Hello everyone longtime reader, first time poster! I see how much information there is to everyone and at times it's a blessing but also can make things confusing. I know you guys are here to help. So here will be my first question that should be simple enough, and I will then elaborate on what I want personally and if it's a good idea or not. I think you guys can answer and just clarify my workflow.
My intended workflow:
1. Capture VHS to digital format. I am only concerned with the VHS tapes that were recorded 25 years ago on my dad's VHS camcorder.
1a. The digital format will be in UT Video AVI Lossless. So with that 1 gigantic file (depending on video length I guess) I would like to create a master copy. I am capturing the VHS in 720x480 NTSC 29.97fps. I would like to have that 1 master copy so I can do wtv i want with it in the future.
1b. From that Master copy, I would like to make a viewable video file for TV or Computer viewing or on the smartphone. I would like an h.264 MP4 file. That's it, that's all.
A little backstory:
I created 10 DVDs in 2005 from 10 VHS tapes (Tape 1 through 10) and they turned out alright. But i'm an idiot and I deleted the capture files and only kept the mpg file and the DVD files to conserve disk space. I'll come back to this later.
September 20 2022 NOW!:
So I decided to tackle the rest of the VHS tapes seeing how technology improved and how disk space is not really an issue. With my "Intended Workflow" in mind, I started capturing 3 more VHS tapes (Tapes 11 12 and 13). So I have 3 new Master copy AVI files. But I want to spruce them up a bit and do some cleaning. I want to remove the bottom noise and left and right side need about 4 pixels cropped on each side. Thinking about it I think I have 2 options that I can do. This is where i would like your opinions or take on this.
Option 1 ( which I kind of started and need some input on it)
I captured using virtualdub2 (Tape11a.avi - the master copy). I made a 2nd file Tape11b.avi that I deinterlaced with virtualdub2 using yadif and doubled the framerate to 59.94 fps (not sure I like this yet but it is smooth). I cropped the left, right and bottom and then resized back to 720x480 which I read i shouldn't do. I also increased audio by 3db. From this Tape11b.avi I can now proceed to make an mp4 file with h.264. I would basically have 3 files for each VHS tape (master avi, processed avi, mp4 for viewing)
Option 2 (I think this is what I want to do now but need input)
With the already captured Tape11a.avi file - the master copy, couldn't I just do crops on the master avi file but restore the 720x480 back with letter box black borders? I don't have much to take off I think I would do 8 left 8 top 8 right and 8 bottom. That way I am masking right? and not affecting quality of the master? the black border when viewed on the tv or computer screen looks great. Then I would proceed to deinterlace, increase audio and make my mp4 file. I would have 2 files for each tape, the master avi which is cleaned and the mp4 file.
Is option 2 the best route??? With option 1 I also need some clarification on the 2nd avi file (Tape11b.avi) that is double the framerate at 59.94. My mp4 will be encoded at 29.97, is that further reducing quality???
+ Reply to Thread
Results 1 to 30 of 48
Last edited by ngnaw; 25th Sep 2022 at 16:39.
Changes I would make to your workflow.
1. Forget about lossless. This is massive overkill for VHS. Instead, capture to an "intermediate" codec. Many years ago Cineform was the professional standard, but I think there are more modern versions. Do a few tests, just to convince yourself, but I defy you to find any difference between Cineform at one of its higher quality setting and a true lossless codec. The difference in file size, however, will be massive. Even if you don't care about filling up hard drives, larger files tend to slow down every step of your workflow.
2. Make sure you are using a REALLY good VCR. All your other steps are a waste of time if you don't have good equipment. I would say that 90% of what makes a great VHS final product is the capture, and 10% is all the stuff you do later on. OK, that's a bit of an exaggeration, but not by much.
3. When you do the capture, make sure your deck is set with the Edit Switch turned on, and all processing turned off. If your deck doesn't have an Edit Switch, re-read #2 above.
