Go Back    Forum > Digital Video > Video Project Help > Restore, Filter, Improve Quality

Reply
 
LinkBack Thread Tools
  #1  
12-05-2016, 05:42 PM
jnielsen jnielsen is offline
Free Member
 
Join Date: Nov 2015
Posts: 53
Thanked 8 Times in 6 Posts
I have done som testing. Here are my notes. Comments are highly appreciated.

Background
I want to convert interlaced MPG2 (from my VHS-captures) to interlaced MP4. Many capture devices can capture directly to interlaced MPG2, and it is the ideal source for making DVD-records. (Because the files on DVD is similar to interlaced MPG2). However I want MP4 for distributing on USB-RAM (for playback on computer), because not all computers support MPG2. I also want deinterlaced because not all players deinterlace by default on playback. It means I want deinterlaced MP4

Deinterlacing with Handbrake
Setting deinterlace = slower ( "yadif" ). Handbrake deinterlaces much better than "AnyVideoconverter" and other "converters" I have tried. AnyVideoconverters built in interlace seems to be like "fast" in Handbrake. A time saving alternative to making non-interlaced MP4 could be just distributing MPG2 interlaced files, almost directly as captured. And then let the player on the computer do the interlacing.

Deinterlacing by player software
Windows Media Player: MPEG-2 format video isn't supported by default on Windows 8.1. You can install the free "ffdshow" encoder. It seems then to deinterlace automatically. However the quality is lower than VLC Player.

MPC-HC: MPC-HC is a light-weight, open source media player for Windows. It deinterlaces MPG2 by automatically, seems to be in "blend" mode, which is fast, but not the best way.

VLC Player: VLC Player is a good player, however it does not interlace automatically. You have to turn it on ("video", "deinterlace", "yadif" ) or permanently under "tools", "settings". Deinterlace quality is good.

VLC is the best player for deinterlacing MPG2 at playback. However it is complicated for many users to set deinterlacing up, and the player must be installed first.

Handbrake vs. Player deinterlacing
Handbrake vs. deinterlacing by playback in VLC Player ( setting = yadif ). Handbrake is visible better. It means that you cannot depend on players deinterlacing on a computer if you want the best quality.

A note on playback by smart-tv
My experience by testing two identical filmclips in MP4 (deinterlaced) and MPG2 (interlaced - from capture ): MP4 needs a fast smart TV for good playback ( seems like it needs more computerpower). If not MPG2 will look better. TV´s seems to handle MPG2 more easily. However not all models cut off the overscan area, it means you will see the headshifting noise in the bottom 10 pixels and the black sides 4-8 pixels. On a fast TV there seems to be no difference between interlaced MPG2 and MP4 deinterlaced by Handbrake.

Conclusion
I find that the best is to make a good deinterlaced MP4 file for computer playback. Of the converters I have tried Handbrake is far the best, and the result is that MP4 interlaced look considerably better at playback than MPG2 deinterlaced during playback on pc.


It means my workflow is
* Capture in MGP2
* Trim the ends by a simple non-destructive "edit"
* In Handbrake: Convert to MP4, deinterlace and crop the overscan area ( eg. 8 pixels at the bottom)

I have read about Avisynth, and plugins for editors like Vegas. However time and simplicity in workflow is also important for me. What is your opinion, do you prefer MP4 or MPG2 or both distribution on USB-RAM?

Last edited by jnielsen; 12-05-2016 at 06:03 PM.
Reply With Quote
Someday, 12:01 PM
admin's Avatar
Ads / Sponsors
 
Join Date: ∞
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
  #2  
12-05-2016, 05:52 PM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
Everything posted above poses serious quality issues and contains what amounts to fictional claims of high quality from certain software products and methods, as well as more inaccuracies and misconceptions than I would find time to correct. Others can chime in as they wish, but digitalfaq usually aims for much higher quality work than you're aiming for, and I feel that the unsupported claims you're making will mislead and/or disappoint others who look for more and better. Cropping overscan? That one alone is a dead giveaway.

