A Little Guidance with Improving VHS Captures

[Micheal81]

I tried many different pieces of software to capture some some old VHS tapes. I never could consistently get Virtualdub to capture without audio sync issues. Amarectv does the job perfectly. Now that I know how to capture without any audio sync issues, I'm figuring out what to do next.

I'm using the Roxio VHS to DVD Plus 3 USB stick and a basic 4 head VCR with a composite connection. I'm capturing at 29.97 fps, 720x480i with huffyuv video and PCM audio. I will not be spending money on a TBC, other VCR, etc. I know that is not ideal, but it is just the way it is and I'm OK with that. I want to make the video as good as possible without spending anymore money. I know the basics of using Virtualdub, Avisynth, Megui, Handbrake, and some other common tools.

I'm thinking I first need to calibrate my PC monitor. It is a HP 2009M. What is a good site or guide to use? I know doing it by eye is not ideal, but that is OK.

Next, I'm thinking I need to make sure that the colors are being captured as accurately as possible and/or use filters to correct after the capture. I need to learn how to make sure of that and how to make adjustments. I found this site http://www.trevlac.us/colorCorrection/colorTools.html Is it a good guide for doing that? I have briefly scanned it and will more carefully go over when I get a chance.

I also want to mask the edges of the video that would normally be covered by overscan. I know how to make an avs script with crop and addborders to do this. I can either do this while encoding with MeGui or I can run it thru VDub (adding any other filters needed) and resave it as a lossless huffyuv and then encode with x264 or maybe as a high bitrate MPEG-2. I know I could just crop, but I'm trying to maintain the full 720x480 frame and use anamorphic to display at 4:3 like needed. Am I going about it right?

I also am thinking about using QTGMC to deinterlace and denoise. I need to learn how to use it first. Any guides that are recommended?

Is there anything I'm missing or anything I'm not thinking right about? Thank you for the help.

Edit:

I have attached a short sample. I have been an amateur magician for around 25 years. When younger, I would record everything to do with magic that came on TV. I have over 100 hours VHS to capture. This will all be for my personal collection. I know it will take a while, but that is perfectly ok. I also know some things may not be able to be fixed, but I want to fix all I can.

[johnmeyer]

Fastest and most bullet-proof way to transfer a lot of VHS tapes is to get a VHS to DVD recorder.

However, since you are going to use your existing hardware, make sure to find the "edit switch," if your VCR has one, and make sure to use it. This turns off all the sharpening and other enhancements which are supposed to make the video "look better," but which totally screw up the signal when you are trying to capture it to another tape (the original reason for having the switch), or when you are digitizing the analog video, as you are. The difference is not subtle. This is important.

The best "low end" calibration tool that I know of is the Spyder colorimeter. I think they have versions for under $100. Since the color on NTSC video stinks, and this then gets made much worse by the VHS technology, I wouldn't get too hung up on this. Calibration is far more important for digital video and still photography.

If you want to mask the edges in order to hide the blanking signal at the bottom of the screen, the easiest (and best, IMHO) thing to do is simply overlay a black bar (i.e., a mask). There is no re-sizing involved which will avoid all sorts of other issues. I highly recommend this approach.

You never need to deinterlace, and I always recommend against doing it. Deinterlacing is the most misunderstood process in all of video restoration. Deinterlaced video is NOT better than interlaced, and the act of deinterlacing always degrades your video. Always. If you are going to watch on a TV set, don't deinterlace.

Your capture looks pretty good, although a TBC would indeed improve the result. On the portion of your sample that was actually taken with a video camera (the head interview, as opposed to the 8mm film segment), you'll see that the vertical sections on the wallpaper behind him wobble a little. Having said that, it actually isn't too bad, so if most of your video is of the same quality, you won't be too bad off without the TBC.

The biggest issue, of course, is the tearing at the left side of the screen. This looks like a tracking issue. Make sure you know how to reset the auto tracking. With many VCRs, this is done by simultaneously pressing both the channel up and channel down buttons, on the VCR itself. Then, if the problem doesn't disappear, try adjusting tracking manually. On many VCRs this is done by pressing the channel up and channel down buttons on the VCR while the tape is playing. There is no simple post-production fix for those tears, and they are pretty significant in your sample.

As for denoising, I have posted my VHS restoration starting script (you always have to tune any script) many times. It uses an old plugin, CNR, to reduce the chroma noise, and MDegrain2 for the noise. I just tried it on your sample and it worked well.

