Test Machine 2:

MacBook Intel Core Duo (not later core 2 duo) 1.83Ghz, Tiger, 2GB RAM

Source file: DV PAL clip, 720x576, 25fps. 23 seconds long, interlaced.
File loaded into quicktime 7.1.6

Exported to the following formats. First time is without elgato
hardware, 2nd time is with the Turbo 264:

Apple TV preset: 2 mins 08 sec, 0 mins 43sec
Ipod 640x480 preset: 1 mins 02 sec, 0 mins 31 sec
Ipod 320x240*: 0 mins 40 sec, 0 mins 17 sec

* Apple dont offer a default 320sx240 ipod preset anymore, wheras
elgato turbo 264 offers it as an option for the ipod preset. So for
the non-hardware encoding time, I used manual settings, h264 baseline
single pass. The time for non-hardware assisted multipass 320x240 on
this macbook was 1 min 15 seconds.


As expected, the difference the Turbo 264 makes on this machine is
less, as the machine has a much better CPU than the G5 iMac, so the
non-hardware assisted encodes are much quicker. Wheras the turbo 264
encodes are roughly the same speed as when done on the G5, sometimes a
bit faster, sometimes a bit slower (perhaps due to slower hard drive
in this laptop than the iMac?).

Filesizes and deinterlacing are, as expected, the same as with the
iMac tests. 

Have done some brief tests on what difference this stuff can make to
people's energy use. Two factors make this an electricity-saving
device. Firstly encodes are much quicker, so you can maybe have your
machine switched on for less time. Secondly the CPU in your mac does
much less when letting the turbo 264 do the encoding.

For example this macbook, with battery already charged, and screen
brightness turned off (I use an external monitor), uses around 20
watts when the CPU isnt doing much. When using the turbo 264 this
rises to around the 27-30 watt level. When encoding using quicktime &
no turbo 264, th CPU is under heavy load and around 43 watts are being
consumed, plus it quickly starts to sound like a small hairdryer. I
shall get round to 'writing this up' in video form and sticking it on
carbondown at some point this month.

Other things to think about when considering whether to get a Turbo
264 and reviewing my results:

My tests were done with quicktime, its likely if you use a different
software encoder at the moment, it may already be quicker than the
non-turbo264 times Ive posted.

If you like precise bitrate control (& thus control over filesizes),
so far the Turbo 264 gives you no control of these things, so its just
as bad as Apples presets in that respect.

Although I havent done quality tests yet, some multipass software
encoders could be expected to give better results. Probably more true
at higher resolutions, will test. 

Likewise if you already make 640x480 videos and use a very good
deinterlacer, you may be best to keep using your current deinterlacer
before encoding with the turbo. But if you struggle with this
deinterlacing stuff, the turbo264 provides an extremely easy way to
fix this stuff in an ok way without even thinking about it.

Cheers

Steve Elbows

Reply via email to