MP4 is not a format. It's a container. Please be more careful and precise about your terminology.
Reply With Quote
  #3  
12-08-2016, 05:49 PM
lordsmurf's Avatar
lordsmurf lordsmurf is offline
Site Staff | Video
 
Join Date: Dec 2002
Posts: 13,503
Thanked 2,448 Times in 2,080 Posts
Best = QTGMC, or a complex Yadif-based method. That's been true for a few years now, and will probably not change in the near future.

Best lazy method = Yadif alone. You'll notice many interlace artifacts with animation, anything high definition, or anything with lots of geometric shapes (especially linear diagonals). For SD(ish) live-action material, Yadif will not be overly noticeable. There will be quality loss, so don't trash the master. I sometimes do a quick-and-dirty Yadif for testing or quickly showing a clip to somebody online.

Most encoders/etc have crap deinterlacers -- everything from the expensive payware Premiere Pro CC, to the freeware Avidemux or Handbrake.

MP4 isn't a format, but a container. I'm going to guess that you're referring to H.264 or XviD, though format doesn't really matter for this conversation. It was just extra details. That said, both support interlacing poorly.

The statement "files on DVD is similar to interlaced MPG2" is false. DVD-Video *is* using MPEG-2 (interlaced or progressive).

I have no idea what a "USB-RAM" is supposed to be. Did you just add "-RAM" on reference to thumb/flash drives? aka small solid-state USB sticks.

Junkyware/Chinaware like AnyVideoConverter is based on freeware. So the fast deinterlace setting in Handbrake is probably what they used.

WMP sucks at deinterlacing, and just blends. For players, VLC's Yadifx2 is currently the best, outside of hardware. VLC is not complicated; some users are just lazy dullards (that spend too much time on Facebook and Youtube). If you want complicated, that's Avisynth or PFclean.

And MP4 has nothing to do with a TV. The specs of the player determines what can be played. But honestly, most are standardized, and it's badly-encoded video that makes those barf -- it's not the fault of the player.

Why are you deinterlacing?

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
The following users thank lordsmurf for this useful post: jnielsen (12-16-2016)
  #4  
12-12-2016, 06:36 AM
koberulz koberulz is offline
Premium Member
 
Join Date: Feb 2016
Location: Perth, Australia
Posts: 453
Thanked 3 Times in 2 Posts
Regarding the TV, the OP may be plugging the USB drive directly into the TV. In that case the TV is the player, and thus becomes relevant to playback.
Reply With Quote
The following users thank koberulz for this useful post: jnielsen (12-16-2016), lordsmurf (12-13-2016)
  #5  
12-12-2016, 09:23 AM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
The container and codec are irrelavant, in that smart TV can play MPEG2 and mp4/h.264 alike whether they're interlaced or not (as long either is encoded within standard specs). The dubious aspect here is deinterlacing for no apparent reason other than the O.P. is using software and hardware that doesn't deinterlace well. This comes down to the owner having to deinterlace and re-encode everything that comes over the cable channels or is sold as commercial video in stores.

Lossy capture followed by lossy re-encoding involves quality degradation especially with an SD source that's inferior to begin with, and poor deinterlacing and unnecessary rescaling have a cost as well. Many sources aren't interlaced to begin with, as the number of tapes and animation that are telecined, field blended, and work with frame duplication cannot be deinterlaced without serious damage. Multiple stages of degrading a source is a bad idea from the start. Processing like this looks no better than sloppy UTube work and is a long way from "ideal". There is nothing new about any of this, which no reputable tech forum would recommend.
Reply With Quote
  #6  
12-15-2016, 04:05 PM
jnielsen jnielsen is offline
Free Member
 
Join Date: Nov 2015
Posts: 53
Thanked 8 Times in 6 Posts
Quote:
Originally Posted by sanlyn View Post
Everything posted above poses serious quality issues and contains what amounts to fictional claims...
Thank you for your response.

"Fictional claims" I do not know exactly what you mean. If you refer to my comments on deinterlacing methods in Handbrake vs. different players, it is not fictional but based on hours of testing on the same short clip. My conclusion is that deinterlacing by the player is not as good as deinterlacing by Yadif in Handbrake. Among the players VLC is best ( but unfortunetely it does not deinterlace by default ). And because deinterlacing by player software is not so good I want to deinterlace for computer-viewing, instead of keeping it interlaced.

"Cropping overscan?" Well I do not like that too, but why is it so important, if it is for computer-playback or smart-tv playback only, and not for making a DVD?