The other big problem with your video, and it is a little unusual, are the horrendous halos around everything: his head, his shirt, etc. Even if you haven't engaged the edit switch, this is a LOT of edge enhancement. You need to look every carefully at your VCR and also your capture settings to see if you have introduced this. One reason I always recommend DV capture over capture cards is that you don't introduce these kinds of problems. Capture cards can produce really good results, but there are far more ways to go wrong, because of all the codec choices, and all the settings available for both the codec and the capture software. You already found out how bad things can get with the audio sync issues introduced by the settings you used in Virtualdub. So, see if you can adjust something to get rid of those halos. Your capture looks pretty good in many other ways, but these do major damage to the resolution, and make everything look out of focus.

[johnmeyer]

P.S. I thought at first that was Mark Wilson, who I used to watch back in the 60s, but I looked at him current day (he is still alive), and it isn't. Who is the guy being interviewed?

[LMotlow]

@johnmeyer: good points.

BTW, that video sample isn't interlaced. It's telecined. The first part with the boy and baton is old 18fps or 20fps movie film with dupes, blends, and telecine.

[Micheal81]

P.S. I thought at first that was Mark Wilson, who I used to watch back in the 60s, but I looked at him current day (he is still alive), and it isn't. Who is the guy being interviewed?

The sample is from a TV show from 15-20 years ago called Before Stars Were Kids. The boy is David Copperfield and the older man is his father.

[Micheal81]

@johnmeyer: good points.

BTW, that video sample isn't interlaced. It's telecined. The first part with the boy and baton is old 18fps or 20fps movie film with dupes, blends, and telecine.

This is the part that I don't understand apparently. It was shown on TV many years ago and I thought that meant it has to be interlaced (480i)? Media Info and Gspot say it is. I get that the older 18fps or 20fps had to be modified to be broadcast, but wasn't the program broadcast as interlaced? Isn't it capture interlaced? Am I misunderstanding something here?

[johnmeyer]

BTW, that video sample isn't interlaced. It's telecined. The first part with the boy and baton is old 18fps or 20fps movie film with dupes, blends, and telecine.

It depends on your definition of interlaced. Yes, it is telecined, but that word simply means that frames or fields are repeated in order to be able to play the 16 fps film at the correct speed when shown on a 29.97 fps TV. However, the way in which telecine is usually done -- and the way it is done on this clip -- is to repeat fields rather than frames. It is better to do it this way, rather than repeating entire frames, because it reduces the judder. However, when the telecine process repeats just one field, and then goes onto the next one, you end up with a frame which has fields from two moments in time.

This is the definition of interlaced: the odd and even fields in a video frame come from two moments in time.

So, in that sense, the first 3/4 of this clip, which come from old amateur film footage, is interlaced because there are many frames that are made up of fields from two moments in time (easy to see if you walk through the clip using separatefields). If I wanted to edit the film portion, you are correct that I would use inverse telecine tools rather than deinterlacing tools in order to recover the original 16 fps progressive footage. However, any attempt to modify this footage while treating it as progressive will cause all sorts of problems. For instance, if I wanted to apply MDegrain2 denoising, I would have to first separate the video into fields, and then , as a two distinct operations, I'd first denoise the odd fields, followed by the even fields. When finished with these two operations, I'd weave them back together.

[usually_quiet]

I'm thinking I first need to calibrate my PC monitor. It is a HP 2009M. What is a good site or guide to use? I know doing it by eye is not ideal, but that is OK.

Since you don't want to pay for equipment to improve your VHS captures, I assume that you don't want to pay for the software and hardware to do a thorough calibration either.

If you do a search you will find free calibration software you can try. If you have not already used any, Windows built-in monitor calibration software would be a place to start. It is at least easy to use.

I have a different inexpensive LCD monitor and found it difficult to try to calibrate it with better software due to the awkward controls and limited settings.

[Micheal81]

You never need to deinterlace, and I always recommend against doing it. Deinterlacing is the most misunderstood process in all of video restoration. Deinterlaced video is NOT better than interlaced, and the act of deinterlacing always degrades your video. Always. If you are going to watch on a TV set, don't deinterlace.

