Method
The method involved in obtaining this kind of data is simple. Video cameras are chronometers, in a way. They record each frame, that composes the videos, on regular intervals. For example, on a video recorded with 50 FPS ( frames per second ) , each frame was captured 1s/50 = 0.02s after another. With established reference marks, one can measure how much time it took from a vehicle to travel from point A to point B.
The precision depends on having a flawless video ( the frame flow must be uninterrupted and without any glitch) and on how good and reliable the chosen reference marks are.
With the typical relatively low frame rates that videos are captured on ( 25 to 50 FPS ) , it's not very common to have the cars exactly on the chosen reference mark, every time. Usually, the car is a small amount behind the reference mark in one frame and, on the next one, it is already ahead of it. But that's not a problem that can't be bypassed. All it takes is to make a fairly easy and accurate estimate. For example, if a car is around 25cm behind a mark on one frame and, on the next one, it's around 0.75cm ahead of it ( meaning the car travelled around 1m , in that interval ) then it's possible to deduce that it took the car an additional 1/4 of the time between the frames to reach the chosen mark. So, in case such video was recorded with 25 FPS , the interval between the frames is 0.04s, and the additional time would be of 0.01s, on this example.
Over the initial and final reference marks, I estimate a total error of around 0.5 of the interval between the frames. That is, I estimate that a car is no more than a '0.25 frame' off the initial mark, and another '0.25 frame' off the final mark. On a typical 25 FPS video, that margin of error corresponds to 0.02s
I have done such measurements, literally, dozens of times and validated the results with the speedometers data available on some videos. By getting the speedometer data ( carefully synchronising it with the images ) and calculating the average speed of several car throughout a specific segment, then it's possible to verify if the average speeds and the time taken, through such segment, correlates, as they should, given that the space travelled was roughly the same ( minor differences in the travelled space occurs because the racing lines are not identical ).
My time measurements never yielded more than 3km/h difference between the extrapolated and the actual average speeds, taken directly from the speedometers ( which also aren't 100% accurate ) but, almost always, the discrepancy is no greater than 1 or 2 km/h, when there is one.
The segment I measured on Campsa is between these marks:
I could not find any pole or fast qualifying lap prior to 2002. The best that I was able to gather was an early 2001 qualifying lap from Häkkinen, which was 0.7s slower than his best, and a 1997 FP2 lap from Frentzen, which was 0.9s slower than the best he did on qualifying.
It's not easy to make an accurate measurement on Frentzen's lap but his time through the segment is around 2.83s, and that yields an average speed of 229km/h. Frentzen drove a very good line through the corner and used most of the apex kerb but, obviously, that's not the best that the 1997 Williams FW19 could do, as Villeneuve's pole was 1.2s faster, but nonetheless it's a great performance considering that the FIA significantly reduced the aerodynamic allowances of the cars ( decreased the diffuser's volume, raised the ride height among other things ) after Imola 1994. On a 1998 edition of the Race Tech magazine, it was written the following::
"Before the Ayrton Senna tragedy of 1994 led to a succession of measures which progressively reduced downforce, F1 cars generated in the region of 4000lbs of downforce at 165mph in Monaco trim. In low downforce trim at Monza the cars were producing something approaching 4000lbs at 200mph. The lift(read:downforce)/drag ratio was as good as 4:1.
This year (1998) in Monaco trim Formula one cars are producing around 3000lbs downforce at 165 mph. At 200mph at Monza they are generating in the region of 2500lbs. At best the lift:drag ratio is something like 2.75:1"
Considering that the 1993/1994 cars had around 33% more downforce than on 1998, it's feasible to imagine that they were going significantly faster than Frentzen's 1997 Williams.
Concerning Häkkinen's 2001 qualifying lap, he also drove a good line and ran over the apex kerb but his lap was 0.7s from his best time in that session, so this likely doensn't represant all that his car was capable of. I measured a time of 2.84s which is only on par with the previous one, of Frentzen, and shows the effects of the treaded tyres and narrower track of the cars.
2002
I found 2 laps from the 2002 qualifying session, Button's best lap and Schumacher's pole. Button's lap is just as fast as the 2001 of Häkkinen but Schumacher was going significantly faster with a time of 2.7s ( extrapolated average speed of 240km/h ). Both drivers went slightly over the apex kerb.
2003
I measured 3 qualifying laps. Schumacher's from Friday's Qualifying, Alonso's from Saturday's Qualifying and a partial lap of Villeneuve which I cannot know the session it's from. Only Schumacher missed the apex while Alonso was already carrying race fuel onboard. Villeneuve's time is 2.87s, Alonso's 2.82s and Schumacher's 2.71s.
2004
This is the year which I found the most qualifying laps to measure. They are : Baumgartner's, Bruni's, Pantano's, Alonso's and Schumacher's laps from both qualifying sessions as well as the final qualifying lap of Sato and Trulli and the first qualifying lap of Panis, Heidfeld and Massa.
I'll not post the times here ( check the overall ranking, on the next page, for that ) but will rather describe the driving lines.
_Alonso used all of the kerb on his Q1 lap but missed the apex on his Q2.
_Bruni used very little of the track, on the entry, and missed the apex during his Q1 lap. On his Q2 lap, he used more of the track but still slightly missed the apex
_Pantano was conservative on the entry but went halfway into the apex kerb, on his Q1 lap. During his Q2 lap, Pantano made a big mistake on the previous corner and went into the gravel. Consequently, he had very dirty tyres when cornering through Campsa besides leaving a lot of track to be used on both the entry and the apex.
