Originally posted by gosman PTTL is exactly as you explained Sinus. Pre-flash and then the big flash. That is what PTTL is!
Except that the interval between the preflash and main flash is very short and just barely discernible with the unaided eye. What Sinus is seeing is probably the red-eye reduction flash.
I had done some reverse engineering on the P-TTL flash system a recall that the interval between the pre-flash and main flash was of the order of a few tens of millisecond or less and is just barely discernible as being two flashes. Normally the interval exists mainly because of the time it takes to lift the mirror and actuate the iris and for the shutters first curtain to open fully - the mirror needs to be down and the aperture blades fully open for the exposure meter to measure the pre flash but then the mirror needs to get out of the way and the aperture blades closed to the chosen aperture before the shutter is released. Then once the first curtain is fully open the main flash burst will be fired,
The camera will of course need to do some calculations based on the light measured from the preflash burst and then instruct the flash at what intensity to fire the main burst. That however should take far less time than the mechanical activity that has to take place between the preflash and main flash so it does not really add any further delay to what already exists when not using the flash.
The odd thing with the K-3 is that having the flash switched on introduces a delay between pressing the shutter the mirror lifting. That is an interval during which there is absolutely nothing that needs to be done except ordering the pre-flash to go off. So the camera is just sitting doing nothing for several tenths of a second or else making itself busy doing something useless.
It is worth noting that on two other cameras I own, the K-r and K-30, there is no noticeable delay introduced by the flash. Some have mentioned that perhaps the K-3's delay is because it does more complex calculations and gets the exposure right rather than hurrying and getting it wrong. That the K-3 gets it right far more often than previous Pentax DSLRs is true but the flaw in this argument lies in the fact that any complex calculations needed would be done in the interval between the preflash and the main flash burst, which is determined solely by the time needed to physically move the mirror, aperture blades and shutter and not between pressing the trigger and issuing the preflash so it cannot possibly explain the delay we are experiencing.