The biggest issue, of course, is the tearing at the left side of the screen. This looks like a tracking issue. Make sure you know how to reset the auto tracking. With many VCRs, this is done by simultaneously pressing both the channel up and channel down buttons, on the VCR itself. Then, if the problem doesn't disappear, try adjusting tracking manually. On many VCRs this is done by pressing the channel up and channel down buttons on the VCR while the tape is playing. There is no simple post-production fix for those tears, and they are pretty significant in your sample.

As for denoising, I have posted my VHS restoration starting script (you always have to tune any script) many times. It uses an old plugin, CNR, to reduce the chroma noise, and MDegrain2 for the noise. I just tried it on your sample and it worked well.

The other big problem with your video, and it is a little unusual, are the horrendous halos around everything: his head, his shirt, etc. Even if you haven't engaged the edit switch, this is a LOT of edge enhancement. You need to look every carefully at your VCR and also your capture settings to see if you have introduced this. One reason I always recommend DV capture over capture cards is that you don't introduce these kinds of problems. Capture cards can produce really good results, but there are far more ways to go wrong, because of all the codec choices, and all the settings available for both the codec and the capture software. You already found out how bad things can get with the audio sync issues introduced by the settings you used in Virtualdub. So, see if you can adjust something to get rid of those halos. Your capture looks pretty good in many other ways, but these do major damage to the resolution, and make everything look out of focus.

I know that deinterlacing technically degrades the video. I do think it can sometimes make it look better. I will probably test out QTGMC and see what I think. It will probably only ever be watched on a PC or mobile device. I will try out your script. I will also try tweaking any settings to see if I can get rid of the halos.

How do I get the color, brightness, contrast, etc. correct or as close as I can before capturing? Is this site http://www.trevlac.us/colorCorrection/histo.html a good guide for correcting after capture? Here is a screenshot of the Device settings. I'm assuming these are the controls to use for adjusting before capture.

[johnmeyer]

I know that deinterlacing technically degrades the video. I do think it can sometimes make it look better.

That's somewhat true. If you feed interlaced footage into a video player (usually on a PC) that doesn't know how to deinterlace, then the results may look better if you deinterlace. In that sense, it can make it look better, even though the video has been degraded, both by removing half of all the temporal information (would you remove half the spatial resolution, i.e., throw away half the pixels?), and by creating artifacts from whatever algorithm is used to estimate the new field that is being moved in time to match the other field.

Most video players, upload sites, and certainly your TV, are capable of deinterlacing if it is needed, and usually (not always, of course) do as good a job as you can do.

[PuzZLeR]

Good points made here, but I will add.

As per de-interlacing, don't, unless a display calls for it. And even if you do, 1) use a good de-interlacer like QTGMC, 2) de-interlace at double the fps for much better quality, and 3) most important, KEEP the Source separately anyway and keep it interlaced to always keep your options open at the highest quality possible for the future.

As per VirtualDub going out of sync, it does not happen to me, so can't help you correct it, but I know it happens to some people, and much may depend on their system, PC or settings. I do still recommend it for high-end capturing. In your case, you can still use it to capture the video, and you can capture the audio separately, such as with Audacity, and simply mux to sync with VirtualDub later.

I know it's more work, but not really much. Audio capture is simple, and unattended work. Just feed the system and come back when it's done. And I do it anyway since some VCRs I have have higher quality audio outputs, but others do better with video - even with the same tape. You may want to test this if you have other VCRs to make another capture worth it.

My biggest quality booster however is in multiple captures and median methods if you're willing to do the extra time. It will remove all sorts of random errors, take care of frame drops, chroma dancing, and calm many snowstorms. You don't need extra hardware, or need to pay for any software, and the difference between captures in the end is a dramatic improvement.

More on it here:
https://forum.videohelp.com/threads/362361-Median%28%29-plugin-for-AviSynth

[johnmeyer]

My very first attempt at VHS restoration, more than a decade ago, was the multiple capture technique PuzZLeR refers to. When I did this in 2003, it was before many of the nifty software tools we now use were available.

The good news is that when everything is set up correctly, averaging multiple captures produces noise reduction that has virtually no side effects, and the results look really natural.

However ...

If your tape has any issue that causes it to lose sync, you have to go through your NLE (if that's what you use to combine the multiple captures), find the point at which sync between each capture is lost, and re-align. This same issue can also cause spatial errors. The extreme example is if the tracking isn't working perfectly, and one capture begins to lose tracking and starts to get snowy or break up. So, for this to work, you have to have a really stable tape, without any major playback problems.

