Log in

View Full Version : Can anyone tell me what encoding speed to expect with a 3.2Ghz P4


fagankev
19th February 2004, 19:39
I have just ordered a new system off dell that has a 3.2Ghz p4 along with 1024Mb of Ram.
Do any users have a similar setup and if so what encoding speeds can I expect when using CCE.

Thanks.

Pyscrow
19th February 2004, 20:56
Dunno about CCE, but my P4-3.0 with 1024mbyte is pretty much real time speed with TMPGenc.

Kedirekin
20th February 2004, 14:16
I've seen posts that say your encoding speed in CCE will be roughly the same as your processor speed in GHz. I think that's a maximum speed though; encoding speed can vary a lot depending on filters and what-not.

Also, this doesn't necessarily mean that you'll encode a 2 hour movie in 40 minutes. If you do multi-pass (N passes) for example, it will take N+1 times as long (i.e. 3 passes on 2:00 movie would take about 2:40).

P0l1m0rph1c
22nd February 2004, 00:25
Well, i have a P4 2.0 and I have ~1.25x realtime doing mpeg2 -> mpeg2 direct transcoding (with avisynth).

Logically, you should achieve sth like 2x realtime in your box, probably more.

r6d2
22nd February 2004, 03:00
Originally posted by fagankev
I have just ordered a new system off dellSince it's a Dell, I guess 0.1 RT on average. :) (just kidding)
Do any users have a similar setup and if so what encoding speeds can I expect when using CCE.There are plenty of reports here (http://forum.doom9.org/showthread.php?s=&threadid=66134&perpage=20&pagenumber=1) too in case you want to check.

OvERaCiD23
23rd February 2004, 07:17
Originally posted by P0l1m0rph1c
Well, i have a P4 2.0 and I have ~1.25x realtime doing mpeg2 -> mpeg2 direct transcoding (with avisynth).

Logically, you should achieve sth like 2x realtime in your box, probably more.

I believe he should have a much higher speed than that. I encode ~2.3x real-time on a 2.3ghz AMD (OC'd, of course). This is only on 512MB of RAM as well (although speed is really limited by processor here). I would think 3x real-time is possible on a standard AviSynth script (no filters, encoding 24fps material). This is using CCE 2.66 as well, which has optimizations for P4s. I guess we'll see the speeds once you break that bad boy in. ;)

LB
23rd February 2004, 21:50
Originally posted by OvERaCiD23
I believe he should have a much higher speed than that. I encode ~2.3x real-time on a 2.3ghz AMD (OC'd, of course). This is only on 512MB of RAM as well (although speed is really limited by processor here). I would think 3x real-time is possible on a standard AviSynth script (no filters, encoding 24fps material). This is using CCE 2.66 as well, which has optimizations for P4s. I guess we'll see the speeds once you break that bad boy in. ;)

Well not if he is transcoding :p (can CCE even do that :confused: ) But yeah, as the post far far above said, your CCE speed is roughly the same as your CPU speed, or it should be. If it's not, then you're doing something wrong.

fagankev
24th February 2004, 00:46
cheers lads. Being dell and all i'm still waiting in the system to arrive 2 weeks later. Ill let you know how i get on.

moon1234
21st March 2004, 07:01
Encoding speed will vary based upon encoding bitrate, final resolution, how clean the source is, what the source is (AVI or MPEG), resolution of the source, etc.

I have a P4 2.53Ghz machine, simple two drive RAID 0, 512MB of Ram.

I make both SVCD's at 480x480 and DVD's at 720x480. Here are some examples of my encoding speeds.

From Divx/Xvid Avi 576x432

1st source is fairly noisy and uses the following AviSynth 2.5 script:

AviSource(path to file,false)
ConvertToYV12() #used with DIVX 3.11 sources
BlindPP(cpu=4,quant=4)
BicubicResize(480,480,0,.75)
AddAudio()
ConvertToYUY2()

For an SVCD resolution of 480x480 I normally see an encoding speeds of around 1.1-1.4. For DVD Resolutions of 720x480 I see encoding speeds of around .9-1.2.

If I substitute the noisy DIVX avi for a pristine HDTV XVID transcode and feed that into CCE via the same script from above I see encoding speeds between 1.8-2.5.

So my point is that unless everyone who tests uses the EXACT same source to feed into CCE, the results will be all over the board.

-Moon1234