- August 25, 2015 at 9:59 pm
I’ve seen a lot of really poor tutorials and explanations for digitizing VHS tapes. I’d like to gather some information from professionals regarding how they would handle such a project, and how they would measure the quality of their output. This is basically a series of questions, but someone with experience can most likely answer a number of them in one go.
Specifically, I am looking at three major aspects of the analog-to-digital conversion process:
– Color quality
(Note: I’m not looking for a specific product recommendation, just information on how to make the best informed decision regarding items with which I am not particularly familiar. Also, regarding resolution, I’m aware that VHS’s are analog – and yet, I want the highest quality digital output, and one way to measure quality is to count the number of lines of the digital file that is created by the digitization process.)
So, there are two major ways that I’ve seen people suggest performing the conversion. Both begin with buying a VHS player. For best quality, people usually recommend a “professional” VCR, specifically one with TBC (timebase correction). A professional-grade VCR is also less likely to chew up the VHS, due to better parts in a better product.
My first series of questions surrounds this – how do I know if I’m buying a “good” VCR? For instance, when looking for a good DSLR, someone might mention the flash card format, the ability to switch lenses or other peripherals, if it can do fancy things like HDR, color quality, and maximum resolution of the sensor. What are similar questions for a VCR? What separates a good VCR from a bad VCR?
My second series of questions is about output – what is the best way to capture the output of a VCR? Should I use RCA, S-Video, or coax? Which of the three standard outputs from a VCR will get me the best picture in the end – where “best” means the least noise, most accurate color quality, and highest resolution?
This is where the two ways of digitizing VHS’s diverge. One method suggests taking the output of the video and feeding it through a small device that will convert the output to USB (something like this: https://thumbs4.ebaystatic.com/d/l225/m/mRZZOlmPIRm_v389CP8uTog.jpg), and then using software to convert that. Another method is to use a capture card – something that plugs in to the motherboard itself, usually through PCI or PCIE (something like this: https://i.kinja-img.com/gawker-media/image/upload/s–bvz-BA6g–/182apxoyw35vqpng.png).
My third series of questions revolves around this step – regardless of the answers to previous steps, assume that I now have some kind of analog output, which I’d like to bring in to a computer. What’s the best way to do that – where “best” again refers to least noise, most accurate color quality, and highest resolution? What are the pros and cons of using a small USB device vs. a full-fledged capture card? The capture cards appear to be the more expensive option, do I get significantly better quality out of one? Do they have features that a small USB device doesn’t?
Oddly enough, there are a ton of tutorials and forum posts online regarding these kinds of questions, but very, very few offer in-depth answers to these kinds of questions. I have yet to find any that answer these questions in the depth I’ve come to expect when it comes to video projects – you can find PHD-level theses about the differences between H.264 and MPEG4, but I can’t very much at all when it comes to questions like these. If you could answer any or all of the questions, or point me to resources that can answer these questions specifically, I would truly appreciate it. Thanks!
- August 26, 2015 at 2:18 pm
Johan, let’s start with VHS itself. An SD consumer format, the resolution of the VHS picture is 720×486. Actual lines of resolution for standard VHS will be around 240 or so. It is, by modern standards, a very low rez format.
As a color-under format (color information on a subcarrier frequency) it usually can carry a fair amount of noise in the color signal.
I do agree with the recommendation of finding a pro deck (usually Panasonic or JVC, although there are others) that has time base correction. This is will adjust for all the little timing errors in a mechanical tape system, assuring every scan line appears exactly timed with those above and below it.
How do you know if it is a “good” VCR? A more relative question. Buy known brands from reputable sellers. Buy from someone who also refurbishes and repairs these decks. You aren’t going to see huge numbers of options that matter any more on these decks, with the exception of output.
In the analog world, component output (separate RGB) is best. Next in line would be S-video, separating lumininance and color information. Third would be composite (what you called “RCA”), with luminance and color already “mixed”. Last in line would be RF over coax, which is video and audio mixed and converted to a broadcast channel frequency.
For digitizing VHS, in my humble opinion, a USB-style converter will be fine. If you haven’t done the most you can with the steps above this one, a fancy digital encoder will not save you. Garbage in, garbage out. I use an older adapter that converts analog video to Firewire, and record direct into Premiere Pro.
