Mac Pro 2.66ghz 4gb

Post your testing results with HandBrake.
Post Reply
distortedloop
Posts: 23
Joined: Wed Mar 28, 2007 4:00 am

Mac Pro 2.66ghz 4gb

Post by distortedloop »

Wasn't sure if I should start a new thread or not, but I figured my machine was different enough (double the ram) and my rip settings different enough (double the bit rate) from the other Mac Pro thread to not be a meaningful comparison:

Mac Pro, 2.66ghz, 4gb ram
MP4 x264 (h.264 iPod)
640x368, deinterlace, 1-pass, AV bit rate 1500kbps,
AAC 128kpbs

All four cores get used, about 35-40% on each core for a total CPU usage of ~155% average.

I'm getting about 33 frames per second rip time on my current DVD collection (BSG season 1). Speeds are similar ripping to internal drive and external firewire drive, but I imagine both are 5400rpm drives.

These results are with MediaFork 0.8.0b1. I've only ripped a few titles on this new mac at all, my first few rips were with the last version of Handbrake (hadn't heard of this yet) and different settings. Handbrake made the machine roar - it was using 350% CPU speed, and the fans on the pro got very very loud (all those processor cycles heating the machine up!) and it took quite a while after rips for the fans to kick down.

I'm going to re-rip these same titles in Handbrake to see if all that processor time results in faster rips, or the original program just less efficient.

Lots more testing to do, looking for best combination of video quality vs file size. I'm trying to rip files that will play on both my iPod and my AppleTV products. Current settings look great on the big screen TV...I may up the bitrate to 2500 if iPod and aTV can both handle that.
baggss
Moderator
Posts: 886
Joined: Tue Jan 02, 2007 8:21 am

Post by baggss »

One thing you can do to get more performance out of your machine is to run multiple instances of HB/MF (they have to have different app names). I've done this and seen CPU usage on my Quad go from 60% to 95% with no noticeable lag on the rest of the system. You simply have to make sure the file destinations are different discs to avoid wearing your drive mechanism down. Both copies only lose a max of 10Fps when doing this and I get double the ripping capacity to make up for it. Definitely worth a shot.
distortedloop
Posts: 23
Joined: Wed Mar 28, 2007 4:00 am

Post by distortedloop »

That's a good idea. I've only got one internal DVD drive, but I do have a firewire external I can hookup and get it going with.

Fantastic to have an open source app that actually uses all this hardware firepower.
distortedloop
Posts: 23
Joined: Wed Mar 28, 2007 4:00 am

Re: Mac Pro 2.66ghz 4gb

Post by distortedloop »

distortedloop wrote:Mac Pro, 2.66ghz, 4gb ram
MP4 x264 (h.264 iPod)
640x368, deinterlace, 1-pass, AV bit rate 1500kbps,
AAC 128kpbs

All four cores get used, about 35-40% on each core for a total CPU usage of ~155% average.
Same machine, same DVD, all settings the same EXCEPT used FFmpeg ripper.

Slightly higher CPU usage, ~180%, double the frame rate, though, 85-90fps. Resulting file 1.48gb compared to h.264 file of 1.16gb.
jbrjake
Veteran User
Posts: 4805
Joined: Wed Dec 13, 2006 1:38 am

Re: Mac Pro 2.66ghz 4gb

Post by jbrjake »

distortedloop wrote:Same machine, same DVD, all settings the same EXCEPT used FFmpeg ripper.

Slightly higher CPU usage, ~180%, double the frame rate, though, 85-90fps. Resulting file 1.48gb compared to h.264 file of 1.16gb.
Are you really sure all the settings were "the same"? Because it looks to me like you used a different average bitrate...
distortedloop
Posts: 23
Joined: Wed Mar 28, 2007 4:00 am

Re: Mac Pro 2.66ghz 4gb

Post by distortedloop »

jbrjake wrote:Are you really sure all the settings were "the same"? Because it looks to me like you used a different average bitrate...
I was PRETTY sure when I posted that, but the way you ask by emphasizing your question with a REALLY sure makes me wonder.

All the files I am ripping I wanted to be playable on both iPod and aTV, and I believe the maximum bitratge is 1500 for mutual compatiblility, so that's what I was using. (correct me if I am wrong on that - I haven't done a lot of searching to find a definitive answer on the highest quality mutually playable settings.)

However, I was tired, ripping many files to test with both handbrake and mediafork, and deleted most files after testing, so I can't pull the files and check them to be sure.

Bottom line, I'll have to do another rio with the FFMpeg converter and ensure all settings the same as h.264 were, and I'll repost.

I'm curious though, the results did not surprise me, because I've read hundreds of times that the FFMpeg encoder was considerably faster than the h.264, while h.264 produced smaller high quality files but at the expense of longer rips. Is that not correct?
jbrjake
Veteran User
Posts: 4805
Joined: Wed Dec 13, 2006 1:38 am

Post by jbrjake »

You said you used the same settings on both.

If the sizes are different, how exactly could you have used the same bitrate for both? Bitrate: the number of bits per second. If you were encoding the same video both times, the number of seconds stays the same. But the output size was different. That implies to me you aren't using the same number of bits...
distortedloop
Posts: 23
Joined: Wed Mar 28, 2007 4:00 am

Post by distortedloop »

jbrjake wrote:If the sizes are different, how exactly could you have used the same bitrate for both? Bitrate: the number of bits per second. If you were encoding the same video both times, the number of seconds stays the same. But the output size was different. That implies to me you aren't using the same number of bits...
Ah, of course. Yes, that makes perfect sense, and further testing with more careful attention to what I was doing shows this to be accurate. Where I confused myself was the higher quality with smaller files, but that means you can use a lower bitrate to get the same visual quality, as opposed to the same bitrate giving higher quality, correct?

Thanks for setting me straight.

I ran 20 different rips of the same file just to compare final products to decide what's best for me. Of the rips I watched, in general, at the same settings of everything other than encoder type, h.264 produced ever-so-slightly smaller files, and had about 1/3 of the fps of ffmpeg while ripping. I've yet to visually compare them all, though.
jbrjake
Veteran User
Posts: 4805
Joined: Wed Dec 13, 2006 1:38 am

Post by jbrjake »

Yep, you've got it down.

The whole ffmpeg/xvid vs. x264 thing ends up being a personal value judgment. Me? I'll use x264 and just let it run overnight. Hell, sometimes I'll turn on advanced options that slow me down to 1 or 2 fps and let it run for over 12 hours. But ffmpeg is certainly a *lot* faster, and will, in some situations (movies with lots of grain or noise), even give better quality. Plus, it's a lot easier to decode, so it won't stress your processor as much. Another thing is scalability. In my opinion, x264 compression artifacts, when scaled up on a 1080p display, are a lot less jarring to the eye than ffmpeg or xvid compression artifacts. Other people will feel differently. It's all subjective, so just trust in your forthcoming visual comparison and make your own cost-benefit analysis.
Post Reply