You also need to understand which VHS quality problems this technique will improve. Multiple captures will reduce spatial noise (those annoying little random dots) and may smooth some types of time base errors and, maybe, some types of dropouts. However, it won't help for flagging, tearing, tracking issues, OTA noise (like ghosting or temporary chroma sync errors) and dozens of other artifacts common on VHS tapes. [/edit]It will also only reduce noise that is introduced during playback: if the noise is on the tape to begin with, it won't touch it.[/end edit]

Finally -- and this is obvious, but still worth mentioning -- it takes a really long time to do this! If you have a two hour tape and you want to capture it three times, that will take six hours! I'll let you do the math for a six hour tape. The noise reduction improves as the square root of the number of captures, so the incremental improvement after three captures is pretty small. If you are doing this for hire, and it is for the National Archives, or if millions of people will be watching, then this technique is definitely worth the investment in time. Otherwise, you'll have to decide if you can afford the hours needed to achieve the result.

It has been twelve years since I first did this, and I described my work in this long post in the Sony Vegas forum:

My "ultimate" VHS tape restoration recipe

That thread title is full of the hubris that comes from not having yet been humbled by doing enough restoration projects to learn all the pitfalls and, eventually, to realize that there is no "ultimate" restoration recipe.

[PuzZLeR]

Interesting post, and interesting guide. Allow me to comment since someone may wish to go this direction, so I share some of my methods too.

The good news is that when everything is set up correctly, averaging multiple captures produces noise reduction that has virtually no side effects, and the results look really natural.

Care is needed here as "averaging" is not synonymous with median methods.

Averaging has different advantages, and different problems. But again, if everything is setup properly with median methods, there are indeed virtually no side effects as well, however...

However ...

You said it.

If your tape has any issue that causes it to lose sync, you have to go through your NLE (if that's what you use to combine the multiple captures), find the point at which sync between each capture is lost, and re-align. This same issue can also cause spatial errors. The extreme example is if the tracking isn't working perfectly, and one capture begins to lose tracking and starts to get snowy or break up. So, for this to work, you have to have a really stable tape, without any major playback problems.

Yes, care is needed when merging the captures. I look for such errors, like even tape drops, and it's especially more problematic when there are "seams" in the recording (like a TV recording that's been stopped and restarted, or recorded over at a point, etc.)

Such problems will cause lots of sync with your multiple captures. You need to make sure they're synced throughout otherwise your final capture will have horrible artifacts and look like it's dissolving underwater.

The way I correct this is with AviSynth. First of all they have to start at the same time - you can adjust the starting point with Trim(). Then, to view them together on my monitor I use StackHorizontal(), and scrub through looking to see if they line up, particularly targetting scene changes which easily show if they're not lined up. If not aligned, I back up to see where the problem is.

If frames were dropped, I correct it with duplicateframe() where needed. Conversely, to eliminate extra frames, or inserts, use deleteframe().

This is a bit time consuming to learn to do this manually, but very easy and effective when you get used to it, and haven't investigated many automated methods really. But yes, if you're dropping a frame every five minutes, a two hour production can take a while to fix, and an automated method would be better here.

Regardless, the much higher quality result is worth it from multiple captures, especially for an important production.

You also need to understand which VHS quality problems this technique will improve. Multiple captures will reduce spatial noise (those annoying little random dots) and may smooth some types of time base errors and, maybe, some types of dropouts. However, it won't help for flagging, tearing, tracking issues, OTA noise (like ghosting or temporary chroma sync errors) and dozens of other artifacts common on VHS tapes. [/edit]It will also only reduce noise that is introduced during playback: if the noise is on the tape to begin with, it won't touch it.[/end edit]

True. Only random errors are fixed. In other words, occurences that happen to be different on each playback, like random snow, and other glitches. It does correct some tracking bursts, even some tearing, but, again, ONLY if they're random, and likely ONLY if it just so happened on one of your playbacks, where correction is similar to handling drops or inserts.

If something is exactly the same each time you play it, whether it's something your VCR is doing repeatedly on a scene, or some scratch over a frame on your tape, that's where the magic ends. Median methods will leave it there. Then again, this is also a good thing as median methods will not remove any detail either from your original recording.

And, again, what it does correct, is plenty, and does dramatically increase the quality of the final production.

