how long should dvd-rb take?

[nick_harambee]

hi,

i have been using dvd-rb 0.70 with CCE 2.5 for a couple of weeks now, and the average time for processing/compressing a dvd with dvd-rb is typically 5.5-6.5 hours. a couple of other posts i have read suggest that i should be able to encode more quickly than this.

i am using a dell laptop with 3 GHz processor and 1 GB of RAM, with plenty of space on the harddrive and no other programs running.

i would be grateful if someone could let me know if these processing times are typical or whether there is something that i can do to speed the process up,

thanks

nick.

[DvPs]

hi

do you use avs filters? they can make the encoding process very slow.
how many passes with cce do you make?

and what is your cce speed?


mfg DvP

[nick_harambee]

not sure what you mean by avs filters, but i have the following options checked in dvd-rb: convert to YUY2, audiodub (blankclip(()), deinterlace with DECOMB, MPEG2Source ("source",idct=7).

i am making 2 passes

not sure what you mean by CCE speed. I am using CCE 2.50 basic.

thanks

nick.

[DvPs]

you can add filters in the avs setup by rebuilder.
i think you can find it under options-->avs-->filters

then a text window should open where you can add filters.


i have also a question: MPEG2Source ("source",idct=7).
what does this comment?

the speed by cce: by the encoding window by cce (where you can see the total frames, the encoded frames, the elapsed and the time left) you can see the speed.

mfg DvP

[jptheripper]

why are you de-interlacing ? that will slow you down, and is usally not needed

oh and the main slow down is read and write to a laptop drive. that is your bottleneck

[MaxT]

OPV with CCE SP 2.67.0.23, convert to YUY2, MPEG2Source ("source",idct=7) = Max 1.5h @ 2h DVD

OS: WinXP Professional 5.1 Service Pack 2 (Build #2600)
CPU: Intel Pentium IV, 3.01 GHz (800FSB)
MEMORY: 1024MB DDRAM 3200 (800FSB)

[jdobbs]

@nick_harambee

That sounds slow for that processor -- but then I believe some notebooks have automatic speed reduction when the processor heats up... and CCE is very processor intensive. I have an Athlon XP 3200+ and it usually takes about 1 hour per pass (and then time for PREPARE/REBUILD).

How many passes are you doing?

[DvPs]

hi

what is the funktion of ("source",idct=7?


mfg DvP

[wmansir]

It uses a more accurate decompression routine. It may not have much or any visual impact, but mathmatically the image is closer to the DVD decoding specifications. It doesn't add much in processing time, so if you are a perfectionist you might as well turn it on.

[nathanaa]

I actually came here to look for info on this. It is now about 12:30a CST in the US. I started DVD-RB on Yes, Minister disc 1 at about 11a (over 13 hours ago!), and it is still running.

I have it in QuEnc mode, everything default from the DVD-RB Installer, but I think I chose "High Quality". I believe it is doing 2 passes.

My set up is a Fujitsu P5020D laptop, 900mHz processor, 1GB RAM. This obviously seems to be taking WAY too damn long.

Any ideas on what is taking so long or what is up with this, anyone?

[DvPs]

Originally posted by nathanaa

My set up is a Fujitsu P5020D laptop, 900mHz processor, 1GB RAM. This obviously seems to be taking WAY too damn long.


i think this is the reason why it takes so long...
900Mhz isn't really fast any more for video encoding...


mfg DvP

[artoor]

I've been using DVD-RB 0.72, CCE 2.70.01.05 according to Max's (http://www.maxt.dk/dvd9.to.dvd5/) tutorial (with a small modification according to this (http://forum.doom9.org/showthread.php?s=&threadid=89537) thread - very helpful information if You are going to use CCE 2.70.01.05 without eclcce 1.81 :) )

My config: Windows XP Prof + SP2, Pentium 4 2.4 HT (FSB800), 768 MB DDRAM (FSB800)

10 passes = at about 10 hours

--== EDIT ==--

What a pity... it was some distortion, because currently 10 passes = 14 h...ehhhh. Sorry for wrong information :(