4. I am not a big fan of deinterlacing unless you are resizing. You are talking about cropping off the head noise and then re-sizing upwards so if you do that, you do indeed need to deinterlace. Another option is to just mask the noise with pure black, and don't re-size. You avoid the re-sizing (quality loss), you avoid deinterlacing (another quality loss), and you speed up the workflow. I doubt anyone will notice 3-4 lines of extra black at the bottom of the screen.
Always do you own tests, and never assume that "turning it all the way up to 10" is going to provide a discernible improvement.
Awww I was testing Option 2 and I screwed up... In VirtualDub2 using crop and resize with letterbox is not what I intended, I should be using Fill and tick letterbox so I can add the black borders... They both achieve what i want but Fill is not destructive. Is this correct?
However, if I crop and then replace the pixels back with black with resize letterbox back to 720x480, doesn't that not stretch and not alter the image? Or does it? I see how fill is basically a mask on top of what I don't want to see. I just want an explanation about the crop and resize and how it affects quality of the capture.
Like i can picture how if you crop and then resize by just stretching the image to fit the 720x480 frame then ya it messes things up. But cropping and then adding black borders with resize, isn't that ok? Or just stick with Fill lol?
Last edited by ngnaw; 25th Sep 2022 at 23:04.
If you merely overlay black pixels, your video is still 720x480. There is no stretching and no need for resizing. If the video changes dimensions, then you are not doing the overlay properly.
Resize = Quality Loss
Overlay = Same quality (other than the loss when re-encoding)
Deinterlace = Quality Loss
Overlay is just like putting black tape on the bottom of your TV screen: the image size and aspect ratio are unaffected, but you no longer see the head noise at the bottom of the video.
Originally Posted by John Meyer
Even if you don't care about filling up hard drives, larger files tend to slow down every step of your workflow.
I did try to find the cineform codec; the only site I could find was a spam-laden eastern-european site. I note that it isn't hosted on Videohelp. That throws up a warning flag for me...
Last edited by Alwyn; 26th Sep 2022 at 02:46. Reason: Cineform codec info added.
If I wanted to capture in a lossy format, I would go for some M-JPEG codec that gives you control over how much quality loss you are willing to accept (most professional codecs don't have this option or are very limited in it). But I recommend to stick with a lossless codec. If you are going through several iterations during the filtering stage, lossless gains more of an advantage with each iteration.
Cineform can be anywhere from 80% to 120% the filesize of utvideo on content such as VHS. Yes, sometimes larger but always lower quality (but quality loss not really visible at higher quality levels). There is 6 quality levels in the SDK versions (original ones only had 5 with filmscan2 begin the highest). One negative is no AR flagging, nor interlaced encoding or flagging in most of the free SDK versions (But those are similar issues with most other lossless AVI codecs). For 4:2:2 interlaced it's usually not an issue if the AR and field order interpreted correctly in the recieving application, because the chroma is horizontally subsampled, not vertically. One benefit of cineform YUV, is it's usually treated as YUV in NLE's instead of being treated as RGB compared to "lossless" YUV codecs. There is a cineform version based on the SDK bundled in vdub2 (cineform native) you can test it yourself
Quick test 720x480 8bit 422 from VHS capture .
UT video 72Mb/s
Cineform Filmscan 3 81.2Mb/s
Cineform Filmscan 2 78.3Mb/s
Cineform was sold to GoPro over a decade ago. GoPro then included this software, which previously sold for over $1,000, into their "Studio" software line which you could download for free. I just checked, and you can no longer find this on the GoPro site because they have discontinued Studio. Therefore you will have to use one of the codecs you described.
One thing for sure: if your files are larger than 50 GB for one hour of standard definition video, you are wasting disk space for a visually imperceptible gain in quality.
If you are going to put this much effort into a project, you need to be smart about the decisions you make.
If you are a perfectionist, it would not be my first choice.
Yes, double rate deinterlace such as yadif2x to 59.94p such as the OP mentioned in the 1st post.
"Inferior to what" ? When you view it (if it was kept interlaced), it's being deinterlaced somewhere. And there are artifacts in the interpolated fields when you deinterlace. That's the key point.