Finally -- and this is obvious, but still worth mentioning -- it takes a really long time to do this! If you have a two hour tape and you want to capture it three times, that will take six hours! I'll let you do the math for a six hour tape. The noise reduction improves as the square root of the number of captures, so the incremental improvement after three captures is pretty small. If you are doing this for hire, and it is for the National Archives, or if millions of people will be watching, then this technique is definitely worth the investment in time. Otherwise, you'll have to decide if you can afford the hours needed to achieve the result.

Timewise yes, work, not really.

The first capture is always the most work, setting it up, making your adjustments, proc amp, which VCR works best, etc. This is something you do anyway if only capturing once. The second, third, etc, is just rewind, re-capture, walk away and come back at a known time. I've even gone out and did errands, or went to work, during subsequent captures.

Yes it takes time, but I've managed to fit it in somehow. Then again, so's VideoHelp. Labor of Love.

As for how many captures, 3 is almost always good enough, especially for the cleaner tapes (and still a dramatic improvement over just one capture). I use 5 captures for super-important stuff, or maybe some particularly troubling tapes. (The number must be odd for median methods.)

Actually, the real time consuming part is setting them up, and merging them, after a particularly difficult tape that had lots of recording "seams", lots of drops, or inserts, instabillities, etc. Mercifilly, this is a minority tape for me.

[johnmeyer]

Care is needed here as "averaging" is not synonymous with median methods.

Averaging has different advantages, and different problems.

Ah yes, my inability to finish a statistics class comes back to bite me again. I was pretty good in calculus and other math subjects, but never connected with statistics. Also, I received similar criticisms and suggestions in that old 2003 thread I linked to. These comments tried to help me find the proper way to weight the multiple captures.

There were many really good, insightful comments in that thread that are worth understanding if anyone is going to use this multiple capture method, and I highly recommend skimming through the posts.

As a side note, I miss the quality that old Sony forum had back then. It has gotten so pathetic, that I don't even post there anymore, after almost 12,000 posts spanning over a dozen years. I learned a lot there before it started going downhill about four years ago.

[Micheal81]

Thanks for all the replies. I will go over them more carefully a little later tonight. I am going to think more about how I want to do this. No matter what, I'm not putting anymore money into it. The multiple capture thing sounds like it may come in handy for some of the more important tapes. I will not do them all that way. It would take too much time. I have more than 100 hours of tape.

Right now I'm thinking that I just want to keep it pretty basic.

1. Calibrate my monitor as good as I can with free tools.

2. Continue to use Amarectv to capture since it has no issues with audio sync. If I can get Virtualdub to work after a little more tinkering, then maybe I'll use it.

3. Figure out how to adjust the brightnesss, sharpness, color, etc before/during capture as close as possible.

4. Figure out how to make any more required adjustments to color, brightness, sharpness, chroma, luma, etc on the lossless capture.

5. I'll probably not deinterlace. I will probably denoise, so I'll have to find a good filter for that. I'll also detelecine if necessary.

6. Encode with x264 or to a high bitrate MPEG-2 which I already know how to do.

7. Be glad to be done.

[ajk]

Such problems will cause lots of sync with your multiple captures. You need to make sure they're synced throughout otherwise your final capture will have horrible artifacts and look like it's dissolving underwater.

The way I correct this is with AviSynth. First of all they have to start at the same time - you can adjust the starting point with Trim(). Then, to view them together on my monitor I use StackHorizontal(), and scrub through looking to see if they line up, particularly targetting scene changes which easily show if they're not lined up. If not aligned, I back up to see where the problem is.

I have been playing with a version of Median() that automatically aligns the input clips up to a selected offset. It slows down the processing more the higher you set the search radius, so it's still a good idea to roughly match the clips, but it certainly saves time on the manual frame-by-frame fine tuning. I can post a trial version later if you want to give it a try.

I ran out of patience with some of my tapes, no matter what I do there's always a bit of misalignment somewhere with 5 captures

[PuzZLeR]

Care is needed here as "averaging" is not synonymous with median methods.

Averaging has different advantages, and different problems.

Ah yes, my inability to finish a statistics class comes back to bite me again. I was pretty good in calculus and other math subjects, but never connected with statistics. Also, I received similar criticisms and suggestions in that old 2003 thread I linked to. These comments tried to help me find the proper way to weight the multiple captures.

Yes, there was some confusion between the two, average and median, as I noticed as well in that thread. And yes, I can see your understanding of calculus even in this thread as you described the asymptotic nature of incremental benefits of subsequent captures.