But anyway I got the point, the quality should be better. So in the last days I have startet experimenting with AVisynth and MeGUI, not easy for me, but now I have managed to understand Avisynth a little and I tried to modify the standard Avisynth script in MeGUI with a crop + addborder command and the QTGMC deinterlacer script. And yes it looks much better. I hope to make it even better, when I learn to use more scripts.

Quote:
Originally Posted by sanlyn View Post
The container and codec are irrelevant, in that smart TV can play MPEG2 and mp4/h.264 alike whether they're interlaced or not (as long either is encoded within standard specs).
My experience ( based on testing the same clip on an old and new smart tv) is that MPEG2 interlaced plays more easy than MP4 deinterlaced on old slow smart tv´s. More fluently, loads quicker, less pixelation. I maybe wrong but I imagine that the deinterlaced MP4 needs more processing. Some old smart tv´s cannot even play MP4 but only MPEG2. My MP4 is made with Handbrake Codec is named "H.264 (x264)". I still just dont get that codec - container - difference.

Last edited by jnielsen; 12-15-2016 at 04:36 PM.
Reply With Quote
  #7  
12-15-2016, 05:26 PM
jnielsen jnielsen is offline
Free Member
 
Join Date: Nov 2015
Posts: 53
Thanked 8 Times in 6 Posts
Quote:
Originally Posted by lordsmurf View Post
Best = QTGMC, or a complex Yadif-based method. That's been true for a few years now, and will probably not change in the near future.

Best lazy method = Yadif alone. You'll notice many interlace artifacts with animation, anything high definition, or anything with lots of geometric shapes (especially linear diagonals).
Thank you for your response and for this great forum.

In the last few days I have tried Avisynth a little (I find it complicated) an also tried the QTGMC deinterlace script. I see what you mean, it is much better, especially with small moving patterns like, in my test-video, the red white squares on an italian restaurant style tablecloth. And yes Yadif in eg. Handbrake is not so bad, and maybe the best choice if you just find Avisynth scripting too difficult.

Quote:
Originally Posted by lordsmurf View Post
The statement "files on DVD is similar to interlaced MPG2" is false. DVD-Video *is* using MPEG-2 (interlaced or progressive).
I guess I do not fully understand this. Is MPG2 and MPEG-2 different things? I have just noticed that it is very easy to make the vob files on a DVD-video to MPG2 files, it seems like almost no conversion is needed.

Quote:
Originally Posted by lordsmurf View Post
I have no idea what a "USB-RAM" is supposed to be. Did you just add "-RAM" on reference to thumb/flash drives? aka small solid-state USB sticks.
Yes thats what we call it here

Quote:
Originally Posted by lordsmurf View Post
WMP sucks at deinterlacing, and just blends. For players, VLC's Yadifx2 is currently the best, outside of hardware. VLC is not complicated; some users are just lazy dullards (that spend too much time on Facebook and Youtube). If you want complicated, that's Avisynth or PFclean.
It is also my experience, that VLC is the best of the players, however I think I used only Yadif and not Yadif (2x). Maybe it can be even better. I think that MPC-HC also was not much better or similar to WMP.

Quote:
Originally Posted by lordsmurf View Post
Why are you deinterlacing?
Good Question. My case is that I used to make videotape to DVD, but now many people want video on a "USB-stick". If I just give them interlaced MPEG2, some cannot even play MPEG2, and others only interlaced. I guess I could instruct them to install VLC, but many has no clue about players, they just want to click on the file and then it should play. Also the MPEG2 files meant for DVD has the overscan area and headswitching noise.

It means I need to do something with the files at least crop the overscan area and at the best add some black borders. I like MP4 (H.264, x264) because it plays on all computers. I like deinterlaced becaues the players (except VLC) deinterlace poorly. But If I give them deinterlaced MP4, I have problems about the quality degrading in the interlace process.

What would you give to people who want video on a "USB-stick"?
Can I just crop and add border to a MPG2 file and tell people to use VLC player?
Can I use MP4 interlaced (only MPG2 to MP4 conversion) or MPG2 deinterlaced? (only deinterlacing MPG2)
Or will I have to make a better deinterlaced MP4 (when I get to know Avisynth)?

