A year ago I decided to digitize all my family memories that were still hidden inside that little transparent boxes, because I’ve never saw them and I also read somewhere they are at risk of being lost. Every year that passes the plastic tape can be ruined by the temperatures, the magnetic coat the plastic tape can get detached or lose magnetic field by an incorrect storage or simply by the unstoppable time-related degradation.
The problem is that, as probably you that are reading this post, I was a totally newbie using this kind of tech, because the last time I watched a VHS I was a kid, and I had no video cassette recorder (VCR) available to read my Video Home System (VHS) cassettes.
This is the story of how I digitized my home VHSs, thanks to this community and a lot of try and errors. I used the best technology available right now under 400 dollar, that is the minimum amount to spend to have the best quality right now (if your tapes are well preserved).
My aim is to help you to delve deep into this ancient and tricky world, that is so fascinating but can make you waste so much time. My setup was bought to acquire PAL VHS, but my VCR also supports NTSC signal.
And now some theory to go down the rabbit hole!
VHS
VHS cassettes are rectangular plastic containers designed to house magnetic tape used for recording and playing back analog video. The outer shell is made of durable plastic, typically a mix of ABS and polypropylene, ensuring the cassette is lightweight yet sturdy. Inside, two spools hold the magnetic tape, which is wound from one reel to the other during playback or recording. The tape itself is coated with iron oxide or chromium dioxide particles, which store the video and audio signals. A sliding flap on the front protects the tape, and this opens automatically when the cassette is inserted into a VHS player.
The signal on the tape is analog. This means that it doesn’t store pixel information, but line information. In analog video recording systems like VHS, a video frame is stored as a series of horizontal lines, with each line representing a portion of the image. These lines are stored sequentially as magnetic signals on the tape.
VCR
VCRs (Video Cassette Recorders) are electronic devices used to play and record video tapes, specifically VHS cassettes. They consist of several key components: a tape transport mechanism to move the magnetic tape across the playback heads, video and audio heads for reading and writing signals, motors to control the reels, and an electronic control system for functions like play, rewind, record, and eject.
VCRs come in different types:
1. Basic VCRs: Standard models for playing and recording VHS tapes, common in households. It usually uses a port called Video-composite.
2. S-VHS (Super VHS) VCRs: Designed for higher video resolution, used for semi-professional or home video enthusiasts. It has a better connection port called S-Video.
3. D-VHS (Digital VHS): is an advanced type of VCR that records and plays back high-definition video using digital data stored on magnetic tape. It uses MPEG-2 compression to store video, which allows for HD content (up to 1080i resolution) to be recorded.
While D-VHS VCRs are very rare and pricey, the other types of VCRs are more common. To properly digitize a VHS you can’t use a VHS VCR because the quality of the head are probably not the best, and it’s not able to display all the real colors and resolution of the image due to bandwidth limitation caused by the composite port and cable. For this operation, a S-VHS VCR is mandatory, and you can find useful information here (
https://www.digitalfaq.com/forum/vid...ing-guide.html).
I bought a JVC HR-S7700 on
https://vcrshop.com/, considered a trusted vendor by this community. He was really kind and helpful even in solving a problem I had with a first unit.
The JVC has some internal settings you need. First, you must set the language. If you buy at vcrshop’s it’s Dutch, and I don’t understand it. Then set the S-video port or SCART port if you use the S-video-to-SCART adapter. I don’t attach any picture of th because I don’t have it.
What I have is the Function view in the JVC menu.
B.E.S.T must be ON, because it makes the image better, and the Image Automatic. Time Autom… is better to be off, you don’t need it to digitize. O.S.D. is the function that displays the pause/play button. It’s annoying, keep it off.
I set Digital 3R off. It is a smoothing filter that adds haloing noise. You don’t need it because the deinterlacing QTGMC algorithm is way better.