But just for the readers, the difference between average and median in this context is:

Say you have five different captures. Let's take five pixels at a common point, at a given frame, and a certain exact point in the cartesian (x,y), and to make it simple, say the pixels have the color 2, with one outlier: 2,2,2,7,2. The 7 in this case could be a white spot randomly occuring on the 4th capture.

Averaging them gives: 3. In other words, you've changed the color entirely, even made the 2, which is likely the "normal" pixel, brighter. If you use your imagination you can see how this could be disasterous, particularly on flat backgrounds.

Median gives: 2. You only kept the "middle" pixel. You completely disregard all others, and hence, you eliminate the 7 which was the white spot, and not change the 2. Very different result.

There were many really good, insightful comments in that thread that are worth understanding if anyone is going to use this multiple capture method, and I highly recommend skimming through the posts.

As a side note, I miss the quality that old Sony forum had back then. It has gotten so pathetic, that I don't even post there anymore, after almost 12,000 posts spanning over a dozen years. I learned a lot there before it started going downhill about four years ago.

Oh, this can be more rule than exception on the Internet.

I have seem many sites that I once loved go south. It could be different ownership, different users, spam, bots, trolls, or implementation of social media features (which can suck), etc. It's always sad when this happens. Worst of all, sometimes it can take weeks or months of frustration to finally figure out that the good ol' days are gone forever.

Interestingly, heading towards my 10th year here on VideoHelp, everything seems almost exactly the same as it was during my first post. Video tech has changed, but the Forum still keeps delivering the same way. I am grateful to the Admins for that.

[PuzZLeR]

Such problems will cause lots of sync with your multiple captures. You need to make sure they're synced throughout otherwise your final capture will have horrible artifacts and look like it's dissolving underwater.

The way I correct this is with AviSynth. First of all they have to start at the same time - you can adjust the starting point with Trim(). Then, to view them together on my monitor I use StackHorizontal(), and scrub through looking to see if they line up, particularly targetting scene changes which easily show if they're not lined up. If not aligned, I back up to see where the problem is.

I have been playing with a version of Median() that automatically aligns the input clips up to a selected offset. It slows down the processing more the higher you set the search radius, so it's still a good idea to roughly match the clips, but it certainly saves time on the manual frame-by-frame fine tuning. I can post a trial version later if you want to give it a try.

I ran out of patience with some of my tapes, no matter what I do there's always a bit of misalignment somewhere with 5 captures

Hey AKJ, thanks for chiming in, and thanks for the plugin - I use it all the time. Awesome! (I even have long forgotten about the older cruder methods of the past.)

Frame drops occur almost all the time, even if VirtualDub does not report them, especially when they come from analog shortcomings, such as tape drift, tape stretch, VCR heads, etc. Even the weather can play a role I believe.

If it's an odd single drop, or even a couple, every 60+ minutes, on five captures, you can get away with ignoring it. Any "damage" is frivolous. However, if it happens frequently, like from a certain VCR or tape, you HAVE to go through the video to align it, and it can be time-consuming to scrub through manually in such cases just to zero in on those drops to implement an articifial duplicate to reset sync (or to delete an insert/dupe in some cases).

I would love to beta this new plugin when you're ready. I say you should upload it on your thread (if readers are interested - the one I linked to earlier). At any rate, I will watch for it, and hopefully I, and any others, can give it a good test.

[jagabo]

Most video players, upload sites, and certainly your TV, are capable of deinterlacing if it is needed, and usually (not always, of course) do as good a job as you can do.

No TV, Blu-ray player, or other device can deinterlace as well as QTGMC() at this point in time. The real issue is that if you deinterlace with QTGMC now you will be locking in the quality of that deinterlacing. If TVs and other devices get better in the future you'll be stuck with your QTGMC flaws. So if you deinterlace now, keep a copy of your interlaced source in case you need it in the future.

[PuzZLeR]

So if you deinterlace now, keep a copy of your interlaced source in case you need it in the future.

+1 For better or worse.

[johnmeyer]

Say you have five different captures. Let's take five pixels at a common point, at a given frame, and a certain exact point in the cartesian (x,y), and to make it simple, say the pixels have the color 2, with one outlier: 2,2,2,7,2. The 7 in this case could be a white spot randomly occurring on the 4th capture.

