but I can't follow the math behind them. I have struggled in every math course since the sixth grade and it's the only thing that's kept me from having straight "A's" throughout High School. (I had one "B" every marking period and it was always in math. It fell to a "C" when I was hospitalized for a week prior to a test for a 24-hour EEG and that really annoyed me for a long time.) I also considered engineering and art majors and Video Production seemed to be the major that pulled everything I enjoyed together.

My Professor (who's also my Academic Advisor) for TV I, TV II, and in the Fall TV III and TV IV in the Spring is really pushing me to become an editor since he's seen my work and knows I enjoy editing and he was able to use some of my projects as examples for certain types of editing throughout the semester. In truth, I'd like to work in multiple fields eventually.

Science & Technology are a hobby of mind (my cousin's an Astro-Physicist so it seems to run in the family) and since Restoration combines technology with video production it's something that I enjoy doing in my spare time. Likewise I'm considering a Minor in theater when I enter my Junior year simply because whenever my video production classes need someone to be "talent" I'm always chosen since I'm willing to do it. My Professor has actually had to tell people that they can't use me as talent so that I can learn certain technical and artistic aspects of video production.

I admit it, I enjoy Science & Technology, as well as Art and History. For me, Video lets me work with all of these fields and while I've yet to be able to do anything with History yet, my final project for TV II was a documentary of sorts.

This project is just something I want to do in my spare time. It's not intended to be groundbreaking. If it is, that's great, and I'll have gained some extra knowledge in the process. If it isn't, that's fine too. It's just something to keep me busy and from what I'm understanding it should fufill that need just fine. I admit it, I could easily go back to college two more times and get two more degrees if I had the time and the money.

I agree with that. Preserve what is there and fight another day. Technology will improve but the VHS copy will degrade.

Ironically enough, this is why my first step was to remaster and preserve the footage on high quality MiniDV Tapes in DVCAM Mode. I don't know how much longer these tapes will last and from what I'm understanding, if I take care of the MiniDV Tapes (rewind/fast-forward every six months, only use two brands, stick with my pro-level Panasonic tapes that are known for reliablity etc.,etc.) they'll last me quite a few years and if the tape begins to wear I can do a lossless firewire dub from camcorder to camcorder. Using my PUT (Preserve, Upconvert, Transfer) method this is step one. I think I can do this quite easily once I get my hands on a few more pieces of equipment. Upconverting and Transferring to a viable HD Format are something I'm just trying to gather information on right now; Blu-Ray and HD-DVD might not even catch on. I might be better off using HDV eventually or even the old D-VHS Standard.

My primary goal right now is to preserve these in a format suitable for archival until more HD Upconverts have been done successfully and the HD world isn't as insane as it is right now.


I was thinking of Photoshop for use on the Preserved MiniDV tapes after capture, but from what I'm understanding it'd be best to preserve these for now and worry about future technology in the future.

@vhelp

I admit the DVD Recorder wasn't my first choice for capture. I have two PCs and one is perpetually down since I'm troubleshooting Linux and a friened is recompiling parts of the kernal on it. (Dual-boot machine mind you.) This machine has the capture card from hell -- it drops more frames than it captures and I wish I had known that back in 2002 when I bought Pinnacle Studio 7 for the M-JEPG Capture card it came with. It only handles 640x480 or lower and I'm one step from just ripping it out of my PC.

My next idea is grabbing the JVC HR-S9911U, some hardware filters, and passing the video through my PDX10 and into my laptop. I've experimented with a friends hardware fitler set-up and the video is fine for SD leaving only an audio issue which SoundFORGE fixes quite nicely. I'm going to try and mirror his set-up except I'm going to use my PDX10 to pass the video into my Mac rather than record to a DVD first. Then I'll edit the audio on my PC and drop it back onto my Mac. Render the video and output onto the PDX10. It's lossless to the best of my knowledge (iMovie for simplicity's sake here) and would preserve the SD Footage. Then I could worry about HD later.