Last edited by jnielsen; 12-15-2016 at 05:39 PM.
Reply With Quote
  #8  
12-15-2016, 09:36 PM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
Quote:
Originally Posted by jnielsen View Post
Is MPG2 and MPEG-2 different things? I have just noticed that it is very easy to make the vob files on a DVD-video to MPG2 files, it seems like almost no conversion is needed.
MPG and MPEG, when referring to video file names (aka "containers") means a video file that has been encoded using one of two codecs: MPEG2 or MPEG1. DVD cannot be created using any other codecs. MPEG-2 is the codec standard for retail DVD and is a standard used in digital TV broadcasting. When MPEG2 is authored for DVD disc, the .mpg or .mpeg containers are reorganized into .VOB containers.

MP4 is a container whose contents can be encoded using several codecs, including MPEG-2 Part2 for BluRay/AVCHD, h.264/AVC (aka "MPEG-4"), MPEG-4 ASP, h.263, VC-1, and others. It was designed with progressive internet streaming in mind. The MP4 container cannot be used for commercial standard formats such as DVD, BluRay, or AVCHD.

Quote:
Originally Posted by jnielsen View Post
Also the MPEG2 files meant for DVD has the overscan area and headswitching noise.
Nonsense.

I don't know how you have time to watch videos if you're so busy re-encoding and deinterlacing everything you buy in DVD and BluRay and everything you watch from cable TV on bad TV sets and rare media players that don't how to play interlaced or telecined video. Consider getting a better computer and learning to work with low-quality sources to make them a little better, or at least viewable by people with two good eyes and some sense of visual discrimination. Dumbed-down mediocre video for dysfunctional users who are already irretrievably dumbed-down is being done worldwide, 24/7/365, by anyone who can click a couple of icons. There's nothing new here.
Reply With Quote
  #9  
12-16-2016, 08:20 AM
jnielsen jnielsen is offline
Free Member
 
Join Date: Nov 2015
Posts: 53
Thanked 8 Times in 6 Posts
Quote:
Originally Posted by sanlyn View Post
Nonsense.

I don't know how you have time to watch videos if you're so busy re-encoding and deinterlacing everything you buy in DVD and BluRay...
The files I refer to when i write "MPG2 files meant for DVD" are the files I capture from VHS video. They (at least mine) have the overscan and headswitching noise area.

Last edited by jnielsen; 12-16-2016 at 08:50 AM.
Reply With Quote
  #10  
12-16-2016, 06:05 PM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
The appearance of the side borders and head switching noise has nothing to do with the codec used to capture, and nothing to do with overscan that would be unique to MPEG2. You shouldn't be capturing to lossy codecs if you wish to modify your image. Not recommend anywhere in this forum. Also, your recommendation to remove borders and head switching noise by cropping and then resampling the image size is idiotic. Only a rank beginner with very little knowledge would resize frames in that manner.

Last edited by sanlyn; 12-16-2016 at 06:18 PM.
Reply With Quote
  #11  
12-16-2016, 08:12 PM
jnielsen jnielsen is offline
Free Member
 
Join Date: Nov 2015
Posts: 53
Thanked 8 Times in 6 Posts
Quote:
Originally Posted by sanlyn View Post
The appearance of the side borders and head switching noise has nothing to do with the codec used to capture, and nothing to do with overscan that would be unique to MPEG2. You shouldn't be capturing to lossy codecs if you wish to modify your image. Not recommend anywhere in this forum. Also, your recommendation to remove borders and head switching noise by cropping and then resampling the image size is idiotic. Only a rank beginner with very little knowledge would resize frames in that manner.
I know that the side borders and head switching noise has nothing to do with the capture codec.

I know lossless would be better, but my capture program can only capture MPG2 or AVI. Some day I will try Virtualdub or whatever it takes, but that is the next step. I like MPG2 because it does not take much space and is easy to convert to DVD.

I do not resize the frame, just crop it. Eg. from 720x576 to 704 x 556

I have now installed Avisynth and MeGUI and QTGMC deinterlacer script, so now I can crop and add borders, and deinterlace in a better way with QTGMC. It already looks a lot better, but it is very slow.
Reply With Quote
  #12  
12-17-2016, 12:57 AM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
I should know by now not to have gone deeper into this thread, but every time you go into detail there are more revelations about sloppy or incorrect processing. Sorry to have to have to keep after this but.....