That "2,2,2,7,2" example you gave does a great job of making the point. I understand completely why you don't want to give any weight to "outliers." While there might be a more sophisticated algorithm or function, I can see where median would work well in many situations, and it is extremely easy and fast to calculate. Also, it can completely eliminate the noise pixel. As a result, it would seem to me that the increment in noise reduction that you would achieve by doing more captures might decline much faster than the square root function that someone mentioned in that Vegas thread. Once you have enough captures to "vote off the island" those pixels that don't conform, additional captures might not change anything.

However, the math required to make this determination is precisely where I failed with statistics: I just couldn't get my head around these sorts of problems, and could never set up the problem.

BTW, I read the thread you linked to, as well as the parallel thread AJK started over in doom9.org, and it looks like his median function works really well for this situation.

I have seem many sites that I once loved go south. It could be different ownership, different users, spam, bots, trolls, or implementation of social media features (which can suck), etc. It's always sad when this happens. Worst of all, sometimes it can take weeks or months of frustration to finally figure out that the good ol' days are gone forever.

In the case of the Vegas site, which is designed for the support of a specific product, rather than simply being a place to discuss video editing, another reason for the decline is the horrendous Sony support of both the product and the site. The site's decline has also been hastened by Sony's abysmal failure to provide any useful enhancements in the past four releases over the last six years, despite a huge backlog of requests from hard core users. As a result, they have lost market share, and the pros have gone elsewhere. I'm staying with the product because I'm too old to learn anything else, and also because the scripting built into Vegas is still absolutely brilliant and saves me hours on every single project. I can't live without it.

[Micheal81]

Is it really necessary to use inverse telecine on telecined material?

Can I use QTGMC on telecined material?

If I don't deinterlace, will denoise filter be affected by telecined material?

I'm just trying to figure out if I really need to worry about whether or not any of the material is telecined. If I really don't have to worry about that, then I'll just use QTGMC on it for deinterlacing and denoising. I don't care if 1 day something better comes along.

I'm also trying to find out what other filters to use after cpature to adjust color, brightness etc.

[manono]

Is it really necessary to use inverse telecine on telecined material?

Necessary? Interesting choice of words. No, it's not necessary. You can do whatever you like. But it's well worth it for encoder efficiency, if for no other reason. Imagine a quality comparison between the same two videos both encoded using the same bitrate. One's progressive 23.976fps and the other's 29.97fps where 40% of the frames are interlaced, with the understanding that interlaced frames require a 50% higher bitrate for the same quality as compared to progressive frames. I've worked with lots of caps where the fields didn't line up properly and I've wound up bobbing with QTGMC before then decimating (removing duplicate frames).

Can I use QTGMC on telecined material?

Again, do whatever you like. I mentioned I do it on occasion. But I'm sure to remove the duplicate frames afterwards if it had been telecined. Some people use QTGMC mostly for its filtering properties - the denoising, smoothing, and sharpening it does.

If I don't deinterlace, will denoise filter be affected by telecined material?

If it's been telecined, IVTC it first and then run your denoising filter. But some denoisers have an interlaced mode allowing them to denoise interlaced material. Or you can bob your video first, denoise, and then reinterlace it. That's not a good idea, though, unless it was shot on video.

I'm just trying to figure out if I really need to worry about whether or not any of the material is telecined.

It takes 30 seconds or less to figure it out. That shouldn't be a consideration, especially when you should be spending a whole lot more time than that studying your videos to decide what needs doing and how to do it.

Edit: Apparently the video is a mix from different sources. I was trying to answer questions specific to telecined tape material.

[LMotlow]

I'd go with johnmeyer's earlier suggestion to consider the whole sample interlaced. It has duped-field, telecined, and interlaced sections here and there. He suggested his version of MDegrain for interlaced video that works just fine using SeparateFields() and then re-weaving when you're finished. Or you can spend the next five months disassembling, denoising, and reassembling that video scene by scene with all the mixed media in it. The best denoisers are in Avisynth, but Virtualdub has some good tweaks. The most effective will use progressive or field-separated video.

Levels are adjusted during capture with brightness and contrast controls. This was already been mentioned elsewhere. Brightness controls blacks, contrast controls brights., They interact, so fiddle a bit to get your input values as close as you can inside RGB 16-235 borders. There might be some over- or undershoot now and then, but it's OK as long as you don't blast levels out of the ballpark altogether. This is best done using a histogram. I don't think Amarectv has a capture-time histogram or other graph, maybe those who use it can tell you. If not, you'll have to make quick caps here and there and test them later, then settle on your settings and use it for the whole tape. Or get complicated and create a GraphEdit setup. Scenes will vary, but once you get an idea of the worst-case areas, you're safe. You can make finer corrections later.