INTERLACING AND TBC
In VHS and many other analog video systems, interlacing is a method used to display video frames. Instead of showing a full frame of video at once, interlacing breaks each frame into two fields: one containing the odd-numbered lines and the other containing the even-numbered lines of the frame. These two fields are displayed in rapid succession, giving the illusion of a full frame when viewed at normal playback speed.
A Time Base Corrector (TBC) is a device used to correct timing errors in analog video signals, particularly those from VHS, Betamax, or other tape-based formats. These errors, known as "time base errors," occur due to inconsistencies in the movement of the tape, causing visual distortions like jitter, unstable images, or color shifts.
A TBC stabilizes the video signal by temporarily storing video frames and then re-timing the output to ensure smooth, consistent playback. This process eliminates distortions, producing a cleaner and more stable image.
TBCs are commonly used during the transfer of analog video to digital formats, ensuring that the video signal is properly synchronized before digital encoding. There are two main types:
1. Full-frame TBC: Corrects timing issues across the entire frame, offering the most comprehensive stabilization.
2. Line TBC: Focuses on correcting timing errors line by line, improving image stability without addressing full-frame issues.
The most common type of TBC you can get is the line TBC, because it’s cheap and effective. A Full-Frame TBC can cost up to 2000 dollars nowadays and usually hard to find, but if you need a good one ask to lordsmurf or find it in (
https://www.digitalfaq.com/forum/marketplace/).
A line TBC resolves line-related synchronization problem, avoiding line shifting and inconsistency of the frame along the borders..
In this forum, the most discussed TBC is the Panasonic es10, defined by lordsmurf (here
https://www.digitalfaq.com/forum/vid...ing-setup.html) as
“The ES10/15 is a strong+crippled line TBC, with non-TBC frame sync. It's best used only for anti-tearing, as tearing is a harsh error that S-VHS VCR TBCs usually cannot fix. The ES10/15 is a DVD recorder with heavy aggressive NR (even when "off"), posterization, etc.” I’ve bought one but is difficult to find one in good condition and at an affordable price, and it also adds very annoying artifacts, loosening your video quality. It’s good only in case of tearing apparently, but it’s a very borderline case. Let’s find a JVC VCR with internal TBC and you are going to obtain the best result.
CAPTURE CARD
The capture card is the essential device that makes your computer obtain the video signal from the VCR. It basically converts a radio signal to a video signal. In this forum the most appreciated cards are from the brand ATI (
https://www.digitalfaq.com/forum/vid...der-hacks.html).
This is a very important piece of your gear, NEVER buy cheap capture cards, NEVER buy modern ones. The cheap cards make your image degraded, with poor color and lost details. Unfortunately, also the best capture cards are bad, like the Elgato Video Capture. A great comparison is made in this Youtube video (
https://www.youtube.com/watch?v=eCK-...sonablyBritish) (watch at highest resolution possible, also to avoid YouTube compression). A modern model that is considered good is the GV-USB2 (here a review
https://www.youtube.com/watch?v=flc8...sonablyBritish), but I still discourage you to buy it.
As pointed out by the first video, one of the best devices is the Intensity Pro (not 4k) that also I own. Unluckily it’s not the one used for the video comparison, but is better than the Intensity Pro 4k, that is the best capture card in the comparison. If you can buy an ATI card, but this exact model is top tier for Blackmagic.
ACQUISITION SOFTWARE
The software that lets you store the video signal captured by the capture card. The state of the art now is VirtualDub1.9.11. You can find the download files and the discussion here (
https://www.digitalfaq.com/forum/vid...lters-pre.html). I have a Blackmagic capture card, so I could use Blackmagic Media Express as acquisition software. I don’t use it because
VirtualDub with
HuffYUV is way better.
If you need a great guide to install
HuffYUV you can find it here (
https://www.digitalfaq.com/forum/vid...-settings.html).
Some additional tips to use it. Set the device in the Device tab and in the Audio tab. In the Audio tab also check “Enable audio capture” and “Enable audio playback”. Start the recording only in the Capture tab>”Capture video”. Never use the F5 or F6 buttons because it can cause errors.