I should mention my Mac is a Laptop, it's in need of some repair (error in the OSX undercoding) and a few upgrades (need 1GB of RAM and a larger hard drive) so it's not your state of the art G5. It does however perform above any of the PCs I own and I've yet to hit a file I can't get to play back on it so it's reliable. I'll be getting a new computer in a year or so anyway which means that I'd like to have the tapes ready by then. That way I can build the machine to handle video and try to finish this task.

Sorry for the long post everyone, I just wanted to address several things at once. Thank you all for your help. I'm just wondering if using my PDX10 to pass the video through to my Mac is a good idea. I'd like to avoid spending too much more money on Standard Def equipment sans the S-VHS VCR. I'm only using iMovie since this isn't something that's going to require excess effects the way a new production would. I'll use FCP HD when I get it with a newer Mac. From what I understand iMovie uses a lossless DV Codec which means that I could preserve in iMovie and then save to MiniDV.

Thanks again everyone.

~Cyrax9~

[vhelp]

and I wish I had known that back in 2002 when I bought Pinnacle Studio 7
for the M-JEPG Capture card it came with. It only handles 640x480

Oh. You have the DC10+ (aka, DC10Plus) card. Actually, it's a great card.
Another one of my favorites. It produces ZERO line noise. All my other
analog capture cards exhibits line noise from to some degree, to the point
that I gave up on them (temporarily) and went with the ADVC-100, which
also produces ZERO line noise.. which I often refer as the Holy Grail
of current capture cards.. mainly on account of the clean source and no
audio issues, and no drop in frames. But, the DC10+ card is also good
card. Just the resolution was too low. 640 x 480 is too low, IMHO.
Only the PAL version were 720 x 576 resolution.. lucky for them.

From what I understand iMovie uses a lossless DV Codec which means that
I could preserve in iMovie and then save to MiniDV.

Well, DV is not lossless. It is lossy.

The lossless part is where you are transfering from DV CAM / device to
the pc. The *transfer* via firewire is what's lossless

Oh, and don't forget, DV is YUV color space, with 4:1:1 sub-sampling.

Only PAL dv is 4:2:0 sub-sampling. I think the stream type layout is
using YV12, but I could be wrong, maybe.

-vhelp 3754

[Cornucopia]

Yes, and while DVCam is arguably more "pro" than standard DV format, they're both still 4:1:1. This means some chroma info is getting lost, even from the VHS masters. NB: As analog decks, (S)VHS vcrs output as full a chroma as their bandwidth and comb filtering allows. This is equivalent to 4:2:2 (maybe even 4:4:4)
For you to retain all the chroma, which I would consider quite important when doing "restoration" work, you'd want to capture and save in a format that used 4:2:2, such as DigiBeta or DVCPro50.

Suggestion:
If there ARE some seasons (albeit "cut up/censored") that are available via DVD/commercialVHS--Buy those and "add back in" the missing footage. Therefore, do ONLY your processing on the footage that gets cut. That will save a great percentage of the space requirements.

I think there are a number of filter functions available for AVISynth and Vdub that DON'T blur the way you fear, you just have to learn them well/play around with them. Then leave P-shop for the few static adjustments that can't be done on the whole stream.

Good luck...

Scott

[edDV]

Yes, and while DVCam is arguably more "pro" than standard DV format, they're both still 4:1:1. This means some chroma info is getting lost, even from the VHS masters. NB: As analog decks, (S)VHS vcrs output as full a chroma as their bandwidth and comb filtering allows. This is equivalent to 4:2:2 (maybe even 4:4:4)
For you to retain all the chroma, which I would consider quite important when doing "restoration" work, you'd want to capture and save in a format that used 4:2:2, such as DigiBeta or DVCPro50.

I just can't buy that. We are into sampling theory here and even 4:1:1 is oversampling VHS chroma. NTSC DV 4:1:1 (like DVCPro NTSC and PAL) maintains pixel spatial overlap with the first of the four luminance samples. This eliminates generation losses associated with 4:2:0 (as used for PAL DV and DVD MPeg2).