Quote:
Originally Posted by jnielsen View Post
I know lossless would be better, but my capture program can only capture MPG2 or AVI.
Not just better; lossless is as accurate and as versatile as you'll get, given decent playback hardware and capture cards. Lossy codecs like MPEG for capture are not the "ideal" you advertise, especially at mainstream or lower bitrates. As you learn more about video you'll realize how difficult it is to repair lossy compression noise and artifacts while maintaining any level of the original image integrity. At the outset you throw away a large percentage of visual data, yet you already know that there will be another stage of more re-encoding loss. So this method is a long way from ideal. It's self crippling from the start.

Quote:
Originally Posted by jnielsen View Post
I do not resize the frame, just crop it. Eg. from 720x576 to 704 x 556
Wait a minute. You can't say at one point that "I like MPG2 because it does not take much space and is easy to convert to DVD", and then say that you crop to a non-standard frame size that's not valid for DVD. If you're trying to un-confuse users who are already clueless about video, then you are confusing other users who are more advanced. What do you mean by "DVD"? The DVD format has rigid specifications not open to free-for-all processing. If by "DVD" you just mean an mpeg file, that's not a DVD. It's just an .mpeg file of dubious specification.

704 x 556 is a non-standard frame size for both PAL and NTSC. It's a mod-4 dimension, meaning that filters, software, and players programmed for mod-8 or mod-16 standard frame sizes will have to do some special processing if you expect to get a 4:3 display aspect ratio from 704x556. What most software will do -- that is, "most" software that will even attempt to handle it correctly -- will be to stretch the image vertically in order to display it at a 4:3 display aspect ratio. So your nonstandard frame size is distorting the original image proportions. As for users who are too clueless or too lazy to disable overscan on their TV, you have just moved 15% to 20% of the image into the masked overscan area on their TV's.

What if your projected image is supposed to be 16:9, not 4:3? Again, you'll have some distortion. And for standard definition BluRay you cannot use a 704 frame width for DAR 16:9 video. But these odd-size mpeg's are a far cry from BluRay, which has even tighter specs that DVD has.

There are other problems. Player devices accustomed to standard frame specs are making exceptions with this material because it's an unexpected pixel aspect ratio (PAR). More specifically, many encoders/decoders, players, and Avisynth filters expect at least a mod-8 image ratio (pixels that can be processed and even encoded in 8-pixel blocks). If you don't know what "MOD" means, you should have looked it up before now. Look at the standard frame sizes in SD and HD video: 352x576, 352x480, 704x576, 704x480, 720x576, 720x480, 1280x720, 1440x1080, 1920x1080, 3840x2160....these are all mod-16 frame sizes (except for the last three formats, which are mod-8 vertically). Does this tell you something about the way many encoding, filtering, decoding, and playback devices are optimized? What it tells me about your methods is that your criticism of various players really talks about the way they handle nonstandard pixel aspect ratios, not the way they handle properly processed and formatted standard video.

Quote:
Originally Posted by jnielsen View Post
I have now installed Avisynth and MeGUI and QTGMC deinterlacer script, so now I can crop and add borders, and deinterlace in a better way with QTGMC. It already looks a lot better, but it is very slow.
Of course it's slower. It's trying to clean up your noisy lossy capture, which you shouldn't be using in the first place. You could speed it up by using a faster QTGMC preset. But at least it's an effort at progress, although undone by a second lossy encode.

What are you going to do about incoming VHS sources that are not interlaced? What if they're telecined or use various pulldown setups? What about blended-field NTSC-to-PAL conversions? What about film sourced vidoes that were shot at 24fps progressive and speeded up to 25fps for PAL, and are already progressive but encoded as interlaced (and which if deinterlaced will give you two duplicate frames for every original frame)? What about animation, which is seldom interlaced and which already consists of many duplicate-frame progressive sequence techniques? What about non-PAL users who have NTSC, which has even more production permutations on tape that PAL does? I realize you haven't mentioned these but have focused only on interlaced home-video tapes made with consumer cameras -- which pretty much abandons a great many users and tape formats. So, again, your suggested method isn't "ideal" for many people.