Yadif wouldn't be my first choice either, because it's not a good deinterlacer. The interpolated fields are not good, and there are artifacts. I would consider it pretty average, typical of what you see most TV' s. Similar artifacts. Which is why people resort to higher quality deinterlacers like QTGMC, if they don't have a higher end TV with better processing chips
The decision to deinterlace some version separate from the "master" should be partially based on what the intended target is, or how it's being viewed - ie. if your display chain has a better or worse deinterlacing capability than what you intend to use for software. Most TV's have low quailty deinterlacers, like yadif and produce similar artifacts, marching ants, buzzing lines. Higher end TV's tend to have good deinterlacing, and good upscaling
Ok guys, I did many tests yesterday and gave Hybrid another go and i'm glad i did.
My first AVI capture is basically the MASTER with the noise at the bottom and I used UT Video T2 YUV422 BT.601 VCM UMY2
I was able to use FILL filter (if needed) to make a minor black border to mask the noise in Virtualdub2 to create the 2nd AVI prefixed with MASTER FILL, re-encoding again with UT Video T2 YUV422 BT.601 VCM UMY2. On a 34 minute video it's about 17 GB. Not bad.
Then I used Hybrid with QTGMC to deinterlace to make my 29.97 fps h.264 MP4 file. I also previously made this mp4 with Handbrake and I have to say Hybrid looks waaaaay better, and the size of the file is significantly smaller, I think because Hybrid is using variable bit rate? I have no clue how to change this. I did make a change to use CRF and 14 quantization factor, so if someone wants to guide me on where to make it use constant bit rate, I would like to test it out. But I am super happy with the end product.
Question though, I now have 3 files
MASTER.avi 17.4 GB
MASTER FILL.avi 16.9 GB
FINAL.mp4 1.23 GB
Can I delete MASTER.avi? The only thing I did in MASTER FILL is re-encode with the same lossless codec and added a fill mask. What do you guys think? I can be left with MASTER FILL and FINAL.mp4 It would tremendously lower the hit on HDD space. Also if anyone can point me or suggest settings to do in Hybrid, I'd appreciate it.
Random jumps/stutters?? You should post a 10 second clip to show an example.
Does your VCR have a TBC? If not, that could be the problem. As I said in my first post, all the things you are doing are "roundoff error" compared to using a good VCR and doing the capture correctly. The codec you use, the post-processing, and all the other later stuff cannot undo or fix a bad capture.
The quality difference between using a good VCR (and using it properly!) and using a lesser VCR are night and day.
VCR -> panasonic DMR-ES25 -> UCEC capture card -> PC
Im not sure if IN1 rear or IN2 front does TBC? I have to test that.
VCR -> Canon ZR80 AV2DV setting -> PC
For number 1 I see the random stutters/jerks, but very random.
For number 2 it's smooth at the same place in the video and all the way through.
Either the ES25 is not doing what it's supposed to (I will test front IN2 soon) or the UCEC capture card is not so good. I ordered a VC500 will arrive tomorrow and I will see. The problem with number 2 setup is that virtualdub will not let me pick my codec, it basically captures it as DV... I don't want that but I do like that it's smooth capture. I mean the end product im satisfied with but now that i know that i can do better, i've gone down that path to investigate lol! Anyone know if the Panasonic DMR-ES25 has tbc on front or rear? I had it plugged in the rear IN1 ports. Using the S-video out, but the vcr is yellow composite to the dvd recorder.
The stutter occurs at 3 seconds in example 1. example2 is the dv capture from the canon minidv camcorder as passthrough
Last edited by ngnaw; 26th Sep 2022 at 20:50. Reason: adding examples
What is the point of having the DVD player in the chain?? I don't see that it adds anything.
You didn't provide the make/model of your VCR. THAT is the key to this whole thing.
The UCEC capture card is a pretty bad choice for capturing.
I suggest you look at Lordsmurf's DigitalFAQ site for instructions on how to do a better capture. If he is around, he may have a few words to post in this thread.
