It is fairly easy to raise a rasterized signal on any oscilliscope so long 
as you can squirt a Z-axis into the scope, and tehn set up the X and Y 
axis from soem kind of external raster timing generator.

This is lot limited to video signals however, and can be applied to a 
wide variety of digital and analog signals, a good example being the use 
of raster generators to hunt down the timing of digital or multi-level/gray 
level signals. Virtually anything with a repeating sequence, data 
streams, encrypted phones, etc can be examined, and you can resolve 
at least some limited parameters while perhaps not reconstructing the 
signal.

Also, very often the signal coming off your receiver has to have the 
signal levels adjusted to a usable amplitude, or as you say the polarity, 
but we also have to be sensitive to the making sure the scope has 
enough bandwidth to "draw" the raster, but also a fast enough phosphor 
(but not too fast) so you can actually see the image.

With some of the newer PC based TV recivers such as WinTV you can 
actually use the card to sample the suspect video and then adjust the 
image timing and reconstruct the images right to the hard drive. A quick 
and dirty way to do this is to simply apply the output from your radio to 
a RF modulator, and the feed the RF signal (usually channel 2, 3 or 4) 
into the WinRadio card as it adjust the amplitude for you but you have 
to give it a signal of the correct polarity.

Remember that over-the-air video will be one polarity, and the opposite 
polarity for conducted video. But, either way it is CRITICAL that you 
measure the signal to determine what polarity it is using, and then 
perform some basic time domain measurements. Then take these time 
domain measurement, drop them into the WinTV software, and be 
amazed as you image starts to show up.

It is normal for the image to roll as you will usually have an imperfect 
sync lock.

Most modern spectrum analysers such as the ESA and PSA actually 
have video capabilities where you can define the time of a signal, then 
set both the video line sync, and amplitude triggers so that you can 
obtian a perfectly sync'ed image.

If you try this you could get it to work with a scope with onely a few 
MHz of bandwidth, but something with a bandwidth of over 25 MHz is 
best, and over 50 Mhz is better yet.

-jma