[mythtv-users] Video Noise and VHS-to-DVD conversion (long - only read if interested)
Brian Wood
beww at beww.org
Fri Jan 5 17:09:21 UTC 2007
(Or: Why does my transfer look like crap?)
This subject was brought up a few days ago, and I've had some private
discussions with a few folks about it. This post is an attempt to
give some practical tips that will at least help to minimize the
perceived noise in a DVD made from a VHS tape. Most of these tips are
pretty obvious, but you'd be surprised how often some of them are
overlooked.
It is really more of a theoretical discussion, since most of the
factors involved are not within the control of the average (or even
above average) consumer. It's written from an NTSC/60 hertz power
perspective because that's where I live, but the principles are
pretty much the same everywhere.
"Noise" is really a very complex topic. Ultimately the true noise
floor of any system is limited by the laws of quantum physics and
the background of cosmic radiation left over from the Big Bang (or
"creation", depending on your particular religious beliefs).
Practically it is limited by the noise figures of the electronic
systems in use and the absolute temperatures of those systems. But
that's not of much interest here, and is certainly beyond our control.
In video systems "noise" is really anything that is not the picture
you want to see. This includes true "noise" in both the luminance and
chrominance channels and "interference" of several types, which is
generally but incorrectly called "noise".
---------
VHS-to-DVD transfers are really a worst-case scenario, since you are
dealing with video in both the analog and digital domains. We'll
start with the VHS machine operating in the analog domain:
Since you will no doubt be starting with an existing tape it won't do
too much good for me to tell you that VHS recordings should be made
at the highest possible speed on good quality tape and the tapes
should be stored properly in a controlled environment. If you are
using a tape recorded at 6-hour speed and stored in your garage for
20 years of summer and winter there's not much you can do about it now.
The heads on VHS machines should be clean and demagnetized. This
includes not only the video heads but the control-track and audio
heads as well as the entire tape path. Unfortunately most consumers
never clean their heads and if they do they use a "cleaning tape"
which can often do more harm than good. The best thing to do is
manually clean the machine, but it is easy for an inexperienced
person (or even a veteran tech) to damage the heads while cleaning them.
All tapes, but especially ones that have been stored for a while,
should be shuttled back and forth a time or two to eliminate
"stiction" which causes the tape to stutter, which causes various
problems like wow and flutter, and makes the servo systems work too
hard and thus "hunt".
-------------
Use good quality cables and connectors. This does *not* mean spending
way too much on overpriced "monster" cables, but just using decent
cables with good shielding. Gold connectors are used by some, not
because gold is a better conductor than copper (it isn't) but because
gold is very resistant to corrosion, which will make for erratic
connections. The best conductor is actually silver, followed by
copper and then aluminum. Silver oxide actually conducts almost as
well as silver itself but it looks "dirty" and can hold moisture
causing bad performance.
Unfortunately some "gold" cables and connectors sold contain not a
single molecule of gold. They are simply "gold colored" because
ignorant consumers know enough to equate gold with quality, but
nothing more. It is marketing, not engineering.
Stranded cables are better for several reasons, including resistance
to breakage and the "skin effect" where RF energy tends to flow along
the outer "skin" of the conductor. While you might think of video as
"signal" and not RF, the 3.5Mhz frequency of the color subcarrier has
been used for decades by amateur radio operators to span continents
under the right conditions (80 meter band), it is really RF in every
sense of the word.
The RCA connectors used for consumer video really suck. They were
originally designed for low-level audio and have bad impedance
characteristics at RF frequencies, which includes the chroma
components of video as well as high-frequency luminance information.
Unfortunately most consumer gear uses these connectors because they
are cheap, so you don't have much choice. A "high-quality" RCA
connector is right up there with "military intelligence" and
"microsoft works".
The BNC connectors used for professional video systems are much
better, but they cost more. The solder type are a real pain to
install properly and the crimp type require an expensive crimper.
They are available in both 50 and 75 ohm varieties but most are 50-
ohm impedance. Remember that the impedance is determined by several
factors including the matching components, so a 50-ohm connector can
actually present a 75-ohm impedance in a properly designed circuit.
75-ohm BNC connectors are normally only encountered in expensive test
equipment, and the sockets are usually damaged by some idiot jamming
a 50-ohm plug into them.
The best connectors for video are actually the PL-259 "UHF"
connectors (don't know why they call them that, they are not rated
for actual UHF frequencies (defined as 300-3000 Mhz.)). They are
hardly used at all anymore though, because they are a pain to install
and use, and take up a lot of space.
-----------
Make sure all your equipment is properly grounded. That sounds easy
but it is not. The "grounding" section in Tremaine's Audio Cyclopedia
is hundreds of pages long. It is real easy to get ground loops since
each cable has it's own ground, as do the power cables, TV cables and
any USB or firewire cables. In NTSC systems AC hum from 60-cycle
power shows up as two horizontal bars floating slowly up the screen.
It would be stationary if the 3.579545 color subcarrier and the
horizontal scan rate derived from it was 3.6000. The reason there are
two bars and not one is that most power supplies are full-wave
resulting in a 120-cycle hum signal and not 60-cycle. Switching
supplies with inverters running anywhere from 15-60 khz. cause their
own problems, as these frequencies fall within the video bandwidth.
You might have to violate some electric codes by lifting the third-
wire ground from one or more power cables to eliminate AC hum. The
best solution is an isolation transformer but few consumers have one
of those. "Humbuckers" and optical isolation units can break up
ground loops, but again not many consumers have access to those.
Unfortunately about all you can do if you get grounding problems is a
trial-and-error process of changing things around to try and make it
go away. Very few consumers have good grounding systems in their
homes (buried radials and 6" copper strap every 2 feet would be a
good start).
-----------------
So: You are now playing back your well-recorded, well-stored high-
quality VHS tape on a clean machine using good-quality cables and
connectors, all properly grounded, right?
In a pig's eye.
In fact you are playing back a slow-speed recording made on 2-dollar
K-Mart-special tape stored in your garage for 20 years, on a machine
that hasn't been cleaned since you purchased it sometime during the
Reagan administration, connected to your computer using whatever
cables you found in your sock drawer and grounded to a steam radiator
with a piece of fishing line.
But at least now you know *why* it looks like crap, even for a VHS
recording.
It's time to enter the Digital Domain:
------------
End of Part 1.
Honestly I don't know if this is of any use or even interest to anyone.
But at least it is different from:
"I am new to Linux and MythTV. Why is this not working for me? BTW,
what is Gossamer?"
Brian Wood
beww at beww.org
More information about the mythtv-users
mailing list