BE CAREFUL: If you have two monitors attached to your pc, disconnect one because if you move the
VirtualDub windows (or sometimes you use an app in the other monitor) it will continue to record just the audio, and the video will freeze. You will never notice it until it ends capturing and you rewatch the video.
After you press Capture video you should use the pc only to stop the capture.
TOP TIP: nobody knows or maybe nobody says it, but there is a solution to the annoying problem of audio recording latency that VHSs acquired with VirtualDub suffer, no matter which version of VirtualDub you have installed. You must go to the Windows Mixer and lower the volume of VirtualDub. I put it at 50%. For what I have understood looking on the web, the root cause is the audio clipping.
ENCODER
A video encoder is a software that converts raw video data into a compressed format to reduce file size while preserving quality. This process typically involves using a codec (compression-decompression algorithm) to encode the video into formats like H.264, MPEG, or others, making it easier to store, transmit, or stream.
HuffYUV is a lossless video codec, meaning it compresses video without any loss in quality. It uses Huffman coding for compression, making it much more efficient than uncompressed video formats. While HuffYUV reduces file size significantly compared to uncompressed formats, it maintains the exact original image quality. We chose HuffYUV for this features, it is not compressed like AVI Motion JPEG and it doesn’t consume a lot of storage space, opposed to uncompressed codecs.
To install HuffYUV download ONLY the file named HuffYUV.rar (from here
https://www.digitalfaq.com/forum/vid...lters-pre.html) and follow the instructions to install it.
CABLES
The cables needed are: one S-Video cable and one composite cable for the audio signal. Try to buy the best cables you can buy, with good screening. Some VCRs have no S-Video port, especially the European models. You should use an adapter from SCART to S-Video. Be careful, these adapters are sometimes not very good because they introduce noise. If you can, buy a VCR with S-Video port, and use it. It’s possible you must adjust the settings to force this port.
Regarding this problem, the user sharc from VideoHelp (
https://forum.videohelp.com/threads/...range-problems) affirms that:
“A reminder of caution regarding these SCART adaptors: Internally they are open wired using wire straps of say 1....3" length, and sometimes a PCB and an IN/OUT switch. Due to capacitive coupling between the unshielded wires the Y signal will be contaminated by the color carrier (C, modulated chroma) which is present on the C wire (and vice versa).
Means the luma signal is no longer a clean grayscale signal but rather a "muddy" composite signal which represents a low saturated pale "colored" picture.
Attached oscillogram shows the Y at the output of a SCART adaptor. One can clearly see the residual chroma (C) which is superimposed on the luma due to capacitive coupling (crosstalk) happening within the SCART adaptor. It's like a "weak" composite signal.
Same (even much worse) happens with low quality poorly screened SCART-to-S-video adaptor cables. It can also be a source of "ghosts" due to the phase shift between the true signals and the crosstalk interference. "
COMPUTER
The computer required to digitize VHS needs Windows (they say XP is the best, I used W10 flawlessly) and an SSD (best if NVME). Considering you digitize a lossless signal, you need a very high writing-speed drive, so the file you capture maintains its integrity having no writing errors.
Some capture cards require PCI Express port, so be careful which PC you choose.
If you have a second monitor unplug it, it’s mandatory to have a working capture.
DEINTERLACING AND HYBRID
As pointed out, interlacing is a technology that nowadays is not used anymore, but was useful years ago to compress a video, even if it has some downsides.
In order to resolve them, you need to deinterlace your acquired HuffYUV video using the state-of-the-art algorithm called QTGMC, capable of extracting single frame lines (remember every frame is made by only half of its line), obtaining the frame with only half lines as it was recorded, then it recovers the missing lines from the next frame, performing a temporal and spatial interpolation. It comes with various performance presettings. You could customize them, but the default ones are so perfect.