VHS, S-VHS, 8mm and Hi8 VCRs all use the color under analog modulation technique to record color components U and V. Broadcast quality composite NTSC (PAL is similar) has color components quadrature modulared around the 3.58 MHz subcarrier with maximum U and V bandwidth of ~1MHz each*. The VCR will extract some of this bandwidth during filtering prior to the recording process. The 3.58 MHz subcarrier is then downconverted to 629 KHz and in the process U and V bandwidth is reduced to ~500 KHz each. During playback the 629 KHz subcarrier is multiplied back to 3.58 MHz but the bandwidth of U and V remains at 500KHz, a very low value.

Ref: http://catalogs.infocommiq.com/AVCAT/images/documents/pdfs/TT189%20-%204611.pdf

Sampling theory suggests that sampling the analog signal at twice the analog bandwidth is sufficient to recover all bandwidth (detail) present. In this case, the Nyquist sampling limit would be 1 MHz for U and V. 4:1:1 samples luminance at 13.5 MHz and chroma at 3.375 MHz. This oversamples the color 3 times and is sufficient to extract all color detail present in VHS tape. 4:2:2 or 4:4:4 sampling would get no better results.

Using a S-VHS VCR for playback will allow separate output of the Y (luminance) and C (chrominance) from the tape over a S-Video connection thus avoiding further losses in a downstream color separation filter in the capture device.

The fact that 4:1:1 keeps the U and V samples spatially aligned with the first of the 4 luminance samples makes lossless generation replication and digital chroma filtering possible.

In case someone was wondering, DV format oversamples VHS luminance about 2.3x (13.5MHz.). Intraframe DCT compression similar to JPG is used to compress bitrate ~5x to 25Mb/s. If you wanted to avoid this compression, you would capture to YUV uncompressed or Huffyuv. The slighty higher picture quality from VHS would be due to lack of compression, not 4:2:2 sampling.

BTW: Digital8, DV, DVCAM and DVCPro all have the same recording signal spec. They differ only for tape formulations and mechanical tape issues such as track width and spacing.


* original NTSC design used a bandwidth differentiated form of U and V called I and Q (one using 0.5 MHz and the other 1.5 MHz bandwidth). The idea was to match the eye's color response. For a variety of reasons, this proved impractical and by the 60's the 1MHz UV scheme became common for NTSC and was formally adopted for PAL.

[edDV]

@ Cyrax9

MPEG, as you know, is YUV color space. But, it is aprox one
layer under sub-sampling. That is, when your dvd recorder
gets your VCR's output, (which is YUV 4:2:0 sub-sampling) at
best, your dvd recorder will UPsample the source to YUV 4:2:2
(this is my belief)

VCR S-Video output is analog luminance (Y) + color modulated around a 3.58MHz subcarrier (C). Where do you get the 4:2:0? Are you referring to an internal TBC?

Analog equivalent bandwidth terms would place VHS output around 1.8 : 0.3 : 0.3

VCR's output analog so I would think this is true even if there is an internal TBC.

Now, assuming that is true, and your dvd recorder has a YUV 4:2:2
UPsampled source, then internally, through its processes to MPEG-2,
it re-sub-samples it back down to YUV 4:2:0 color space, to a
final MPEG-2 source video file.

So, your setup situation seems to flow like this:

** VCR[out yuv420] -> dvd_recorder[UPsample yuv422 >> sub-sample yuv420]

My point that I am trying to make here, is that if quality is your
main objective, then shouldn't you be using a setup for a finished
YUV 4:2:2 color space ??

With a YUV 4:2:2 color space source, [ie, Huffy set to YUV 4:2:2] you have
a better chance of a final archival format, (for further process)

Now, considering the fact that you are going to be working with some
form of filter(s) for these videos, and assuming that your goal is to
stay inside your mentioned tools (ie, P.Shop) then you have a better
chance of consistancy with working with an YUV422 -> RGB for P.Shop
filtering around with, than you would, if your final source videos are
in YUV 4:2:0 color space (converted by) YUV420 -> RGB for P.shop.

Thus.. what I'm saying here, is this:

bad or not-so-good:
YUV420 -> RGB

GOOD:
YUV422 -> RGB

Then..

FINAL:
YUV422 -> RGB -> [rgb]filter/process -> (RGB->YUV422) -> [yuv422]New_Container

The above is under the assumption that your tools are RGB color space,
as in the example, P.Shop software -- your preference.

