For all you video editing buffs?

  • Thread starter Thread starter Chris
  • Start date Start date
C

Chris

Hi, so we took 2 systems similar specs both P4 3GHz and used MGI video wave
5 on them both to create a 1.5 hour DVD. This software is dual thread so
will take advantage of the Hyperthreading.
System 1:
P4 3GHz 533MHz bus, Gigabyte GA8-INXP, 2Gb PC1066 RD Ram, ATI all in wonder
Radeon 9700.
System 2:
P4 3GHz 800MHz bus, Gigabyte GA8-KNXP, 2Gb DDR 400, ATI all in wonder Radeon
9700.

Both have Western Digital 1200JB 120Gb 8mb cache hard drives fitted.
We took a 4.02Gb mpeg2 recording with a duration of 1.46 hours and turned
this into a DVD image.
System 1 took 56 minutes 23 seconds,
System 2 took 54 minutes 38 seconds.

We had thought about building an AMD XP3000+ machine as well but as they do
not have Hyperthreading the test would be pointless and no real result would
be achieved. We have tested a P4 2.8GHz 533MHz bus that does not have
Hyperthreading and it was 6 minutes slower than system 2 and we know the AMD
does not match this because none of the software is optimised for it.

Chris
http://www.ckccomp.plus.com/site/page.HTM
(e-mail address removed)
 
I want to get into video editing but I want a system that would cut
those times in half. What would I need to do that? A 4-way Opteron or
4-way Xeon with SCSI drives?
Click to expand...


If speedy encoding is wanted, consider better software. MGI videowave is FAR
from good/speedy video software..
 
Stacey said:
If speedy encoding is wanted, consider better software. MGI videowave is FAR
from good/speedy video software..
Click to expand...
That would not make much difference encoding a mpeg 2 to a, well mpeg 2
 
Hi, so we took 2 systems similar specs both P4 3GHz and used MGI video wave
5 on them both to create a 1.5 hour DVD. This software is dual thread so
will take advantage of the Hyperthreading.
System 1:
P4 3GHz 533MHz bus, Gigabyte GA8-INXP, 2Gb PC1066 RD Ram, ATI all in wonder
Radeon 9700.
System 2:
P4 3GHz 800MHz bus, Gigabyte GA8-KNXP, 2Gb DDR 400, ATI all in wonder Radeon
9700.

Both have Western Digital 1200JB 120Gb 8mb cache hard drives fitted.
We took a 4.02Gb mpeg2 recording with a duration of 1.46 hours and turned
this into a DVD image.
System 1 took 56 minutes 23 seconds,
System 2 took 54 minutes 38 seconds.
Click to expand...

Everyone keeps gushing about RAID configrations. I stupidly wanted to
do this and bought 3 identical drives , and then promptly as usual
forgot all about it.

Then as I was editing video with UGH Pinnacle - incredibly intuitive
and idiotically easy to learn interface is the main attraction to me,
I grew frustrated at the agonizingly long times it took to load video
at various points up from the HD.

By then I had swapped the 3 identical drives for 3 different size by
hugely larger drives. I now have a 200 gig , 120 gig and one 60 gig
left.

Now do people editing see a killer difference in RAID while video
editing that makes it moronic not to use it? Or are the difference in
overall usage not that critical?
 
RAID is very useful for capturing raw, uncompressed video. It's not
very important if compressing on-the-fly to MPEG.

For editing the most beneficial drive arrangement is to have the
source and destination files on different physical drives, on
different IDE channels. Reading from and writing back to the same
RAID array (if a simple RAID 0) is in my experience slower than two
(or more) seperate drives or another drive in addition to the RAID
array. The difference isn't substantial though, and certainly not
time-critical like when capturing. Perhaps the most beneficial to the
user who wants to keep using this same computer while it's running a
video editing job is to have the OS & Application drive seperate from
the video editing drives. In other words, have many drives, each with
their own purpose, on as many IDE channels as possible.
Click to expand...

Thats interesting because Id read stuff at sites how its almost
essential to use RAID for editing at various sites but saw no
objective test about it.

Maybe its Pinnacle. Everyone keeps saying Vegas and other editors are
way better. However they just dont seem very intuitive. In fact they
really look weird when you load them up. It might help if I actually
sat down and read a manual once in a while.
 
kony wrote:



By this you mean raw analog not to be confused with "raw" DV.
Click to expand...

No, there is no difference, all video is written to the HDD digitally
regardless of analog vs digital input, it has the same data rate when
"all else" is equal.

I mean uncompressed video. If you're ripping a DVD or other
compressed source to the HDD, still compressed, then of course the
data rate is substantially lower.


Dave
 
Thats interesting because Id read stuff at sites how its almost
essential to use RAID for editing at various sites but saw no
objective test about it.

Maybe its Pinnacle. Everyone keeps saying Vegas and other editors are
way better. However they just dont seem very intuitive. In fact they
really look weird when you load them up. It might help if I actually
sat down and read a manual once in a while.
Click to expand...