Lordsmurf usually suggests Faster setting, I prefer Fast. I strongly recommend you to do some testing on your specific videos after reading this comparison (
https://macilatthefront.blogspot.com...-settings.html).
QTGMC is available by graphic interface thanks to Selur. Some time ago I followed the settings advice by lordsmurf. Unluckily I’m not able to find that discussion anymore. Anyway, I can attach its screen showing his settings. Those settings are about an old Hybrid version, so old that I couldn’t find it on the web. I also attach my settings for one of the most modern release (2024.03.26.1) that you can find here (
https://www.videohelp.com/software/Hybrid/old-versions). I copied lordsmurf setting, except for some values that are not supported anymore, the Fast QTGMC and the maximum number of reference frames. I reduced it because this avoids weird screen when you pause and resume playing from VLC.
I attached some screenshots to this post. Unzip it. Now I’m going to explain the trickiest settings of every single image. If I don’t explain it, you can simply copy it or look for it in DigitalFAQ.
1. I use x264, if your hardware is fast enough you can probably go for x265 to save so much space without losing details. I keep the audio untouched.
2. I use a low quantization factor, just for archiving reasons. The file size will be higher, but it will have no perceived quality loss.
3. I set 3 maximum number of reference frames, because if you set 5 as lordsmurf does, you will have weird artifacts when you stop the video in VLC and then play it. Lordsmurf does it just to make the size file smaller, and it will not affect the video quality. It’s a question of I-B-P frames, do as you prefer.
7. The first time I digitized a VHS I was convinced the image was the best obtainable, but I was wrong. I discovered a lot of things, also that your capture have not the real aspect ratio it was recorded. Due to the pixel size that is not squared, the image acquired has a wrong aspect ratio, so you can adjust it using bicubic resize method and the final resolution of 740x540 for PAL VHS.
8. My favorite settings for QTGMC (use Vapoursynth version) is Fast. Check bob to do a correct deinterlacing, outputting two frames for every field. Is technical, this should be a good fast explanation. Not using it will create a not fluid video.
Pad mod16 really improves the deinterlacing speed, you could check it. Anyway, selur not recommends it because is a not enough tested feature.
The overwrite input scan type must always be top field first. If you do the opposite, you could be seasick watching the video. There is a minority of VHS using the bottom field. If you are seasick, change this one.
UPSCALER
The upscalers are usually hardware devices that upscale your video signal, the most discussed recently are retroTINK models.
While it seems a great thing, remember that VHS stores an analog signal, so you can’t really use maths to upscale the image as it was made by pixels. This is one of the reasons to avoid upscalers.
They add extra cost and don’t really improve your picture quality. What you really want is to obtain the most pure and untouched image as you can, to preserve its quality, and maybe feed your upscaling AI in the future with the most authentic image.
ARTIFICIAL INTELLIGENCE
A lot of improvements have been made in this field. You can buy a software named Topaz video or upscale using a state-of-the-art neural network. My advice is to wait some years to do this operation, as technology gets better.
Anyway, NEVER EVER throw away your original deinterlaced captures and your VHS.
VHS-DECODE
VHS-Decode is a software capable of converting RF signal from the head of your VCR and obtaining the video file, bypassing all the electronics of the VCR.
While this approach aims to acquire the highest picture quality, giving you also the opportunity to archive the raw data from the head to decode it when the technology will be mature enough, it is not recommended due to invasive hardware modification of VCR, high file size (~300GB / 1 hour long VHS) and high computing times (~8 hours/VHS with high-end PCs).
CONCLUSIONS
This guide wants to be the Half-Blood Prince book, that gives you the theory and practice advices together. I’m sharing the knowledge I acquired digitizing VHSs and reading this forum in the last year.
I’m sure enough I will regret not writing some other tips in this post, but I’m sure that when it comes to my mind, this post will be surely published. This is one of the reasons why I advise you to read the comments. I will probably publish some updates, and you will find useful tips or corrections to my writings by other members of this forum.
Thanks for reading it!
johnEtna