Besides, you miss the major point. With any input source -- but especially with horrible noisy typical VHS tape -- one lossy encode at capture before denoising or cleanup is bad enough. But two lossy encodes are always worse. Always.
Reply With Quote
  #13  
12-17-2016, 01:20 AM
lordsmurf's Avatar
lordsmurf lordsmurf is offline
Site Staff | Video
 
Join Date: Dec 2002
Posts: 13,503
Thanked 2,448 Times in 2,080 Posts
@sanlyn: calm down, be nice. He means well, but just has some wrong info. I don't detect any degree of stubbornness (which also gets my dander up), but just lots and lots of bad understandings.

So let's educate him.

FYI: He's posting from Denmark, so there may be some slight terminology differences. USB-RAM, for example, seems to be the local reference. And I'm betting English isn't his first language, though he has a definite command of it. Something to consider.

I want to read this over later, but something caught my attention while skimming: MPEG capture. I do it all the time, be it to DVD with a JVC recorder, or 15mbps with an ATI AIW card. In fact, I'm doing the latter this very minute. I'm not, nor have I ever been, an AVI purist. (In fact, I'm capturing DV over s-video, which would put some purists into fits.) You use the right tool for the need.

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
The following users thank lordsmurf for this useful post: jnielsen (08-21-2017)
  #14  
12-17-2016, 06:02 AM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
I get your point, lordsmurf, and hope I'm not seeming too strange here. Perhaps I'm using some of the wrong terminology myself and looking as if I'm sitting here smoldering in my chair.

If so I apologize to jnielsen and will be more careful.

I myself have made some VHS transfers directly to DVD at average bitrates , but not with the clear intention of filtering and then re-encoding to another format. Those few DVD recordings were made on a recorder with an LSI chip, played with high-end JVC and Panasonic "AG" players, and were recorded from pristine Hollywood retail issues. But they were not cropped and re-encoded to nonstandard dimensions. A couple of them are favorites that I've watched again. When I do, I wish I had kept the tapes so that i could capture, clean, and process those movies properly.

Principally, I object to the use of the word "ideal" to describe this method as being a wonderful way to archive video that we know beforehand is going to be re-processed and re-encoded. Perhaps "ideal" is not the best description. Expedient or utilitarian, yes. I agree that if we all tried to rework everything along perfect lines, none of us would live long enough to see our own projects completed.
Reply With Quote
The following users thank sanlyn for this useful post: jnielsen (12-22-2016)
  #15  
12-22-2016, 12:43 AM
deter's Avatar
deter deter is offline
Premium Member
 
Join Date: Dec 2009
Posts: 320
Thanked 28 Times in 26 Posts
sanlyn,

Hi, it has been a while since the Michael Jackson video days.

Yea, it amazes me to no end how people have no clue. A guy wanted me to convert a VHS tape to 4K, What? Than I had another guy who captures everything in AVI fully uncompressed and stores the videos and never converts them to mpeg2. What is the point of this madness....

I had to learn from Lord Smurf I used to do WMP files and convert them to 720x576 cause it had more lines, that was my thinking. It took a good few years to get a grasp on how to do this stuff, Lord Smurf can't say enough good things about. he helped me through everything.

To the guy in this post, keep the videos in the format it is in. If you need to mask the overscan you can do that a few different ways, the picture here is just an example of a chroma key boarder for normal NTSC.

VHS you keep them interlaced and you capture the video as such 720x576 25 frames per second. Than you make them mpeg2 videos. Use AC-3 audio 384kbs.

Not sure if you do top field or bottom field first on PAL video, I usually do top field, don't see much difference.


Attached Images
File Type: jpg ChromaKeyBoardersNTSC.jpg (24.4 KB, 123 downloads)
Reply With Quote
The following users thank deter for this useful post: jnielsen (12-22-2016)
  #16  
12-22-2016, 05:09 AM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
That amateur masking technique is more damage than I'm willing to inflict on a video.
Reply With Quote
  #17  
12-22-2016, 07:27 PM
lordsmurf's Avatar
lordsmurf lordsmurf is offline
Site Staff | Video
 
Join Date: Dec 2002
Posts: 13,503
Thanked 2,448 Times in 2,080 Posts
Quote:
Originally Posted by sanlyn View Post
That amateur masking technique is more damage than I'm willing to inflict on a video.
Are you referring to deter's mask? What's wrong with it?