Unless a tape has a really serious but consistent color cast, you're driving yourself nuts trying to correct VHS color during capture. The only practical capture-time color control is a Hue control, but it's limited and might affect other colors in ways you won't like. Once those colors are captured, it's there unless you recapture or work overtime to un-correct your corrections. There are dozens if not hundreds of Avisynth and VirtualDub color filters you can use later. Most of them are more sophisticated than simple Hue controls (but even poor little ole Hue has its place, too). Right now you're getting ahead of yourself. Best to make a single capture to start, then get some help on fixing color, contrast, etc., and learn a step at a time. Then get the rest of your caps and go like crazy.

The problem you have at the outset is that if you capture using the same setup as before, you'll get the same results.

[Micheal81]

The problem you have at the outset is that if you capture using the same setup as before, you'll get the same results.

I'm ok with the setup I have. I know it is not ideal. I can live with it though. I do want to make any improvements I can with denoising, brightness, contrast, etc with filters and the software that I already have.

[jagabo]

Is it really necessary to use inverse telecine on telecined material?

It's not absolutely necessary.

Can I use QTGMC on telecined material?

Yes you can. But if the video was cleanly telecined (the sample you provided wasn't) then IVTC will usually give better results. You will also have a lower frame rate in the end and that will require less bitrate in your final encoding.

If I don't deinterlace, will denoise filter be affected by telecined material?

It depends on the denoiser. Any denoiser with a 2d component will blend parts of the two fields together ruining them.

I'm also trying to find out what other filters to use after cpature to adjust color, brightness etc.

There are many many filters for that. Tweak(), ColorYUV(), Levels(), the SmoothAdjust versions of all those, RGBAdjust(), etc.

[johnmeyer]

If I were being paid $$$, I'd separate the combined video and film capture into two separate videos: the film parts and the video parts. I'd then IVTC the film parts and restore them. I'd do the same, using a different set of denoising tools, with the video portions. At the end, I'd apply telecine to the film portions and then join it up with the 29.97 interlaced video portions. You'd end up with the same thing you have now: mixed film and interlaced video, with the film portions telecined.

Why?

Because restoration tools go bonkers (technical term) when fed telecined material. Any temporal tool (something that looks at more than one frame) sees the telecine repeats as a frame with absolutely no noise because there are no differences from the previous frame. So, nothing happens, and you get no noise reduction on that frame. This is not what you want. I have no idea what QTGMC might do with telecined material. Read the doc, and do a search on the long QTGMC thread in doom9.org to see if someone else has already asked the same thing. It is an amazing script and the author may have provided some settings for this situation.

But, of course, in the real world we don't always have time to do things the "right" way.

As for deinterlacing, why do you want to do that? I always recommend that people not deinterlace unless they are re-sizing, but for this material there is another reason to avoid it: if you deinterlace the video portions and end up with either 29.97 interlaced or 59.94 progressive, what are you going to do with the 16 fps film portions? If you don't IVTC them, then they will still be 29.97 telecined/interlaced, and any attempt to deinterlace THAT will produce a fine mess, I would think.

As already suggested, a reasonable fall-back position is to treat everything as interlaced, but realize that the film portions are not going to be improved by as much as you might like.

[manono]

At the end, I'd apply telecine to the film portions and then join it up with the 29.97 interlaced video portions. You'd end up with the same thing you have now: mixed film and interlaced video, with the film portions telecined.

By 'telecined' I'll assume you mean hard telecine. I'd do it (and have done it) differently - encode the progressive parts progressively, applying soft pulldown afterwards, and encode the interlaced parts as interlaced 29.97fps. Then rejoin the different parts during authoring.

[jagabo]

encode the progressive parts progressively, applying soft pulldown afterwards

But you can't use soft pulldown with 16 or 18 fps sources.

[manono]

But you can't use soft pulldown with 16 or 18 fps sources.

As I said earlier, I haven't looked at the sample, and there are many other kinds of mixed framerate sources. You're right, of course, but one can easily interpolate up to 19.98fps for encoding and later soft pulldown for DVD.