[nick_harambee]

why are you de-interlacing ? that will slow you down, and is usally not needed

oh and the main slow down is read and write to a laptop drive. that is your bottleneck

ok, so i will not use de-interlacing if this is not necessary

i have been reading from and writing to an external hard-drive (USB 2.0), but processing on the laptop. would it be quicker if i was reading from and writing to the internal hard drive on the laptop?, as i probably could create enough space on this drive?

thanks

nick

[nathanaa]

How does one choose to either deinterlace or not? And when is it not necessary? Perhaps this is part of my problems with QuEnc; QuEnc has not worked even once for me, so I've always used ReJig, but this time ReJig isn't processing properly either.

Oh, and the speed of my processor has never seemed to be a problem in the past, at least not with ReJig or DVDShrink/DVDDecrypter, etc. I believe I have the priority set for "High Priority" in both DVD-RB and QuEnc, if those settings are available (I know it is for QuEnc).

I also know I didn't bump and keys or anything while it was processing, in case anyone thinks that might be the source of my trouble, never touched the keyboard.

I know a 16 hour processing is way too much, and then I end up with the Error 0006, but I'm still not sure why it took so long before it got to the error and crashed. So, any help greatly appreciated, thanks.

[dragongodz]

900mHz processor,
you cant seriously expect fast encoding with such a cpu.

the speed of my processor has never seemed to be a problem in the past, at least not with ReJig or DVDShrink
they are transcoders not encoders. there is a big difference.

I think I chose "High Quality"
better quality settings mean longer encoding of course.

I've always used ReJig, but this time ReJig isn't processing properly either.
sounds like the problem is not QuEnc then doesnt it.

[MaxT]

"10 passes = at about 10 hours"

Yeh, seems 1 pass takes about 1 hour on any decent new box.

[nathanaa]

Well, seeing as ReJig didn't work for me due to an entirely different reason (to my knowledge), that is why I was trying QuEnc. AND I never said that QuEnc was the source of this extreme-time problem, but that I was here *asking* if it might be - yeah, it still could be that QuEnc is the problem. It might not be, I know, but just saying so doesn't really help figure out if it is or not.

And as regards my cpu - I agree it isn't the Ferrari of processors, but it is the equivalent of a !.2-1.4GHZ processor, so it isn't a Model T either, and should be able to run this in something shy of 16 hours. Unless someone is actually going to tell me that 12-16 hours is what I should expect it to take - if that is so, then please, somebody say so. I was under the impression it should take no more than 5-7 hours to process, but if my cpu speed actually will make that big of a difference, then it would be great to know.

Any more input is appreciated folks, thanks.

[jdobbs]

How does one choose to either deinterlace or not? That's an easy one.

Not.

Let DVD-RB do it's job -- it will output a DVD that matches the source.

[dragongodz]

AND I never said that QuEnc was the source of this extreme-time problem, but that I was here *asking* if it might be
and you compared the speed to transcoders to which i already answered quite clearly with

they are transcoders not encoders. there is a big difference.
so your assumption the long time is a "problem" is based on incorrect information.

is it normal ? well you see people with much faster cpus than yours encoding a lot faster but you will notice it still takes hours for multipass encoding. also speed comparisons to CCE are not relevant since that is also faster than QuEnc, especially if you turn on high quality options in QuEnc(try it with that turned off and see how much extra speed you get). it doesnt get much simpler than that.

And as regards my cpu - I agree it isn't the Ferrari of processors, but it is the equivalent of a !.2-1.4GHZ processor
uh what ? hmm you have a 900mhz cpu but you say its the equivalent to a 1.2-1.4ghz cpu. ah no, its a 900mhz cpu so its equivalent to a 900mhz cpu. if you are trying to compare it to another brand or type of cpu say so and which. either way compare it to the much faster cpus others are using and as i already said multipass still takes hours.

