different results w/o threads - why?

[mancubit]

i am quite new to x264 (and this board), but i was wondering why are such differences without using threads - if you compare:

w/o. threads:

x264 --bitrate 500 --level 4.1 --keyint 25 --min-keyint 2 --ref 2 --mixed-refs --bframes 2 --b-pyramid --weightb --vbv-init 0.
9 --partitions p8x8,b8x8,i4x4,i8x8 --8x8dct --vbv-bufsize 1500 --ratetol 0.1 --vbv-maxrate 550 --me umh --aud --progress --output "D:\testsequences\small\_outpu
t\flowergarden_h264_500k.mp4" "D:\testsequences\small\flowergarden_uncompressed.avs"
avis [info]: 352x240 @ 25.00 fps (115 frames)
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 Cache64
x264 [info]: profile High, level 4.1
mp4 [info]: initial delay 2 (scale 25)
x264 [info]: slice I:43 Avg QP:41.94 size: 7083 PSNR Mean Y:22.43 U:30.20 V:31.57 Avg:23.86 Global:23.05
x264 [info]: slice P:44 Avg QP:47.44 size: 357 PSNR Mean Y:20.01 U:28.74 V:30.94 Avg:21.54 Global:21.27
x264 [info]: slice B:28 Avg QP:48.94 size: 102 PSNR Mean Y:18.43 U:28.03 V:30.83 Avg:20.01 Global:19.82
x264 [info]: consecutive B-frames: 22.2% 77.8% 0.0%
x264 [info]: mb I I16..4: 11.5% 61.9% 26.6%
x264 [info]: mb P I16..4: 0.2% 0.2% 0.0% P16..4: 48.5% 3.8% 2.7% 0.0% 0.0% skip:44.6%
x264 [info]: mb B I16..4: 0.0% 0.0% 0.0% B16..8: 11.9% 0.2% 0.0% direct: 0.7% skip:87.1% L0:44.2% L1:53.3% BI: 2.5%
x264 [info]: 8x8 transform intra:61.8% inter:55.3%
x264 [info]: ref P L0 73.8% 26.2%
x264 [info]: ref B L0 74.7% 25.3%
x264 [info]: SSIM Mean Y:0.6873443
x264 [info]: PSNR Mean Y:20.530 U:29.110 V:31.147 Avg:22.035 Global:21.399 kb/s:562.00

encoded 115 frames, 44.49 fps, 565.90 kb/s

and w threads:

x264 --bitrate 500 --level 4.1 --keyint 25 --min-keyint 2 --ref 2 --mixed-refs --bframes 2 --b-pyramid --weightb --vbv-init 0.
9 --partitions p8x8,b8x8,i4x4,i8x8 --8x8dct --vbv-bufsize 1500 --ratetol 0.1 --vbv-maxrate 550 --me umh --aud --progress --threads auto --thread-input --output
"D:\testsequences\small\_output\flowergarden_h264_500k.mp4" "D:\testsequences\small\flowergarden_uncompressed.avs"
avis [info]: 352x240 @ 25.00 fps (115 frames)
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 Cache64
x264 [info]: profile High, level 4.1
mp4 [info]: initial delay 2 (scale 25)
x264 [info]: slice I:5 Avg QP:33.11 size: 17105 PSNR Mean Y:29.04 U:33.39 V:33.82 Avg:30.10 Global:30.06
x264 [info]: slice P:61 Avg QP:36.97 size: 3109 PSNR Mean Y:25.83 U:30.93 V:32.35 Avg:27.03 Global:26.80
x264 [info]: slice B:49 Avg QP:39.27 size: 319 PSNR Mean Y:24.99 U:30.69 V:32.36 Avg:26.27 Global:25.97
x264 [info]: consecutive B-frames: 10.9% 89.1% 0.0%
x264 [info]: mb I I16..4: 8.4% 50.1% 41.5%
x264 [info]: mb P I16..4: 0.1% 0.1% 0.2% P16..4: 38.1% 26.9% 19.8% 0.0% 0.0% skip:14.7%
x264 [info]: mb B I16..4: 0.0% 0.0% 0.0% B16..8: 26.2% 1.6% 0.8% direct: 2.1% skip:69.2% L0:24.7% L1:51.9% BI:23.4%
x264 [info]: 8x8 transform intra:48.3% inter:44.3%
x264 [info]: ref P L0 80.3% 19.7%
x264 [info]: ref B L0 81.8% 18.2%
x264 [info]: SSIM Mean Y:0.8950272
x264 [info]: PSNR Mean Y:25.610 U:30.935 V:32.418 Avg:26.837 Global:26.520 kb/s:505.72

encoded 115 frames, 73.77 fps, 508.28 kb/s

with using threads, x264 works far more accurate in keeping the targeted bitrate - whats the reason for that? as far as i know - using threads only affects the encoding speed, not the final outcome - at least i thought that..

[Dark Shikari]

Do note that ratecontrol can be suboptimal on clips of very short length.

[me7]

Multithreading in x264 does affect the final outcome, using more threads is said to lessen the quality a little bit but it's too small to be significant so don't bother.

[mancubit]

Do note that ratecontrol can be suboptimal on clips of very short length.

is it better to limit the bitrate by using -vbv-maxrate?

@me7:
far from it, when you look at the ssim stats, you see that there is a much more accurate bitrate-handling plus a gain in videoquality.

[Esurnir]

Multithreading in x264 does affect the final outcome, using more threads is said to lessen the quality a little bit but it's too small to be significant so don't bother.

Actually from the note about multithreading "It depends" I saw that on certain thread number it was better than single threaded, some other result where degrading.