It certainly depends on the exact task. If you're doing multiple
passes, saving the editing work in uncompressed format, then RAID will
help. If the output of the editing is being compressed, one-pass or
final pass to it's final distributable format, then the compression
(in addtion to the editing task) is more of a "bottleneck", lessens
the benefit of RAID (0) performance increase.

Video editing has large, divided camps, lying many places between
profesisonal and casual use... Some people will claim there isn't even
significant reason to capture uncompressed or at least in MJPEG, while
others will laugh at that notion, which can also depend on the quality
of the source. Personally, I always avoid compressing twice unless at
a very low compression rate.


Dave
 
kony said:
No, there is no difference,
Click to expand...

Stacey is correct. Raw DV is dv that is stored by itself (in a .dv file), as
oppossed to being wrapped in something like avi (audio video interleave).
This is still DV, and as such is still compressed. This is quite different
from raw video coming from an analogue source.
all video is written to the HDD digitally
regardless of analog vs digital input, it has the same data rate when
"all else" is equal.
Click to expand...

That is correct, but with DV you are simply making a digital copy. The term
DV capture is a bit of a misnomer really, all you are doing is copying
digital data from the tape to the hard drive. Analogue capture is quite
different, you are taking an analogue signal and converting it to digital.
In both cases it ends up in digital form, but the result is very different.
In the case of DV you have a relatively small file because compression is
used. In the case of a raw analogue capture you have a bloody huge file, and
in the case of an analogue capture which is compressed (using something like
MJPEG) you a relatively small file. That file will in no way be similar to
the DV file though.
I mean uncompressed video.
Click to expand...

Yep, that's all Stacey was clarifying.

Gareth
 
Stacey is correct. Raw DV is dv that is stored by itself (in a .dv file), as
oppossed to being wrapped in something like avi (audio video interleave).
This is still DV, and as such is still compressed. This is quite different
from raw video coming from an analogue source.
Click to expand...

When I write "raw uncompressed video", I mean just that, not
compressed video, not the output from a camcorder, etc, else I would
have written as much.
That is correct, but with DV you are simply making a digital copy. The term
DV capture is a bit of a misnomer really, all you are doing is copying
digital data from the tape to the hard drive. Analogue capture is quite
different, you are taking an analogue signal and converting it to digital.
In both cases it ends up in digital form, but the result is very different.
In the case of DV you have a relatively small file because compression is
used. In the case of a raw analogue capture you have a bloody huge file, and
in the case of an analogue capture which is compressed (using something like
MJPEG) you a relatively small file. That file will in no way be similar to
the DV file though.
Click to expand...

And thenin lies the key, that HDD rate isn't of utmost importance when
copying something already previously captured. It may take longer
with a slow HDD, but the end result is of no lesser quality. The
format of the file isn't imporant to the HDD of course, except for the
data rate.

However, then there is digital capture, as-in, capture. If it's not
compressed, it has same HDD requirements as analog that's been ADC'd.
Yep, that's all Stacey was clarifying.

Gareth
Click to expand...

Except that "digital video" is not by definition, compressed video. A
".dv" file is digital video, but not all digital video, is a
compressed ".DV" format.


Dave
 
I want to get into video editing but I want a system that would cut
those times in half. What would I need to do that? A 4-way Opteron or
4-way Xeon with SCSI drives?
Click to expand...

It depends on the quality of video that you want to encode.. If you
want good results then an hours wait really isnt that much at all.
 
Hi, so we took 2 systems similar specs both P4 3GHz and used MGI video wave
5 on them both to create a 1.5 hour DVD. This software is dual thread so
will take advantage of the Hyperthreading.
System 1:
P4 3GHz 533MHz bus, Gigabyte GA8-INXP, 2Gb PC1066 RD Ram, ATI all in wonder
Radeon 9700.
System 2:
P4 3GHz 800MHz bus, Gigabyte GA8-KNXP, 2Gb DDR 400, ATI all in wonder Radeon
9700.

Both have Western Digital 1200JB 120Gb 8mb cache hard drives fitted.
We took a 4.02Gb mpeg2 recording with a duration of 1.46 hours and turned
this into a DVD image.
System 1 took 56 minutes 23 seconds,
System 2 took 54 minutes 38 seconds.

We had thought about building an AMD XP3000+ machine as well but as they do
not have Hyperthreading the test would be pointless and no real result would
be achieved. We have tested a P4 2.8GHz 533MHz bus that does not have
Hyperthreading and it was 6 minutes slower than system 2 and we know the AMD
does not match this because none of the software is optimised for it.
Click to expand...

What astounds me is the lack of advantage for both 800MHzFSB and HT.

I assume the 2.8 also used Rdram?

Do I read correctly that 3.02 GHz HT 533FSB = 56:23
2.8 GHz nonHT 533FSB = approx 60:38

Thats only 7% better? And on a clock basis HT is actually even
slower!?

Would a dual cpu system get something 'real' out of the dual thread
of the 'MGI video wave' ? Obviously HT doesn't.


ancra
 
Kim said:
I want to get into video editing but I want a system that would cut
those times in half. What would I need to do that? A 4-way Opteron or
4-way Xeon with SCSI drives?
Click to expand...

A Mac G5
 
Back
Top