Nope! RF isolation transformers will distort the composite video signal.
The coupling capacitor route would be better, but it does cause a loss of the video DC component.
This may not matter if the TV uses DC restoration in the video amplifier.
The TV must have dc restoration in some form because the DC information is not transmitted,
Actually it is!
Analog Broadcast Television systems go to great lengths to maintain the black/blanking level (In PAL, blanking & black level are the same, so I will use that system)
Sometimes DC coupling can be used, but it is usually impractical, so ac coupling must be used.
The time constant of the RC coupling used is long enough that there is no discernible "tilt" in the waveform over the duration of a line.
This is not the case, over a field, however.
To restore the blanking level, a device called a "keyed clamper" is used, which works at line rate.
Keyed clampers commonly generate a narrower pulse from the line sync pulses, which is then delayed till the "back porch", operating some form of switch which connects that point to zero volts for the duration of the pulse.
As the colour burst is also on the back porch , the device requires colour subcarrier "traps" so as to not mutilate the burst.
This regime is maintained up to, & including, the transmitter modulator, although the reference clamped to may be other than zero volts.
Thus, given levels of Modulation are representative of the original video DC levels.
The result of this is that a simple diode detector at the receiver will produce a video signal with extremely low "tilt" over a field period.
Old style "high level modulated." transmitters went a step further, with "Blanking Level Feedback", where video from diode detectors on the antenna feeders was used to adjust the Tx clamp level, so as to maintain "Blanking level stability".
the only question is whether it's before or after where he's tapped in.
Regarding transformers the problem is the low frequency cutoff, and I'll admit to not having calculated it before posting...
The lower end of the frequency range required depends on what content he's displaying on the screen, as a minimum you need to pass the horizontal sync pulses through so you need to pass the line rate (17 kHz or so), but unless the application has a huge brightness differential from top to bottom then you can probably get away without going down to the 50/60 Hz required to pass full (half)frame information without distortion and a minimum pass frequency maybe 500-1000 Hz might give an acceptable amount of distortion (granted so it's not RF). Looking at what he's displaying he's got maybe 30 characters a line which need perhaps 6 pixels horizontally, so the required upper frequency is around 4 MHz. I guess 4 decades is a lot to ask from a transformer and it's probably not that practical.
Which is why they are not used in practice!
Coupling capacitor value would depend on the impedance of the circuit node, and yes a new DC restoration circuit may need adding if the existing one is further back down the chain.
This dependence upon impedance was used to good effect in the early BW TV days, where vacuum tube circuits were mostly high impedance.
Some very cheap TVs relied upon large coupling capacitors to minimise tilt, & didn't bother with DC restoration at all!
More "upmarket" sets used simple,"diode clamp" DC restorers, which only clamped at field rate, but still did quite a good job.
The set under consideration was made around 1984, so it probably does have DC restoration built into a large IC which also does the job of sync separation.
The date of manufacture of the TV set also casts doubt upon the idea of it being a simple "transformerless" unit, with the "chassis" connected to Mains Neutral.
The dubious advantages of such architecture to tube TV receivers are pretty much non-existent for solid state designs, so it is more likely to be a SMPS design.
If it has a "figure 8" power cord, & a Symmetrical Mains filter with caps connected from both sides to chassis, there is every chance that the chassis will be elevated to 60v w.r. t. Neutral.
Grounding that in some manner will cause a pulse of current in the ground,conductor, perhaps sufficient to
trip the RCD/GFCI.