Standard play VHS video cassette recorders offered only about 3 MHz luminance bandwidth by recording the signal on an FM carrier. What is that carrier frequency? Is it possible to confirm the carrier frequency or even the FM deviation by analyzing archived VHS recordings?

Later edit: Since answers are describing the chroma handling in VHS recording I point out that it is only the luminance that I seek information about. Specifically, how is a European Teletext waveform distorted? Teletext is a non-return-to-zero digital stream at 6.9375 Mb/s added to luminance on spare lines in the vertical blanking interval. Recovering Teletext from archived VHS recordings would require detecting the phase of the clock run-in stream at half bitrate i.e. 3.46875 MHz which is below the noise AFAIK.

  • \$\begingroup\$ Is this part of GrimFandango's ongoing Teletext recovery project or are you someone seperate? \$\endgroup\$
    – RoboJ1M
    Commented Dec 14, 2018 at 17:24

3 Answers 3


For video the carrier is around 3.4 --> 4.4MHz (For NTSC)

for sound its 1.3 to 1.7MHz



  • \$\begingroup\$ I award you the bounty! You answered the video carrier frequency and the book reference is very useful. \$\endgroup\$ Commented Jun 21, 2015 at 2:41

I went and looked up the early papers from the JVC engineers. Unfortunately they don't seem to be available outside of IEEExplore. Presumably this information can be found in textbooks on magnetic recording.

For recording the luminance, the video signal is low-pass filtered, given pre-emphasis, then FM modulated with carrier "3.4-4.4 MHz". I'm not clear on why this is a range.

The chrominance is band-pass filtered, boosted by 6 dB, mixed down from 3.58 MHz to 629 kHz, and low-pass filtered. Thus it sits 'under' the luma in the combined signal.

There is a further wrinkle to the chroma that is pretty interesting. The azimuthal recording technique effectively reduces crosstalk from adjacent tracks in the case of mistracking, but only for the luma frequencies. The low-frequency crosstalk for the chroma can be much worse, so additional steps are required. The phase of the chroma signal gets advanced by 90 degrees every horizontal period. During playback, the phase is restored, and the signal added to a delayed copy of itself. (This is for NTSC, the process is slightly different for PAL.)

I don't see any reason you wouldn't be able to see this if you analyzed the signal recorded on a tape. You'd have tap into it before demodulation though.

1: http://dx.doi.org/10.1109/TMAG.1978.1059827

2: http://dx.doi.org/10.1109/TCE.1978.267051

  • \$\begingroup\$ The luminance carrier frequency is literally modulated in the range 3.4 MHz (black or sync tip) to 4.4 MHz (peak white). I don't see how one can tap into a VHS tape "before demodulation". \$\endgroup\$ Commented Jun 21, 2015 at 2:45
  • \$\begingroup\$ i.e. between the read head and the FM demod. \$\endgroup\$
    – mng
    Commented Jun 21, 2015 at 15:22
  • \$\begingroup\$ It seems you ask me to rebuild the electronics of a VHS player instead of using available (highly developed, i.c. based) equipment. What I actually have to work with are digitised records (8-bits, oversampled) from a good player of an archive of VHS cassettes of historical interest that are not readily available and may have decayed by now. \$\endgroup\$ Commented Jun 21, 2015 at 23:29
  • \$\begingroup\$ Sorry, but it wasn't clear what was meant by 'archived VHS recordings'. I mean, I have a archive in my closet of physical tapes. It would have helped if this information, not to mention the interest in Teletext, had been in the original question. \$\endgroup\$
    – mng
    Commented Jun 22, 2015 at 0:12

Note: When I wrote this answer, I assumed, like other signals, that the composite video signal was recorded directly to tape with only the odd quirk of being striped diagonally instead of a continuous timebase. It turns out that that's not true. The color information is actually decoded and re-encoded differently for tape, then decoded again and re-encoded back to the composite standard for playback. So I learned something new this week. See the other answers for more detail.

Original answer is as follows:

VHS video as I understand it is simply a composite (RCA yellow) video signal recorded on magnetic tape, along with a separate analog soundtrack. This signal is divided into 1-frame increments that are stored diagonally across the tape, hence the angled, spinning head if you've ever looked inside a VCR. It's not RF modulated except as defined within the signal itself for color information. Googling "composite video standard" without quotes gives lots of results about how the signal works. Here's a good one from google images:

enter image description here

For black-and-white video, there is no RF modulation at all. Just an instantaneous DC level that gets amplified and sent to the electron gun at the back of a CRT (tube) display. For color, and to maintain B&W-compatibility, the difference from pure white is encoded in two mixed AM waves of the same frequency and different phase compared to the color burst that appears between each line of the picture. This frequency is 3.579545MHz, but it only applies for color video. B&W displays just don't have the bandwidth of their own to respond to that, so they only react to the average level, which is precisely the luminance signal that they always used.

Also to note are the negative-going H and V sync pulses; the only difference is that that V. sync covers an entire line while the H. sync is only a blip between the previous line and the color burst.

In summary, a color composite video signal requires 3.579545MHz, so any equipment that supports it must have at least that much also, but don't be surprised if it's only barely, like maybe 3.7MHz or so.

As a side-note, analog TV broadcast was simply that composite signal on another AM radio wave. That's 2 layers of AM for color. Audio was FM on a different frequency, just like analog radio broadcast, so the TV required two tuners to receive one channel. This is probably why analog TV tuners for computers often include FM radio but not AM radio - the hardware is already there without modification.

  • \$\begingroup\$ From what I understand, some videotape formats like uMatic use phase-modulated chroma, but it requires that the head speed be maintained more uniformly than is practical in home equipment. From what I recall, VHS separately amplitude-modulates the red and blue components using separate carrier frequencies; the red carrier frequency is pretty low, which is why reds tend to "bleed" on VHS. \$\endgroup\$
    – supercat
    Commented Jun 15, 2015 at 17:27
  • \$\begingroup\$ That may be possible, but I don't see why the recording format would be different from the in/out format except for maybe a technical limitation of the medium. Unfortunately, I don't know tape well enough to say one way or another. \$\endgroup\$
    – AaronD
    Commented Jun 15, 2015 at 17:31
  • \$\begingroup\$ If a video source has 227.5 chroma cycles per scan line, each scan line must be 63.556us long; a deviation of even 0.1% on the speed over the course of a scan line would mean that the color phase that represents red at the start of the line would represent orange or magenta at the end. Separating the chroma into two separate amplitude-modulated signals greatly reduces the sensitivity to head speed. \$\endgroup\$
    – supercat
    Commented Jun 15, 2015 at 17:49
  • \$\begingroup\$ Okay. That makes sense. Would you like to take a stab at answering? \$\endgroup\$
    – AaronD
    Commented Jun 15, 2015 at 18:48
  • \$\begingroup\$ "I don't see why the recording format would be different from the in/out format except for maybe a technical limitation of the medium." Exactly. And boy oh boy does magnetic tape have a bunch of limitations in store for it. VHS and other video storage is quite complicated - much more so than it seems at first glance. I've once made an all-digital controller for a VHS deck and boy was it no fun. \$\endgroup\$ Commented Jun 15, 2015 at 18:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.