[jptheripper]

just for reference

on my p3 1ghz, 1.5gb ram, reading to and from different 8mb cache 7200 rpm drives, it takes me about 5 hours per pass.

so 5+1vaf pass = ~30 hours

and its still worth it

[nathanaa]

********
just for reference

on my p3 1ghz, 1.5gb ram, reading to and from different 8mb cache 7200 rpm drives, it takes me about 5 hours per pass.

so 5+1vaf pass = ~30 hours

and its still worth it
**********

See, now this is what I needed to hear - so maybe my situation ISN'T so unusual then. I didn't realize that it might take so long.

What is "+1vaf" though? Are you using QuEnc too, and if so, with the High Quality option enabled? A question just occured to me - if I did NOT use the HQ option, would the result be equivalent or better than what I would get using the ReJig transcoder?

Regading my cpu (yet again), when I got the laptop, the type of processor it is is apparently equivalent to a 1.2-1.4 Pentium 3 (when I was researching the laptop, this kept getting tossed about as a big selling point by the technogeeks I was listening too...well, reading). That's what I was talking about.

[dragongodz]

What is "+1vaf" though? Are you using QuEnc too,
no by saying a vaf pass tells us he is using CCE, which is faster than QuEnc to start with.

if I did NOT use the HQ option, would the result be equivalent or better than what I would get using the ReJig transcoder?
better in what way ? quality ? you need to understand encoders and transcoders work differently. transcoders are generally very good for when only asmall amount of reduction is needed(say down to over 80% of original size). once you start getting to much larger amounts of compression encoders usually(there is always the exception to the rule so i would never say 100% of the time) do a much better job. so even without HQ QuEnc should give better results for large size reduction.

it is is apparently equivalent to a 1.2-1.4 Pentium 3 (when I was researching the laptop, this kept getting tossed about as a big selling point by the technogeeks I was listening too...well, reading). That's what I was talking about.
see now that makes more sense. if you look though you will see a lot of people using Pentium 4's and Athlon XP's, these are also faster than a Pentium 3 aswell.

[jdobbs]

Originally posted by jptheripper
just for reference

on my p3 1ghz, 1.5gb ram, reading to and from different 8mb cache 7200 rpm drives, it takes me about 5 hours per pass.

so 5+1vaf pass = ~30 hours

and its still worth it 30 freaking hours? My processor shouldn't be more than about 3 times as fast as yours (AMD 3200+) with a similar hard drive and less memory... it only takes 1 hour per pass.

I usually do 2 movies between the time I go to bed and I get up the next morning. I think the price of a processor upgrade might be worth that much time.

[nathanaa]

Originally posted by dragongodz
better in what way ? quality ? you need to understand encoders and transcoders work differently. transcoders are generally very good for when only asmall amount of reduction is needed(say down to over 80% of original size). once you start getting to much larger amounts of compression encoders usually(there is always the exception to the rule so i would never say 100% of the time) do a much better job. so even without HQ QuEnc should give better results for large size reduction.

Yes, I meant picture quality. When I've used ReJig I have almost never seen any kind of artifacts on the screen (pixelation, motion blurring, etc.), although sometimes I have been concerned at the huge shrinking the disc went through. That's the main reason I'm so interested in getting QuEnc to work for me, even with the major time it will take (I figure I'll just run it overnight, like most probably do).

[Morte66]

FWIW, it takes me about 6 hours to process a movie with DVD-RB/QuEnc using basic 2 pass settings on my Athlon XP 1900+ desktop. High quality mode adds 4 or 5, and trellis mode adds 4 or 5 more, neither making much difference.

I would suggest that you also take a look at an encoder called FreeEnc (search these forums), which is a QuEnc compatible (you rename FreeEnc.exe to QuEnc.exe). Use the Incredible_KDVD template (rename it to default.ini). In my experience it's a touch faster than QuEnc for the same quality. Also it's much happier about running in parallel with other applications and not hogging the processor, which makes it more pleasant to use on a slower machine.

[jptheripper]

yeah i know the 30 hours is obscene, I get a cce speed of around .8-.9 if im lucky.

I know its time for a processor upgrade but i just blew a car speaker so that has to be shelved temporarily.

oh well.