I am afraid that that is a very low-end VCR. You are likely to get a lot of jitter, especially on EP tapes. It most definitely does not have a TBC.
However, I did a little more research and your DMR-ES25 does have a TBC of sorts, so I now understand why you included it in the capture chain. I don't have the manual so I don't know what settings are needed to produce the best results. Here is a thread at DigitalFAQ where they discussed using it to stabilize the video:
I tested an old MSI TV Anywhere Plus card and the capture from that was smooth, just like from the Canon Camcorder, however the MSI card lets me use the UT Video codecs. So my conclusion is that the UCEC USB capture is at fault for the stutters... Everything else equipment wise is the same. I hope the Diamond VC 500 will be better. I have another SONY VCR Sony SLV-675HF is that any better than the Sanyo i'm using? TBH I don't see anything wrong with the Sanyo for what I have available right now.
You could also use the "DV output" from your DVD player. It encodes using the DV codec, which may offend your sensibilities since it is compressed and, for NTSC, has a limited color space. However, if it is stable, and given that this is VHS, which is pretty awful to begin with, you may find the tradeoff to be worth it.
To capture the DV signal you use a Firewire (a.k.a., 1394) card in your computer. There actually isn't any "capture" because the signal from the Firewire is merely a copy of the bits generated by your DV player's DV codec.
Last edited by johnmeyer; 27th Sep 2022 at 10:59. Reason: punctuation
I'm glad you brought this up cause I was going to ask about DV. The Canon Mini DV camera has the dv out to the pc. Like let's be honest, isn't the recording on the mini dv tape better than the vhs tape? or is the colours better on the vhs? I don't know they look pretty similar to me.
Like you said I could capture a stable but limited colourspace via DV. I could do 2 types of connections
1. VCR composite out -> Canon ZR80 4 pin dv out -> DV in PC
2. This one I don't know if it will work the DMR ES25 has only a 4 pin dv in on the front, there is no dv out, there's composite, component, svideo and hdmi out. I was going to try this to see if it will passthrough on the composite and s video out to the usb capture card
VCR -> composite in Canon ZR80 dv out -> dv in DVD recorder -> composite in USB capture -> PC
The only thing that would help is that the canon will provide stability though? I don't know what will happen to the signal from the canon (DV) through to the USB Capture card??? I will test and let you know. I hope this connection will provide stable video but also allow me to use wtv codec I want. Connection 1. like you mentioned and i've seen is DV AVI.
Last edited by ngnaw; 27th Sep 2022 at 12:26.
I did most of my VHS tape captures by connecting the S-video from my VCR to my Sony TRV-11 camcorder. This was twenty years ago. I also had an ATI All in Wonder capture card, but I had all sorts of stability and levels problems. So even though it should have produced a better result, in actual practice, it did not.
While DV video has JPEG artifacts and limited color space, I found that I never dropped a frame in the over 100 tapes I transferred (probably more than that). Also, there were no stability problems introduced by the capture. It is the easiest format in the world to edit and if your NLE supports no-compress cuts-only editing, you can cut your video without any further degradation. (Actually, if you own Vegas or if you have the Mainconcept DV codec, both will re-compress with imperceptible degradation. Someone in the old Vegas forum posted a 100-generation DV recompression experiment, and the results were amazing.)
What I meant was if you had a VHS camcorder and a Canon ZR80 miniDV camcorder and you used both to record the same exact scene, the miniDV would/should be far superior to the VHS recording no? So using the dv camcorder as passthrough is solid enough to capture VHS tape. Right?
And you hit on my backstory from 2005 from my original post, I have the original captures for Tapes 1 to 10 as DV AVI (I thought I deleted them but I found them on a backup), and from that I made mpg and authored my dvds. So I doubt I would need to recapture those tapes if I already have the DV AVI and can just make my mp4s for those tapes. Correct?
Last edited by ngnaw; 27th Sep 2022 at 14:41.
I received the Diamond VC500 usb capture card and it's light and day better than the UCEC. Like I can't believe it. I already captured 4 videos and all looks good! I still wanna test the DVD DV IN passthrough though just for knowledge.