_I cannot see Heidfeld's line very precisely on his Q1 lap because he had the on-board camera attached to the nose of his car instead of the more typical "T-cam", on top of the car. But it seems he missed the apex by a small amount.
_Baumgartner left a lot of the track to be used on both of his laps (on the entry and the apex)
_Massa's used almost all of the apex kerb on his Q1 lap but left a lot of the tarmac to be used on the entry.
_Schumacher used practically all of the tarmac on the entry and went slightly over the apex kerb, on his Q1 lap. During his Q2 (pole) lap, he had a similar line on the entry but missed the apex by a significant amount.
_Panis used very little of the track on the entry and went slightly over the kerb, on the apex, during his Q1 lap.
_Trulli left a little of tarmac to be used on the entry but went halfway into the apex kerb, on his Q2 lap.
_Sato used all of the track without going into the apex kerb
It must be noted that Q1 didn't define the grid but just set the running order for Q2 and, the latter, already had the cars with race fuel onboard.
2005
I gathered 3 laps from Q1 (no race fuel onboard) : Ralf Schumacher's, Liuzzi's and Räikkönen's ( aggregated "pole" ).
Liuzzi used all of the track on the entry but missed the apex by a very big amount. R.Schumacher used all of the track and went almost halfway into the apex kerb. Räikkönen left a little bit of tarmac to be used on the entry but went over almost all of the apex kerb.
The aggregated times of Q1 and Q2(race fuel onboard) defined the grid in that year, so everybody was pushing 100% on Q1, already. Liuzzi and Schumacher had identical times of 2.72s and that is significantly slower than Räikkönen's 2.62s ( a whole tenth of a second ). As you'll see on the overall rankings, that is one of the fastest laps ever, through Campsa.
2006
I only found Alonso's pole lap from the qualifying session but decided to also measure an FP3 lap of Montoya which had him almost flat-out through Campsa.
Montoya used all of the track and most the apex kerb while Alonso used all of the track, on the entry, but slightly missed the apex. Despite the small error, Alonso's, together with Webber's 2010 pole, was the fastest one measured (2.56s), overall. From the audio, it seems he was flat-out and that is made even more likely considering that Montoya was almost full throttle with a time 0.13s slower through the same segment.
2007
Besides Massa's pole, I also measured Trulli's fastest Q2 lap (his best on qualifying as they carried more fuel on Q3). Trulli didn't use all of the track on the entry but went halfway into the apex kerb. Massa used all of the track but didn't go over the apex kerb. Trulli, despite being on low fuel, had a time of 2.75s versus Massa's 2.71s.
2008, 2009 and 2010
I only found the pole laps from these years. Despite using all of the track and going slightly over the apex kerb, Räikkönen was quite slow(2.78s) in comparison to most other years, on his 2008 pole. Button also used all of the track, on his 2009 pole, but barely touched the apex kerb. His time(2.67s) was much better than Räikkönen's, on the previous year, on a car that was supposed to have considerably less downforce than the 2008 one, but the infamous DDD (double deck diffuser) as well as the slick tyres helped him to have a quick drive through the corner.
As stated before, Webber's 2010 pole was the fastest lap measured, together with Alonso's 2006 one. Also like Alonso, Webber used all of the track on the entry but also missed the apex of the corner.
2011
Two Q3 laps measured. Alonso's fastest and Webber's pole. Both had very similar performances using most of the track and with an equal time of 2.72s. Their actual average speeds, taken from the speedometers, were pretty similar too with 239km/h for Alonso and 240km/h for Webber.
2012
I've analysed 3 Q3 laps from this year. Hamilton's deleted pole, Maldonado's P2 and Räikkönen's P5.
Räikkönen's used almost all of the track on the entry and went halfway into the apex kerb. He also had two oversteer corrections. Hamilton used all of the track but did not run into the apex kerb. Maldonado could've used a bit more of the track on the entry and missed the apex by a tiny amount.. An average speed of 232km/h measured for Räikkönen and a time of 2.8s. Maldonado had a roughly similar time with 2.81s while Hamilton was slightly faster with 2.77s.
2013
This is an interesting case. I've only found 2 qualifying laps from Q3 and, curiously, the pole lap drive through Campsa was significantly slower than another Q3 lap, from Alonso.
Rosberg used pretty much all of the track on the entry but slightly missed the apex while Alonso used literally all of the track on the entry and missed the apex by a tiny amount. I couldn't accurately verify the synchronicity of the speedometers but, as it is, they yielded an average speed of 242km/h for Rosberg and 248km/h for Alonso. The Spaniard's time was 2.62s and Rosberg's 2.68s, which are in very fine agreement with the speedometer data, by the way.
2014, 2015 and 2016
The cornering capabilities of the cars significantly dropped during those years and their times were on par with that of a Formula Renault 3.5!
The driving was pretty similar in all those 3 years with the pole setters using most the track on the entry and slightly going over the apex kerb. Their times and measured average speeds are on the overall rankings and only the 2016 lap was faster than Conor Daly while testing a Formula Renault 3.5, on 2012. You can see more of that on this video
2017
I've measured two of the fastest laps of this year's testings. Bottas' 1.19.310s and Räikkönen's 1.18.634s.
Bottas used all of the track on the entry and went slightly into the apex kerb. Räikkönen did a pretty similar line but used less of the track on the entry.
Their times were roughly the same and very close to the fastest ones, the 2006 and 2010 pole laps.
