Ok .. I spoke to a reallly helpful civil servant today (I know that's rare! He spent a while trying to dig out the specific data for the RS3's when it was "type approved".. This would give the sound readings from the new car. He said to me that, as in my previous post, the limit for cars now sits at 74dba, but that this is more of a guideline than anything else. He said cars were allowed a discretionary amount over this but there was no specific limit. The reading is taken while the car is in motion, with the microphone at a distance of 7.5m as the car passes at it's closest point.. It is taken at a speed of 30mph at no fixed rpm in an "appropriate gear for vehicle speed" across a 20m stretch of particular tarmac with a certain grade/size of stones etc. This is the test that should be at around the 74dba guideline. This is also the data held for type approval. Another test is performed on the car, stationary inside lab conditions with no atmospheric (other! lol) noise to intefere with the testing.. In the stationary test, the car is accelerated in a gear to a certain percentage of it's rev range.. This varies due to how high the car can rev. The criteria is - the car is revved to 75% of its's rev range if the total is less than 5,000rpm - the car is revved to 3,750rpm if it's rev range is 5,000-7,500rpm - the car is revved to 55% of it's rev range if the total is more than 7,500rpm This second stationary test is not a legal requirement, but is apparently performed on all new cars. The data is only available to buy and there is no maximum limit (although this time the guideline is around 90dba) It is done with the microphone about 0.5m away from the exhaust at a certain angle etc... He said that manufacturers exploit the way the test is set out by doing several things to reduce the noise - active exhaust valves are one, auto transmissions that use a much taller gear (or have upshifted to a lower gear than you would've perhaps been in in a manual car) are another. Both of these are based mainly around throttle position, so he said companies like audi are pretty clever as getting the car "set-up" for the test so it can be as quiet as possible, and he had come across several cases where a new car was tested with no defects and found to be considerably louder than the test cars provided! Hope this hasn't bored everyone to tears - i found it pretty helpful! Some of his data came from this document http://www.unece.org/fileadmin/DAM/trans/main/wp29/wp29regs/R051r2e.pdf In particular pages 24, 25, 26. It goes into detail about position of mic's for the test, conditions and loadsa other variables in types of vehicle etc I've got to hear back now from VW at milton keynes tech dept, who have the data for the rs3 specifically in the tests! Every audi we looked at : 3.2v6, r8 v8, rs4, 3.0tdi etc, were all between 70-73dba..