Thanks! My primary source for this was Carl Sagan's book "A Pale Blue Dot" IIRC — don't have the folder in front of me to double check, but fairly certain.
Edit: found it!
Here's the excerpt. According to Sagan they sent these instructions up. Given his details on what had to be done to boost the signal upload, it sounds like this really did happen:
"...while taking a photograph of a street scene from a moving car. This may sound easy, but it's not: You have to neutralize the most innocent of motions. At zero gravity, the mere start and stop of the on-board tape recorder can jiggle the spacecraft enough to smear the picture.
This problem was solved by sending up commands to the spacecraft's little rocket engines (called thrusters), machines of exquisite sensitivity. With a little puff of gas at the start and stop of each data-taking sequence, the thrusters compensated for the tape-recorder jiggle by turning the entire spacecraft just a little.
To deal with the low radio power received at Earth, the engineers devised a new and more efficient way to record and transmit the data, and the radio telescopes on Earth were electronically linked together with others to increase their sensitivity. Overall, the imaging system worked, by many criteria, better at Uranus..."
Thanks for the excerpt. I read a couple of Sagan's other books many years ago and I really should read APBD sometime.
Interesting to me, Sagan's "little puff of gas" was borne out in the paper I referenced (not that Sagan needed being borne out!) and that the resulting "imaging system worked ... better at Uranus" was something I hadn't thought of. Per the paper, the Voyagers originally had minimum thruster pulse lengths of 10 ms. In the lab and then on Voyager 1, the Voyager engineers figured out that they could reduce the pulses to 5 ms, thus allowing finer control of Voyager 2's attitude at Uranus (and later Neptune) and probably better image quality than at Jupiter and Saturn. Very interesting - I really should read Sagan's book!