BDRB or x264 speed drop-off?

[MilesAhead]

Is anyone else noticing a throughput drop-off across the board with one pass abr encoding to BD5?

Seems to me I was consistently getting throughput in the range 1.29x to 1.66x or higher on my quad. On my dual core around .66x optimal.

Now I'm rarely seeing anything over 1.4x and often getting less than 1.0x. I'm wondering maybe if something happened to x264?

Also core utilization is down from nearly maxed out to about 60%. If I do an encode to DVD using FAVC it maxes my cores so I don't think it's the PC.

Just seems odd I'm getting such a drastic reduction across the board.

[jdobbs]

Is anyone else noticing a throughput drop-off across the board with one pass abr encoding to BD5?

Seems to me I was consistently getting throughput in the range 1.29x to 1.66x or higher on my quad. On my dual core around .66x optimal.

Now I'm rarely seeing anything over 1.4x and often getting less than 1.0x. I'm wondering maybe if something happened to x264?

Also core utilization is down from nearly maxed out to about 60%. If I do an encode to DVD using FAVC it maxes my cores so I don't think it's the PC.

Just seems odd I'm getting such a drastic reduction across the board. Not me. If anything I've been getting slight speed improvements as each version of X264 gets released.

[MilesAhead]

Something's really weird. The entire speed range is lower. Like if I was getting 1.29 to 1.75 range for processing now I almost never see 1.4. Last one I just did completed at .71x. Never saw it go that slow on my quad. Try different version of AviSynth etc. and nothing makes a difference. Most peculiar.

It's like something's cutting me down to 2 cores just when I run this. FAVC maxes out as always. Pretty bizarre when my dual core runs BDRB almost as fast as my quad.

[setarip_old]

You might want to check if your system is overheating...

[jdobbs]

Something's really weird. The entire speed range is lower. Like if I was getting 1.29 to 1.75 range for processing now I almost never see 1.4. Last one I just did completed at .71x. Never saw it go that slow on my quad. Try different version of AviSynth etc. and nothing makes a difference. Most peculiar.

It's like something's cutting me down to 2 cores just when I run this. FAVC maxes out as always. Pretty bizarre when my dual core runs BDRB almost as fast as my quad. Things I'd check:

1. Antivirus software. Any recent upgrades? Doing real-time scanning? That cuts my encode speeds in half. I disabled it on my video computer.

2. Indexing enabled on your working drive? This really slows things down. I'd recommend turning it off.

3. Close to full on one of the drives in use? When you get within a small range of filling a drive it really slows down.

4. Run Task Manager and make sure something else isn't using all the CPU time.

[MilesAhead]

Thanks for the suggestions. Trouble with this off the shelf PC, it has some nice features but they put slow memory with a fast CPU. I can't even use high quality settings or the thing will trap out. Limits me to "good" profile.. or at least it did last time I tried. I'll try some other settings.

[chipvideo]

I was wondering if this newer version of bd rebuilder is slower than the previous ones? I noticed that I was getting .75 on my encodes at the highest setting and now I am getting .62 on the highest setting encoding to a bd25.

[MilesAhead]

You might want to check if your system is overheating...

No, it's running cooler. I always have Core Temp in the corner whenever doing any CPU intensive stuff.

I changed out drives to ones with more free space and defragged as jdobbs suggested. Seemed to help a bit. Did a run at 1.24x. At least that's in the same ballpark. I'll know more as I do more encodes.

[jdobbs]

I was wondering if this newer version of bd rebuilder is slower than the previous ones? I noticed that I was getting .75 on my encodes at the highest setting and now I am getting .62 on the highest setting encoding to a bd25. It would be easy to test. The encoding is done by X264 -- and BD-RB has little influence over its speed. You could replace the version of X264 with an earlier one to see if there are any differences in speed.

Frankly I'm not sure I've noticed any speed decreases -- but I haven't really been watching it either.

[jdobbs]

I ran some tests overnight against newer and older versions of X264... identical times within a few seconds. Of course it wasn't scientific and it didn't test every setting etc... but it's close enough.

My guess is that the difference you're seeing is probably just more complexity in the source, or something external to BD-RB and X264.

[chipvideo]

I should have noticed that. Cant compare two different sources. Each source is a different complexity. Only way to know would be to use the same source.

[MilesAhead]

I didn't think to try older x264 since when I tried newer it balked. I just assumed it had to be the exact version. It's easy to forget how big these files are. I think it's likely because I changed out to a HD with less free space and a bit fragmented. I just started an encode with the new release to about 5% done with drives having more room and freshly defragged, to get an idea. Showed about 1.37x which is more in the median(using the same source as I realize that can vary wildly.) Sorry for the false alarm. :)

[jdobbs]

I didn't think to try older x264 since when I tried newer it balked. I just assumed it had to be the exact version. It's easy to forget how big these files are. I think it's likely because I changed out to a HD with less free space and a bit fragmented. I just started an encode with the new release to about 5% done with drives having more room and freshly defragged, to get an idea. Showed about 1.37x which is more in the median(using the same source as I realize than can vary wildly.) Sorry for the false alarm. :)No problem. Switching out X264 is usually a bad idea anyway unless it's just to test something. If you go too far back it will fail because some of the command line options have changed.