Some oscilloscope waveforms and measurements (note that timing measurements collected via Logic Analyzer will be much more precise than those collected from the oscilloscope).
Removing the 75 Ohm load approximately doubles the voltages.
Logic Analyzer Timing Measurements
| ROM Version | H | 3A |
| Pixel Shape | Tall | Square |
| Horizontal Pixels | 720 | 608 |
| Vertical Pixels | 364 | 432 |
| | |
| Using 20.37504 MHz Dot Clock | | |
| As measured at U6B on the CPU Board | | |
| | |
| VSync (U6B-5) | | |
| Low duration | 267.764 us | 212.796 us |
| Period | 16.99948 ms | 16.9614 ms |
| VSync low to HSync low | 3.135 us | 788 ns |
| | |
| HSync (U6B-16) | | |
| Low duration | 28.265 us | 22.764 us |
| High duration | 15.710 us | 14.927 us |
| Pulses per VSync period | 379 | 450 |
| | |
| CSync (U6B-12) | | |
| Low duration | 40.040 us | 34.548 us |
| High duration | 3.931 us | 3.143 us |
| | |
| Vid (U4D-8) | | |
| Front porch (last pixel to start of CSync) | 352 ns | 352 ns |
| Back porch (end of CSync to first pixel) | 4.355 us | 4.361 us |
Here is my attempt to recreate these driver circuits in the Falstad CircuitJS simulator: link (https://www.falstad.com/circuit/circuitjs.html?ctz=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-JAyn7LC-yJcQgIASQAEWggjChwhCcLIqjjDvIVMPZGiSDY6hgmkcAcmQ0DIN4AAjX1KlEMgsEUNBrisGCIX7NsIT-NtxAAGRBeAAH1rAAgBBTje37HBYhU-RTKXdSwF0gyjLorBXQYxyehACjqJguiEAsxxvMAoSoL8H1SmCv99xHcR0kVSKYwiaZREsxTYksod63ESz4tYJIGR9FyW3S7KNnyrLYmCppexy10hh-f04GJYr6GLKq6qsU55QaKsnMwFwIEINBNVNRBTAwNBmBBUaXVmF4HHoKNGDGeAXggO0YQRFE0QiTkOsZGQWsqaoBQS7r2u6sUw2uSyttOyldlKW6wvq3Y8qenIGXidLXSe103oaWDbsHWUXojC7XrVZaTXQ6hokjERih1NUvCueBTEWjAPgoNQpIq+01qdTZfr826axzVRKVKcnJFNUI2gMPyKZkIQnh4CABjIah4FWUZikQbVdVmVR1CK8me3+0ISZqLye2F0JTQZKLWCcnI8pqeXnuWZXwEvMnJAsdhJQoFMrFsOjqtYRxTYqeCPFmYdzDqyU5IEcxrkleNxGUVpkemb2fd9mYBnGaQwDNVRimmUZ+fmDQhZ1uqTb2ut8d2glfo+FMUHiH6ZANqllcpTk-IL-twqpU2RRa9FU-t2O8l4AAPXs2akGYYEocBMSXC51IASwAZ3hAAdXu0AAejrBuITIWILGWkFPVcru+8H3uwAAGiHnv+6H0fvAAeycWIy3b0akxUdB4kQVkLTNDdWnUJcUFiFgM9cEgyF4IA)
It would be a good idea to check to see whether I've copied the schematics faithfully. There is certainly a difference in levels apparent in both in the simulated oscilloscope. (Tip: mouse over the L1_RCA and L210_RCA labels to see which label is which in the simulated 'scope --- note that it may still be hard to tell if you have certain kinds of colourblindness. The L1_RCA has consistently higher voltages everywhere.)
It's interesting to note slightly different circuits to the ones I've transcribed in some motherboard schematics: for the Lisa 1, some different values in this schematic (http://www.bitsavers.org/pdf/apple/lisa/hardware/Lisa_1_Motherboard/lisa-motherboard-enhanced.pdf) as well as a 100R resistor between Q1 emitter and Q2 collector.
Meanwhile the Lisa 2/10 motherboard (http://www.bitsavers.org/pdf/apple/lisa/hardware/Lisa_2_10_Motherboard/050-4032-B.tif) (note: link goes to a TIFF file) shows a single component U2 creating the video signal from CVOUT and VID; I presume this is the hybrid circuit that has been mentioned elsewhere?