I don't agree with your quality assumptions for VHS but I would agree Huffyuv would be a good "preservation" format if you intended future restoration with future digital filters.

But, on the other hand, AVIsynth and it's arson of filter plugins
can keep the color space conversion(s) to a minimum, if any.
And, assuming that you are capable of re-configuring your Video Transfer
setup, you could configure a YUV 4:2:2 color space format capture,
with the proper codec. One popular codec is the HuffyUV or something.
This one is a Windows version. But, since you work inside MAC pc's,
you may have a problem, because I don't think there is support for
this codec on a MAC enviornment. But, surely, there must be a codec
similar for this type of machine

In any case..

In AVIsynth, there are a number of filter plugins that work under the
YUV422 color space. You only need to have a little bit more patiance
and strive harder to learn and understand them, and use them to your
advantage in your source.

An AVIsynth 'For Example'..

if you have a group of pictures (images) in your video source that:

A - [Group A] - need to deal with 'background' noise..
B - [Group B] - need to sharpen blurry edges..
C - [Group C] - need to brighten bad drop outs..

EDIT: - revised..

then you could write an AVIsynth (AVS) script, call it,
"vhs_transfer_ep27.avs"

 # vhs_transfer_ep27.avs
 # EP 27 requires three segments to filterize.
 # sharpen; blur; and levels;

 clip1=TRIM(0,1200).sharpen(1)
 clip2=TRIM(1201,4500).blur(1.58)
 clip3=TRIM(4501,7980).Levels(16, 1, 255, 0, 255)

 return clip1++clip2++clip3

(Note, the above will work as it is, but realize that it is ment to
illustrate how you can string several filters inside one video source
and process those segments accordingly)

The sharpen() function does what it says.
blur() will blur (lets say, denoise here) at its max value, 1.58
levels() in this case will brighten this area of video

The above are all Internal filters of AVIsynth.

My suggestion would be to do some research on all the various AVIsynth
filter plugins that are available, and see what each one does. You
may find various ones that your source(s) will benefit from.

About the only tricky part of filter plugin usage is knowing which
filter to use inside a color space.. ie, YUV422.YUY2 vs. YUV420.YV12

But, just imagine how you could apply this to your library of video
archives and process (filterize) them to your needs.


RESOURCES:

** Internal filters - listing of all (by catagory)
** AVIsynthManual - listing of all keywords/commands/filters etc.
** External filters - listing (maintained by warpenterprises doom9 member)

-vhelp 3750

Capturing to 4:2:2 YUV space may prepare the VHS capture to use these filters but simple upconversion or oversampling will not improve quality. 4:2:2 capture places all the pixels where the filter can work with them but most of the pixels will replicate the same value. Worse case is if you capture higher frequency noise in your VHS capture. This may get interpreted as motion or detail and may confound the filters. Best way to avoid that is to lowpass filter the capture just above the maximum possible VHS bandwidth for Y, U and V. This is expensive to do in analog. Best to digital filter the capture file.

[edDV]

Why are you doing this? As an intellectual exercise?

Better you learn realworld video production techniques while you are in school and try for the Nobel Prize after you have mastered the basics. There isn't enough detail in a vhs tape to create an acceptable 13x pixel (720p) or 30x pixel (1080p) expansion using existing technology. It may well be possible in the future but the technique will require detail knowledge of complex digital filtering.

You need to decide whether you are a video production major or should transfer to a math or engineering career. I'm not trying to put you off. Many discoveries are made when attempting the "impossible".

Video production leverages other's engineering work to enable production of cutting art films or videos. Restoration falls into a different category and is a technical specialty.

edDV--

I'm doing this as part intellectual exercise/personal challange, part stress relief (yes, this is stress relief for me,) and part preservation of unedited footage. Censorship editing; especially the overzealous editing this series was subjected too after 9/11 truely irks me and it hurts to watch what they've done to the series. Many of the most serious and intense lines were removed as were some of the visual effects departments best scenes. The audio editing hurt the plot and the lack of DVDs, edited or otherwise means that this is all I've got. The edited Video CDs I imported a few years ago (PAL format) are actually worse than my VHS Source footage on a PAL Monitor. That's saying something.