VHS is not a particularly robust format. The ideal of time base correction and component out will give you about the best you can get.
If material was shot poorly or in bad light, none of this will make that look good, just less miserable.
Dell Precision T7600 (x2)
Win 7 64-bit
Adobe CC 2014 (as of 7/2015)
256GB SSD system drive
4 internal media drives RAID 5
Typically cutting short form from HD MP4 and P2 MXF.
- August 26, 2015 at 3:02 pm
1. short answer, buy a panasonic 4 head vhs that outputs component and Blackmagic Design Intensity Pro 4k(support capture and playback output and its mjpeg codec)
2. long answer(for when you have free time to read below)
2 words: compression and resizing.
Tip #1. Many capture cards, have a native resolution size and then they just resize to whatever you request. If you have a fast computer, try to capture
their native, even if its way bigger than you need. You can then resize with professional software/hardware later on.
If you have space issues, capture 640×480 or 720×480 29.97fps. but I’d go with 720 so that you have a standard scalable resolution. Don’t capture with
contrast filters on! AE can do better unsharping/contrast in post.
if the software allows it, try 720×486. There’s 486 visible lines sometimes in the source footage. Think “the ring”.
a digital NTSC video frame is 720 pixels x 486 lines, and a PAL video frame is 720 pixels x 576 lines.
other technical considerations for ntsc:
color management: BT.601 16-235 or 0-255 full gamut(test). (don’t use 1953 color)
White point is D65 6500k. Top/even field.
0.91 pixel aspect ratio
Tip #2.Capture at little to no compression so you can recompress later. It’s called generation loss. The more you compress, the more blockiness/artifacts appear. Some of the best capture cards actually can be set to many pixel sizes and you can choose the codec of your choice depending on your hard drive speed.
I used to use an ati all in wonder, cheap picvideo codec with Virtualdub. no droppped frames! 2700KB/s ; it’s perfect for 7200rpm hard drives. supports
Advanced Deblocking of Highly Compressed Video, DirectShow Codec.
Some newer models: all component, not composite:
Blackmagic Design Intensity Pro 4k ProRes and DNxHD, or as fully uncompressed 10-bit video,
Black magic decklink can capture uncompressed, for instance. but it needs special usb renises driver
Diamond Multimedia USB 2.0 High Definition can capture component.
StarTech.com Standalone Video Capture component.
Matrox MXO2 Mini High Definition HDMI and Analog I/O Device component
Hauppauge Colossus PCI Express Internal HD-PVR With Component
Tip #3. You’ll want a 4 head S-VHS with component output and Time base correction to reduce fringing and poor color. If you can find one used in TV stations, they are professional quality. Clean with alcohal on the heads, run dummy tape cleaner. You’ll want to stay away from composite. It is an old, noisy way to combine all singals. s-video is a little better. Use component intead, as it is the best.
If you have a slow computer and little disk space, Picvideo codecs(a motion jpeg) always worked well for me because its intraframe-only compression is easy
on computers, hd space and cpu’s. Prores/DNxHD is also small.
codec capture roundup:motion jpeg’s or jpeg2000 or prores.
mjpeg tolerates rapidly changing motion in the video stream, whereas compression schemes using interframe compression can often experience.
cpu lag, interframe compression can often experience unacceptable quality loss when the video content changes significantly between each frame. temporal
compression can create “artifacts” or remnants of previous R-Frames, until the next I-Frame refreshes the scene
smaller file size than mjpeg
compatible mac, small size, less compression
if your raid can handle it, go for it! nothing beats 100% perfect quality uncompressed.
Tip #6 Output or final render gear:
output: Blackmagic Design Intensity Pro 4k-realtime playback output
supports color management BT.601.hardware Algolith box, terrenax box, topaz enhance,After Effects preserve upscale, DCT artifact removal, resize, flicker,
mosquito noise, color halos, deinterlacing, etc.
Tip #7 Deinterlace with the highest quality deinterlacer before final render/upscale or you wasted ALL previous steps!
