This project is based on the Macrovision FAQ written by Antti Paarlahti.
Many pre-recorded tapes and DVDs are copy protected by Macrovision. This prevents pirates from copying and selling media and actually stealing other people's work. However it is perfectly legal to make a copy for yourself (might depend on the country you live in). But if you try to copy a protected tape or DVD, you'll notice that the picture goes from bright to dark and back to bright continuously. How do they do that?
To anser that question, you have to know how a TV picture is built up. The description below is for the PAL system. NTSC is very similar; only the number of lines and the timing differs.
A TV picture consists of 625 lines and is refreshed 50 times per second. In the early days of television, electronics weren't fast enough to write 625 lines 50 times per second. So they introduced interlacing: the first 1/50th second the odd lines (1, 3, 5, etc.) are written; the next 1/50th second the even lines (2, 4, 6, etc.) are written. This means it takes 1/25th = 0.04 seconds to write a whole picture. Writing a single line takes 0.04/625 = 64us. Lines are separated by a horizontal synchronization pulse; the frequency of these pulses is therefore 1/64us = 15625Hz. This is the high pitched tone you sometimes hear coming from a TV set. A vertical synchronization pulse tell the TV to start at the top of the screen again (starting at line 1 or 2). The frequency of the vertical sync pulses is 50Hz.
The picture below shows a single line.
A line starts with a horizontal sync pulse. Next comes the color burst. The remainder on the line contains the picture data.
The first 20 lines or so are not used for the picture. The picture data section of these lines contain other useful info like Teletext. When playing a video tape or DVD, these lines contain no information at all:
The last few lines before the real picture begins contain the black level. This tells the AGC in the VCR how strong the signal is. ACG stands for Automatic Gain Control. It amplifies a weak video signal and attenuates a strong signal. Macrovision uses this behaviour for its nasty tricks; it adds high amplitude pulses to the last few lines:
Result: the AGC thinks there a strong video signal and will attenuate it, making the picture is very dim. The amplitude of the Macrovision pulses vary over time. Otherwise you only had to turn the brightness control on the TV set when playing a copied tape. To make things worse, Macrovision also adds false horizontal sync pulses, making the VCR loose count. The picture below shows a close-up of a single 'Macrovision-contaminated' line.
The line starts with a real horizontal sync pulse, followed by a color burst just like any other line. But then the close-up clearly shows the false sync pulses and the misleading black level.
TV sets usually don't bother about Macrovision's high amplitude pulses, because they don't have an ACG. If the picture is too bright or too dark, you can simply adjust the brightness. A TV also has its own horizontal sync pulse generator, so a few false pulses in the video signal doesn't upset it.
If we want to get rid of Macrovision, we need to restore the picture data in the last few lines to black level again. Experiments showed that it's OK to eliminate the first 20 lines instead of only the last few lines before the picture starts.
When we want to restore picture data to black level, we first need to determine the voltage of the black level. This is done by U3A, U4B, C6 and U3B, a simple sample-and-hold circuit. U4B is a switch controlled by pin 9. If pin 9 is L (0V), pin 4 is connected to pin 5. If pin 9 is H, pin 4 is connected to pin 3. Pin 9 is connected to pin 5 of U1, a sync separator. Pin 5 is L during the color burst. This means that C6 is charged with the video voltage during the color burst. This is not really equal to black level, but it's close enough. Both opamps (U3A and U3B) serve as a buffer. The false sync pulses also fool U1. This means that pin 9 will also become L during the Macrovision pulses. If U3A's input were directly connected to to the video signal, C6 would hold the wrong voltage. Therefore, U3A's input is connected to the video output (which is of course at black level during the false pulses; that's the whole purpose of this circuit).
The black level voltage is connected to pin 1 of switch U4A, which is controlled by pin 10. We want to switch to black level during the lines that may contain Macrovision pulses, and whithin those lines only during the picture data (not during horizontal sync and color burst). This means that pin 10 must be H if the video signal contains picture data AND only during the first 20 lines.
R10, D1 and D2 make a perfect AND gate. The inputs of this gate are connected to MMVs U2A and U2B. The output of U2A is H during the first 20 lines; U2B's output is H during the picture data. MMV U2A is triggered by a vertical sync pulse. The amout of time the output remains H after the pulse is detrmined by R11 and C4. MMV U2B is triggered right after the color burst pulse (on the positive edge of that pulse). It's output returns back to L after about 50us (length of the picture data), determined by R12 and C5. To make sure the MMV will not be retriggered by false sync pulses in the picture data, pin 9 is connected to pin 11.
Pin 15 contains the Macrovision-free video signal and is buffered by the circuitry around T1 and T2.
You can just take any opamp you like. I used a TL082 for U3.A and U3.B, but I think any opamp will do. Just make sure it has a high input resistance so C6 will not discharge to much.
The DC voltage across R7 and R8 will be about half the supply voltage: 6V. So the power dissipation in R7 and R8 will be about 62/150 = 0.24W. Using regular 1/4W resistors is not a good idea. Use 1/2W resistors instead.
The values given for C4 and C5 worked fine for me, but you may need to increase (or decrease) them. The pulse length may vary by 900% from one 4528 to another! This means C4 and C5 will need to change accordingly.
Instead of a 4528, you may also use the pin-compatible 4098.
Since the Macrovision Killer has its own voltage regulator (U5), you can use a simple 15V/200mA unregulated power supply. U5 may run a little hot. In that case, you may mount it on a small heatsink.
Bill of materials:
# Ref Value ----------- 1 C1 470n 1 C4 47n 1 C5 2n2 1 C6 220n 1 C9 10u 1 R10 47k 1 R5 100 1 R6 2k2 1 R9 680k 1 T1 BC547 1 T2 BC557 1 U1 LM1881N 1 U2 4528B 1 U3 TL082 1 U4 4053 1 U5 7812 2 C7 100u C8 2 D1 1N4148 D2 2 R11 100k R12 2 R3 1M R4 4 C2 100n C3 C10 C11 4 R1 150 R2 R7 R8 6 OUTGND TESTPIN INGND IN VGND OUT VPOS
You can download the PCB (printed circuit board) of in several formats here: JPEG, EPS and HPGL.
The component layout looks like this:
Make sure you don't forget the two wire jumpers (the thick black lines on the component layout)!
The Macrovision Killer is suitable for both composite video and S-Video signals. Composite video is connected like this:
The Macrovision distortion is only present in the luminance (black and white) part of the video signal. So S-Video (S-VHS) connectors can be wired like this:
Pin 4 carries the chrominance (color) and pin 3 the luminance component.
The S-Video connectors are mounted on a separate PCB like this:
The PCB layout is, as usual, available in JPEG, EPS and HPGL.
Turn R11 all the way up (highest resistance) and R12 half-way. You will notice a black band or block on top of your TV screen, because the output of U2A is H for the first few picture lines. If not, try adjusting R12. If you still don't see a black band or block you may need to shift the picture down a bit using the approprate control on your TV set. If your set lacks such a control, you can temporarily solder a 10n capacitor across C4.
If you see a black block, increase R12 until it forms a band over the entire width of the screen. If you already see a band, reduce R12 until it just fills the width of the screen. We now know that the pulse at pin 10 of U2B covers the whole picture data.
Now adjust R11 until the black band just disappears.