Video CD is worse in almost every way.

...

I agree with that. Preserve what is there and fight another day. Technology will improve but the VHS copy will degrade.

Ironically enough, this is why my first step was to remaster and preserve the footage on high quality MiniDV Tapes in DVCAM Mode. I don't know how much longer these tapes will last and from what I'm understanding, if I take care of the MiniDV Tapes (rewind/fast-forward every six months, only use two brands, stick with my pro-level Panasonic tapes that are known for reliablity etc.,etc.) they'll last me quite a few years and if the tape begins to wear I can do a lossless firewire dub from camcorder to camcorder. Using my PUT (Preserve, Upconvert, Transfer) method this is step one. I think I can do this quite easily once I get my hands on a few more pieces of equipment. Upconverting and Transferring to a viable HD Format are something I'm just trying to gather information on right now; Blu-Ray and HD-DVD might not even catch on. I might be better off using HDV eventually or even the old D-VHS Standard.

My primary goal right now is to preserve these in a format suitable for archival until more HD Upconverts have been done successfully and the HD world isn't as insane as it is right now.

I wouldn't worry about the DVCAM tapes wearing out in the next 5-10 years, if then. Use the Pro tapes which are better constructed. Digital recording minimizes most tape playback issues. Analog video is recorded directly and any dropouts or noise will be visible in analog playback. Digital recording has a much wider latitude and will play back full quality even with added noise (e.g. dirty heads) well beyond where analog will be unwatchable. DV camcorders also have extensive dropout and other data gap compensation.

Here is my advise. Go for the best SD capture you can. DVCAM is very good, next up would be an uncompressed YUV capture using the best pro equipment. Talk to a respected dub house, they may give you a student break*. They would probably cap through a frame sync to SMPTE-259M to digital Betacam. That would make a great capture master. Get a YUY2 or YUYV 4:2:2 uncompressed AVI dubbed to a hard disk. Later you can compress that to Huffyuv more cheaply than paying them. That is what you will use for filter reference input.

Go for best SD filtering. Let the playback device upscale to HD in hardware.

Why do you want HD files? Each HDTV will be able to upscale SD and the internal scaler will be optimized for that HDTV's display technology.


* Dub houses won't do a commercial tape without a letter from the copyrite owner.

[Cornucopia]

edDV, of course, you're right!

What was I thinking? Any of the color-under formats would be severely chroma bandwidth-limited, and as such would be completely covered under the 4:1:1 sampling space.

Now, if you're going to be doing multiple subsequent post-processing (which the OP wants to do)--then it would make sense to upconvert to 4:2:2/uncompressed after capping the 4:1:1. But I certainly agree that 4:2:0 with MPEG compression isn't the way to go. And I agree with the dub/preservation/convert suggestions too.

Scott

[vhelp]

VCR S-Video output is analog luminance (Y) + color modulated around a 3.58MHz subcarrier (C). Where do you get the 4:2:0? Are you referring to an internal TBC?
..
VCR's output analog so I would think this is true even if there is an internal TBC.

No. I was not refering to an internal TBC. Sorry for the confusion..
Ok. Maybe it is not 4:2:0 but I was basing that on my eye and how I
view the video's quality's over-all level. And my thinking was was
the 420 sampling (as I thought it was closest to saying lack of detail)
but perhaps 4:1:0 or 4:0:0 is closer to the lack of detail. But, in
any case, I ment well in trying to describe similar output qualities
from this low detail source medium.

But, now that I had time to re-think my response (and yours) I now feel
the 4:1:0 or 4:0:0 is the closest to video detail. Weather I'm wrong
or not, is not really the purpose here. Even if what you say (and post
via link) is true. I'm not arguing these facts. But you seem to have
misunderstood me, and I may have confused myself at first. Anyways.
We both mean well.

Capturing to 4:2:2 YUV space may prepare the VHS capture to use these
filters but simple upconversion or oversampling will not improve quality..

Now here, I think you missunderstood me. Still..
I didn't say upconversion or oversampling will improve quality.

hmm.. I think that what I was trying to say here, was that it would
be better to capture at 422 or receive video in 422 sampling. And
by obtaining video at 420, quality *would* be reduce, (even from VHS)
because info/detail would be less, than if it were 422 sampling,
which has more info/detail.