- August 26, 2015 at 5:08 pm
Deck – look for an S-VHS deck, as VHS decks do not normally have S-video outputs. I’ve never seen Component on a VHS or SVHS deck, maybe they exist, I don’t know, but typically S-video is as good as it gets for VHS output. Pro-sumer edit decks like a Panasonic AG-1980 or JVC 9xxx series SVHS decks will have time-base correction and S-video output.
I’d go with the Black Magic Intensity for capture. On a PC, will record as an .avi using Black Magic codec which is high-quality, high bit-rate 4:2:2 video.
While I’ve never tried a “USB capture device”, these are usually cheap sub-$100 units having Composite input (poor) and then they also often compress directly to MPEG-2 or H.264. These formats will have a lot of quality loss, since it is squishing the video down to a data rate that works over USB 2.0, which is not very fast. You get what you pay for.
Safe Harbor Computers
- September 4, 2015 at 2:57 pm
The company I used to work for had a JVC BR-S622DXU VHS deck that had a TBC with noise reduction and component outputs.
This model here: https://pro.jvc.com/prof/attributes/features.jsp?model_id=MDL100045
The only caveat is that the deck would only play back tapes recored in “SP”. It seemed like half the tapes we received were recorded in “EP” so we would have to use a consumer deck to digitize those tapes.
- March 22, 2020 at 12:53 am
There is no Wrong way to do things; though there are tradeoffs in different methods.
Define your goal clearly. If all you want to do is digitize, usb will work, giving you OK but not necessarily discernable results. If you want something more discernable, you can go DV passthrough, which will force an output at 720×480 in most cases, and it will be interlaced video, giving you only 240 effective pixels in each field and creating a blocky motion in some areas, with a blurry motion in others. Really old VHS tapes shot at 24 frames, in interlaced 48, and had to be Telecined to 29.97 in broadcast, with what is often referred to as a 3:2 pulldown, which used the interlace by pulling down one set of interlace, to create another set for an IN-Between frame, that created the new frames necessary for 29.97 frames. This was originally developed as a means of averaging mechanical function in exactness, in the NTSC standard. Though this also resulted in signals that were rarely similar in color field once dubbed, giving the standard another acronym for a nickname “never the same color”. After a while, only one or two camera makers could actually compensate for the signal differences, either by developing a process to maintain equipment or developing a mechanism that maintained itself. The quality of the images increased, while the size of the data began to decrease.
For home movies, you’ll have a lot of work to do to clean them up. No matter what plug you use to output the signal, you’ll get noise, color degradation, signal degradation, and a host of other issues to contend with, along with the standard INTERLACE issues. For best quality:
1. Use a player with SVIDEO out, it splits the data up into several pathways. Rather than composite, which doesn’t split the video data, just jumbles it together, and makes for noisy transmission, S-video outputs the video information in separate streams that are combined in the next device in the chain, where any added noise can be filtered out. The sound is output as well in most instances, but some require the use of the composite sound (despite the added composite, these are better, more split in the video data and it’s cleaner).
2. Use a device that supports S-video, and selection of standard. This will allow you to set the standard your video will output to.
3. Run the video through twice on interfaces that upscale for you, once using the upscale, once not. Compare the results with an upscale you run in after effects.
Once in After effects:
A. If you know the format was in 24frame format, you can interpret the footage first, and remove a 3:2 pulldown. The first set is the most common. You can try several different ones to see what gets the best result, then delete the others (import the same video several times and interpret each one differently, or just duplicate it).
B. Start with a comp at the same size as the video, run deinterlace if you wish, but you may opt for a motion estimated plugin to fix it properly. Try this twice, and set both an upper and a lower field order. The best looking one is your starting point pre-comp.
C. Place the pre-comp into a comp at your output size, apply any detail enhancement or color adjustment, then upscale. The detail retaining upscaler works well on its own. You may put your first precomp into a second, at the same size, then apply detail work. Afterward, place the second into the final and upscale. This layers the adjustments so they are done in successive passes, retaining more detail, though it will take longer to process out. Once your results show ok in the previews, you should be able to farm it out to media encoder without a problem. You can even compress the new stream, in progressive video format.
Log in to reply.