Lag Measurement

Lag was measured as follows:

A bbc micro was programmed to flash it's screen from black to white for 1 frame every 8 frames.
The 8 frame interval is to ensure that any measurement shows up lags of multiple frames.

It's output was connected to an RGBtoHDMI and the HDMI output was connected to a HDMI to VGA converter to produce an analog VGA output. The output resolution and refresh was set to 720x576 at 50Hz which was supported by the VGA converter I was using and the test was made with no menu on the screen. The output refresh rate must match the input refresh rate (50Hz in this case)
The green video lines from the BBC micro input and the VGA output were then connected to two inputs of an oscilloscope:

Yellow is the BBC source and Cyan is the VGA output
The width of one "burst" is one frame of video and the output is lagging the input by a fraction of a frame.

Zooming in on this shows the output lagging the input by 3.6 milliseconds:
The above value may vary by a few percent in integer scaling mode depending on the source and the screen resolution

This measurement does include the lag of the HDMI to VGA converter so the actual lag of RGBtoHDMI may be less than the above but the lag of a hdmi to vga converter is generally minimal as it is effectively a D to A converter with little storage (no framebuffer).

NOTE: When the menus are displayed, the lag will increase by several frames as the text has to be drawn over the video before sending it out and there isn't enough time to do that without delaying the video. This doesn't affect normal operation as you won't be using it with the menus on screen once everything is configured

It should be possible to use this technique to measure the lag with any HDMI converter.