The turbo option

[max-holz]

I would know in which way the turbo mode affects the video quality.

[Sharktooth]

Almost in no way. Sometimes it can lower the quality but by a ridiculous and unperceivable amount.

[Sirber]

some would say 0.01db ;)

[CEC]

WITHOUT TURBO
---------------------------
"x264.exe" --pass 1 --bitrate 400 --stats ".stats" --ref 3 --mixed-refs --bframes 2 --weightb --analyse all --8x8dct --threads 2 --progress --no-psnr --output NUL "C:\batch\video.avs"

http://img469.imageshack.us/img469/5012/noturbo2rs.png

WITH TURBO
--------------------
"x264.exe" --pass 1 --bitrate 400 --stats ".stats" --bframes 2 --subme 1 --weightb --analyse none --me dia --threads 2 --progress --no-psnr --output NUL "C:\batch\video.avs"

http://img469.imageshack.us/img469/909/turbo6up.png

Look the marked place!

THE "IN" COLUMN CHANGE! (IT MIGHT BE BAD!)
THE "OUT" COLUMN IS THE SAME! (SHOULDN'T IT CHANGE SINCE THE "IN" COLUMN CHANGE??)
AND THE BIGEST DIFFERENCE: THE "TYPE" COLUMN CHANGE!!!!!! (THAT'S VERY BAD FOR ME!)

This thing is happening only in certain places of the stats files!
You can see that out of the marked area, everything (except the "Q" column which is
always different) is going OK!!!

Make some tests yourself!

I don't use turbo mode!!!

[ChronoCross]

that's because it disables some of the advanced decision options in the first pass. Therefore some of the statistics will be different. Placement of P,B,I frames, reference decisions, etc are all affected by the decision making. but like the x264 devs and users have said before the quality difference is minimal. I think turbo is a good option.

[foxyshadis]

All decisions are redone on the second pass, or it would be rather faster. ;) The first pass applies a rough estimate which the 2nd uses to provide a more precise rate control. Second pass will not use those exact values. Turning off turbo is only useful if you are going to keep the first pass; otherwise the only useful comparison is between the outputs of the two 2nd passes (with and without turbo).

Oh, and line by line comparisons aren't entirely useful because the difference between choosing a frame as a p, b, or bref can be very marginal in a low-movement scene, the frames exist in an ecosystem that adds up to effectively the same quality.