It a user (lets say) recieves video data in a final 420 sampling, then
that user has less video info/detail. And, (add to this) if the user
requires to incorporate filtering at some point after recieving the
420 video, (even if VHS) that user would have much less detail to
reproduce or make better (ie, user's planned up-convert to HD) from
this 420 video because during the receiving phase of the 420 video,
the process (internally) threw out data during the capturing and
encoding to a final (in this users case) MPEG-2 source.

As such, this user would have been better off with a unit/device that
were to receive (capture) and finalize to a 422 video source, even
if MPEG-2 was the final source.

So, weather the original source is VHS or other, does not matter.
But obtaining the final source with the most "sampling" is critical
here. So when I said that VCR's output 420 I ment well.. even if
I was in error on this. I still ment well. And it does not change
the fact that the source (ie, VHS) should be captured with a sampling
of 422 as the final source.. forgeting about this nonsense about
VCR's transmitting Analog 'wise. (It's not important)

-vhelp 3756

[Cyrax9]

Video CD is worse in almost every way.

...

edDV--

LOL I wish I knew that before I paid $99.99 for partially edited VCDs with some of the worst ghosting/trailing effects in the history of video! What's sad is I burned a VCD using Roxio Easy CD Creator Version 4 quite a long time ago and my 400+MB VCD looks better than the proffessionally pressed ones! Sure the files are bloated but they rival my VHS master for picture quality even today.

I would never use VCD as an archival format, I'm just noting that's the extent of a "professional" release and it wasn't even a US Release despite the show being a US Show. Sad, isn't it? The only excuse I can think of is that the company doesn't see the show as marketable, even when hacked to bits in the USA. A German company did edited/dubbed versions of a few episodes in PAL for a German release on DVD and the Chinese Video CDs are almost identical to the US TV airings and are legit copies. There were no US releases without some form of editing -- ditto for the few UK releases of the show -- it doesn't help that the VHS tapes they did produce are run together without openings or endings on episodes and titles are nonexistant. Half-baked is an understatement -- how these got past QA is beyond me.

I wouldn't worry about the DVCAM tapes wearing out in the next 5-10 years, if then. Use the Pro tapes which are better constructed. Digital recording minimizes most tape playback issues. Analog video is recorded directly and any dropouts or noise will be visible in analog playback. Digital recording has a much wider latitude and will play back full quality even with added noise (e.g. dirty heads) well beyond where analog will be unwatchable. DV camcorders also have extensive dropout and other data gap compensation.

This is true, the tapes I planned to use were Panasonic AY-DVM83MQ Master Quality Pro MiniDV Tapes. These are about $10 a tape. I'd stripe them in DVCAM Mode (looks better on the PDX10 than DV SP) and then record onto them via the PDX10. Striping would be done so that these could be treated as "insert edits" if need be and supposedly striping makes a tape easier to work with.

Here is my advise. Go for the best SD capture you can. DVCAM is very good, next up would be an uncompressed YUV capture using the best pro equipment. Talk to a respected dub house, they may give you a student break*. They would probably cap through a frame sync to SMPTE-259M to digital Betacam. That would make a great capture master. Get a YUY2 or YUYV 4:2:2 uncompressed AVI dubbed to a hard disk. Later you can compress that to Huffyuv more cheaply than paying them. That is what you will use for filter reference input.

This sounds like a good idea, I'd need to find a dub house and in my limited experiance with digital Betacam it's supposed to be one of the best prosumer formats around. I'd only use an uncompressed AVI if I intended to do further filtering on the video. I've considered Huffyuv compression but for some reason my PC crashes every time I load the codec. I hate using the Studio codec since it's a bit buggy at times.

Go for best SD filtering. Let the playback device upscale to HD in hardware.

I thought about this, and it seems to make sense. I take it I should find a high-end playback device to ensure the most reliable (don't want to say best) hardware filtering possible.

Why do you want HD files? Each HDTV will be able to upscale SD and the internal scaler will be optimized for that HDTV's display technology.