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
  #18  
12-22-2016, 07:56 PM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
The mask has nothing to do with the image content. It's set up for CRT overscan (sort of), which on most CRT's is not adjustable. It discards 12 pixels of the top of the image for no apparent reason, it assumes that there are 12 pixels of head-switching noise (sometimes there are none, usually 4 or 8), it cuts off 16 pixels from each side (that's 8 lost pixels of image content from each side for no apparent reason). The resulting image is cut to 688x456, which is needlessly destructive. And it forces a lossy re-encode of a lossy encode. It's a sledgehammer technique.

Just keep on lowering that bar, folks.
Reply With Quote
  #19  
12-22-2016, 08:24 PM
lordsmurf's Avatar
lordsmurf lordsmurf is offline
Site Staff | Video
 
Join Date: Dec 2002
Posts: 13,503
Thanked 2,448 Times in 2,080 Posts
I think you're confused there.

In VirtualDub, you
- Use the Resize filter -- but don't resize, simply change the letterbox/crop to the source resolution. Don't change anything else.
- In VirtualDub filters view, with Resize highlighted/selected, Crop
- Crop by 2 pixels increment (2,4,6,8,10,etc)

After this is done, the image will re-center in the mask. If the above image was 12 per size, then it could have been 24 from the bottom and 0 from the top. Or, more likely, something like 20 and 4.

The mask may look lopsided if you ONLY cut off bad pixel zones, so balancing it is left to artistic interpretation. Most often, you'll crop/mask about 4 more from each side.

I play it safe, doing to jitter concerns (thus overscan seeping into an already-bad sectino), and thus crop 2-4 extra, just to be safe about it.

Remember that the mask is going to be invisible on virtually every TV,SD and HD both. Only the computer shows this. You can completely crop and resize for streaming, since it's probably being downsized anyway. The H/x264 encoding will harm it just as much as resizing. This is all about the lesser of evils, and varies. Sometimes you can't crop, and must leave overscan, for streaming.

If anything, deter's image show hairline noise in the left/right borders, and I'd just assume mask it too. The left/right overscan is larger than top/bottom overscan anyway. (Note: JPEG compression may be the culprit in this exact image.)

Honestly, I'd not be shocked if that was a re-post of one of my images posted here over the years.

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
  #20  
12-22-2016, 08:56 PM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
Quote:
Originally Posted by lordsmurf View Post
I think you're confused there.

Remember that the mask is going to be invisible on virtually every TV,SD and HD both. .
That's ridiculous and patently untrue for many viewers with new tv's. I'm not talking about "seeing" the mask anyway. It's not a great idea ("artistic license" or not) to cut off chunks of images. Ask the creator or an archivist or the owner how they feel about it. This is just mediocre grunt work looking for praise and justification.

The same author used the same gimmick to lacerate another already-damaged video posted by the same author elsewhere. It's obvious from looking at the video that big chunks are missing from the frames. Do what you want. But not to my videos and not to my clients'. Any tyro can send a video through cheap tricks like this and re-encode ad infinitum. I wouldn't recommend it and don't subscribe to it.

Otherwise everyone is perfectly free to do whatever they want.
Reply With Quote
Reply




Tags
capture, deinterlace, interlaced, mp4, vhs

Similar Threads
Thread Thread Starter Forum Replies Last Post
Help with Replaying MPEG2: Convert Hi8 to DirecTV Steelers__Fan Capture, Record, Transfer 3 11-15-2010 01:12 AM
Problem with mpeg2 daMaddColombian Encode, Convert for discs 5 09-20-2005 11:19 PM
MPEG2 transfer to DVD Sujit Author, Make Menus, Slideshows, Burn 3 05-06-2005 09:12 PM
MPEG2 for DBS admin Videography: Cameras, TVs and Players 0 02-15-2005 10:50 AM
Editing MPEG2 video myron Encode, Convert for discs 4 12-15-2004 08:18 AM

Thread Tools



 
All times are GMT -5. The time now is 12:19 PM