1. The DV sensor is true 720x480 (656x480 square pixel). VHS is analog so even though people talk about "pixel equivalent," that is just a way of talking about it, and doesn't really describe how it is stored. Having looked at a lot of home VHS video over the years, I'd say that most of it is closer to what you'd get with 320x240.
2. Despite DV's NTSC 4:1:1 color, it is still vastly superior to the way the NTSC color carrier is shoehorned into the analog signal. You'll never get decent, non-blooming red from VHS.
3. VHS has a HUGE amount of noise.
4. The imaging device in most VHS camcorder was very primitive and on the early ones (like the Vidicon and Saticon) you'd get smearing and ghosts.
5. Because VHS is analog you always get generation loss when you copy, and this includes the capture process. By contrast, once the DV camcorder records on tape, when you transfer it to the computer it is just that: a transfer. What you get on the computer is bit-for-bit identical to what is on the tape.
I already covered using the DV camcorder in the passthrough mode. Using it like this means that it is "just another capture card," like your VC500 or UCEC. I covered the pluses and minuses compared to those two.
I always recommend that people who are unfamiliar with video capture always start with DV because it is so bulletproof. Then, if they have other capture hardware, I recommend they capture 5-10 minutes of the same material and see if they notice an improvement. DV's weak points are its color and the JPEG compression artifacts. Watch the DV capture side-by-side (if you can) with capture of the same material using one of your other capture devices. Look for:
Color (saturation and "trueness")
Shadows and highlights (look for clipping in the whites, and crushed blacks in the shadows)
Picture stability (jumping or jitter or edge distortion)
Then sit back and, without overthinking anything, ask yourself which you like better. Have someone else watch alongside and get their opinion as well.
Most important: do all this before you spend weeks of your life doing this, only to have it slowly dawn on you that your captures aren't very good.
Thanks for the info and help johnmeyer. It's good to be able to reach out for some guidance. So like you mentioned the UCEC card is crap, well the Diamond VC500 is awesome! Better picture and stable smooth video... I am happy with it. So I will go with the initial VCR -> DVD Recorder -> Diamond VC500 -> PC
I will capture with VirtualDub2 as much as I can for the important tapes, and then basically let Hybrid do it's thing to mp4. I'll post some clips later.
Oh ya, I also decided to recapture the 10 VHS tapes I did back in 2005. The quality difference is worth it.
Vdub 1.9.11 for capturing.
And maybe upload a few seconds of your capture and collect comments/advice from members here.
I am not saying that, and never have.
However, almost every post asking for help capturing VHS tapes, both here and in the doom9.org forum, post horrible results when they provide a capture clip because there are so many ways to screw things up with good capture devices. Often I have found that the better the capture device, the more screwed up the result because of the issues with drivers and settings that must be mastered in order for things to work. By contrast, DV capture is pretty much bulletproof and with no settings to screw up and -- please excuse the insult -- it is idiot proof. This is especially true if you use Scenalyzer (now a free program) as the capture software.
I love how you say it: "... in many cases this means it [DV] is actually the better choice."
I still want to install and use the AJA Kona LHe capture card which someone gave me a year ago. It is supposed to be one of the better capture cards ever sold (I got it from someone who worked on well-known performer's music videos in the 80s and 90s). If I get it working, I will find the best SP VHS tape I have and then capture it from two VCRs: one low-end, and the other a semi-pro model. I'll then make two captures from each of these two VCRs, one using the most professional DV encoder I have (probably my Sony FX1 camcorder) and then other with the Kona card. I then plan to post a few 10-second clips from each of these four captures, along with a bunch of stills to illustrate what differences I find.
I might also do the same thing with one of my Criterion laserdiscs, just so I can use the most pristine NTSC signal available to me. And, if I really have some time, I'll also include captures from my Hauppage 1512 PVR2, just to have a comparison with a "better capture method" using something that is easily available to most people.
Last edited by johnmeyer; 28th Sep 2022 at 19:21.