Hornby BR A4 DCC Sound - how's it done?

Most (all?) UK RTR locos with sound use ESU LokSound chips. Those chips can be run with or without a synchronising sensor (its a software setting in the chip). Most (all?) RTR models are built without a sensor, and instead use a software approximation to the motor rotation speed (known from the motor drive component of the chip). The approximation is usually good, but not perfect.
The cylinder beat will be in the sound file recording and the repeat loop, again in the ESU chip's software. The quality is down to a mix of the sound recording and the skill of the person setting up the ESU software.
The end user has some tweaking available in CV settings, but quite a bit of a sound chip is locked to anyone lacking the original sound project source files.
- Nigel

out of phase with the pistons from the initial installation and would need to be tweaked into phase during an initial run somehow.. And thereafter the correct phase setting would need to be preserved when motive power is off to resume correct operation when next switched on. I suppose they've catered for this. Anyway thanks for the information. Appreciated.
steve

You've missed a few bits: distance of listener from the locomotive, and the human brain's ability to re-synch sound and pictures. So, unless you look and assess things very carefully, the brain is remarkably good at making you believe that things are synchronised. Thus doing all the synch things you mention is probably unnecessary for many installations.
In very round numbers, if you are 100m from a real locomotive, the sound takes about 0.33 seconds to reach you. At a fast walking pace (6m/s), the driving wheels will have rotated 1/3rd of a turn before the sound reaches you. The visual information (speed of light) travels the 100m in near enough zero time. Yet your brain will happily put the two together and suggest the sound and pictures are properly synchronised. (Your brain will actually take quite a while before you conciously hear the sound and see the image - 1 to 2 seconds - but the brain is very good at fooling you into thinking its instant).
There are limits to what the brain will do; get the sound ahead of the image and things go wrong, there is a limit on the delay synchronisation which is accepted, and what happens does vary on the type of sound (lip synch for speech in films/TV has different sensitivity to some other sounds).
- Nigel (Past professional experience researching human perception of audio-synch issues in teleconferencing)

Try low cunning and chocolate.
cheers, Simon