Partly because I want to be able to store them on an HD Format, partly because I sometimes use a laptop (vacations) for video playback or a device that doesn't have a great converter in it. I also thought that the upconvert wouldn't be too bad as technology progresses.

I'm probably going to go the DVCAM route since I just bought a Vidicraft VDM200 Detailer/Proc Amp and I've seen what these do when connected to a JVC HR-S9911U S-VHS VCR and the output is outstanding. I still need a DNR and a standalone TBC but those shouldn't be too hard to find.

@vhelp

The card itself was great when I first got it, don't get me wrong it's not a totally horrible card but it hasn't aged well. The 640x480 limit is obnoxious and for some reason when I had to send my PC back to the factory for repairs since they hadn't (and still refuse to) sent me the proper recovery CD for my computer which is a bit of a lemon, the machine was returned to me "repaired" with several things never running quite right again. Sadly the card was one of them, I tried reinstalling it and it just never worked well after the factory messed around with the machine. Prior to DVD it was a wonderful card but today it doesn't help me with the 640x480 cap. I used it for several captures and it's a wonderful card but as I conmtinued ot upgrade my machine I never upgraded the card and on many tapes I had signifigant frame drops that were inexcusable. I reconfigured the system and it didn't help. I ran the tapes into a DVD Recorder and they came out fine. The only thing I can think of is the card is damaged and while it was wonderful when I started with it it's a bit sub-par by todays standards. I didn't mean to suggest it was always a bad card.


Also, this is probably an off-the-wall/completely insane idea but I figured I'd run it by everyone in here. Remember those "mirrors" they used to make that would let you videotape 35mm filmstrips and slides? -- You projected the slide/filmstrip and then recorded it with a VHS Camcorder similar to the way DuMont would point a Kinescope at a TV to record their programming.

What about this for an HD Upconvert. It's a bit "odd" -- I know. Take the VHS tapes and remaster them in the highest SD Quality possible within my budget, and output to the afore mentioned tapes; archive said tapes for the time being.

Invest in an HDV Camcorder and a projector -- possibly rent the projector and a fairly large screen.

Use the projector to display the output of the SD copies of the video. Using a "mirror kit" similar to the ones used for 35mm and slides, record what's being shot through the mirror kit on an HDV Camcorder effectively "refilming" the video. Since I can do the audio conversions on my computer easily and the pixel aspect ratio is the problem, would it not make sense to just "film/tape" a screen with the original SD video through a mirror and then do some post-production work in vDub?

Is this a feasable idea or just a pipe-dream? It seems doable but I don't know if it'd work.

Here's my current plan, namely due to the budget I have for this project and a small investment I made tonight.

JVC HR-S9911U --> Vidicraft Vidimate VDM200 --> standalone TBC (DNR if needed -- these tapes look like a DNR could actually lower the quality rather than raise it) --> Passthrough PDX10 into iMovie. Edit the video in iMovie (no filtering) and edit the audio on my PC using something like SoundFORGE --> Insert new audio with captured video -- > output to MiniDV tape in DVCAM Mode using the PDX10 and iMovie.

Archive tapes and see what else I can do.

Also, correct me if I'm wrong, but if the tapes source signal is VHS, and I'm using a S-VHS VCR, wouldn't I want to AVOID using the S-Video output on the VCR since the source material is composite? In effect wouldn't I want to run Composite Red/White/Yellow Cables to Composite Red/White/Yellow Cables and then run those into the camcorder? Isn't adding Y/C seperation through an S-Video cable just going to further lower the quality of the video since it wasn't originally an S-Video source?

I've heard conflicting viewpoints on this but there seems to be a consensus that if the source is composite (VHS/LD) the output for capture should be composite and not S-Vid when remastering. Is this correct, or are some people uninformed?

Thanks again guys, and I'm looking forward to working on this -- as much as it's going to take me a lot of time to complete I'm up for the challenge.

It sounds to me like 4:1:1 is okay since the source footage is VHS, and I'm kicking myself as I try to remember whatever possesed me to think it was 4:2:2. These tapes are in good condition but they're still VHS.

I'm just wondering if my "scope" idea is insanity or viable. Thoughts on that are welcome.