08-04-2002, 10:05 PM
|
Free Member
|
|
Join Date: Aug 2002
Location: Montréal, QC
Posts: 19
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
I just finished encoding Star Trek II: The Wrath of Khan (1 h 52 min) with a modified KVCD 2x Plus-704*480 Templates.
What I changed was:
MPEG-2
480*480
CQ 65 (Maximum: 4000, Minimum:1800)
Stereo
Soften block noise 35-35
Quality is absolutely incredible ... and it fits perfectly on 2CDs... I will use it with all my new SVCD now!!!
Thanks for your great templates Kwag !!!
|
Someday, 12:01 PM
|
|
Site Staff / Ad Manager
|
|
Join Date: Dec 2002
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
|
|
|
08-04-2002, 10:45 PM
|
Free Member
|
|
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Thanks Nico...
But I suggest you keep the audio at Dual-Channel, if you plan on using Dolby Surround sound. It's much better that stereo, because of the independent channel separation. Also select Soften Blocks, but set it to 0-0. It works in reverse as it is advertized at least all the tests I did, up to the current TMPEG version 2.57. The higher the number, the more visible the blocks.
-kwag
|
08-05-2002, 10:38 AM
|
Free Member
|
|
Join Date: Aug 2002
Location: Montréal, QC
Posts: 19
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
In fact, the file comes from a divx (I downloaded all the Star Trek movies on Kazaa) so you know sound isn't so great from an MP3 file (by the way, do you know a way to improve sound from MP3 file??? Cause what I have sounds a little bit "metallic")...
For "soften block noise", I have to admit I never tried but I will on my next SVCD and will tell you what I think....
|
08-05-2002, 08:37 PM
|
Free Member
|
|
Join Date: Aug 2002
Location: Montréal, QC
Posts: 19
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
I'm sorry Mr Kwag but I just tried the soften block noise with 0 0 and lots of ugly blocks appear mostly in slow scene (I'm using the latest 2.57 version). So I think, I'll continue with 35-35 .
|
08-05-2002, 09:05 PM
|
Free Member
|
|
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by Nico...
I'm sorry Mr Kwag but I just tried the soften block noise with 0 0 and lots of ugly blocks appear mostly in slow scene (I'm using the latest 2.57 version). So I think, I'll continue with 35-35 .
|
Ok. I'll take a look. Maybe they fixed it
-kwag
|
08-06-2002, 02:19 AM
|
Free Member
|
|
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Hi Nico...,
Try a little test yourself. Encode a same clip at 35-35, and do the same at 100-100. Which has more blocks? 100-100.
Now select 0-0 and do the clip again.
The one with less blocks is 0-0.
Again I can confirm that the "soften blocks" function in TMPEG is reversed. This was previusly tested with various different movies. I just tested it again with a small clip from "Kate&Leopold"
This was tested on TMPGEnc 2.57 Plus.
Please chose the same scene for your test, and pause in different places. Then you'll see the sharp "edge" of the macro blocks more visible, the higher you increse the number on soften blocks.
-kwag
|
08-11-2002, 11:49 AM
|
Free Member
|
|
Join Date: Aug 2002
Location: Montréal, QC
Posts: 19
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Oups! I made the test... you were right Kwag... sorry
|
08-17-2002, 07:22 AM
|
Free Member
|
|
Join Date: Aug 2002
Posts: 27
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
-> Nico
>The Wrath of Khan (1 h 52 min) with a modified KVCD
> 2x Plus-704*480 Templates.
>What I changed was:
>MPEG-2
>480*480
>CQ 65 (Maximum: 4000, Minimum:1800)
if your S/A player does is well there´d be no need to set the lower
limit to 1800. Check it out :
try to set low to 300 and you´d get space left over that could be invested
in a higher CQ (maybe 70 or more). At 4000 max you´re out of
SVCD´s spec. Lotsa people reported bad results at this high speed.
>Stereo
my idea:
both stereo channels got more signals in common than signals are
different. So it doesn´t make sense to encode the same signal twice
and put it into 2 channels with just a little difference.
I allways use joint stereo cause there´s more space left over for better
encoding IF THE ENCODER DOES IT RIGHT !
So that´s the theory:
some people reported TMPGenc´s audio encoder isn´t that good and
joint stereo results in worse coding than dual channel. Maybe
you´d give 2lame a try as external encoder
Obrigado
|
08-17-2002, 09:38 AM
|
Free Member
|
|
Join Date: Aug 2002
Location: Montréal, QC
Posts: 19
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Heu... I tried a lot of tests...
1800 has been chosen because it's, for me, the settings where no macroblocks appears...
4000 has been chosen because when chosing an higher bitrate it does not give an improvement.
65 for the reason it gives me 60 minutes or more...
Maybe you should give it a try... it works extremely well for me...
You're true joint great is great (mostly when when encoding from divx, dual is better to preserve the dolby surround channels from DVDs) and I always use toolame for encoding and SSRC for resampling...
|
08-19-2002, 06:29 AM
|
Free Member
|
|
Join Date: Aug 2002
Posts: 27
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
-> nico
> 1800 has been chosen because it's, for
> me, the settings where no macroblocks
> appears.
shure about that?
usually you'd get block artefacts if the
bandwidth is to low, so it'd be a problem
of the upper limit.
Assume a nearly black movie scene
without any difference between the pics:
in this case you won't get a significant
better image
when raising the lower limit to 1800.
It's just wasting bandwidth, cause this is
the main principle of MPEG's optimization
of interframe encoding.
But it may be a problem of the S/A player
who's unable to decode a fast variation
of bps-rates. In this case you'd really need
some bitstuffing :-(
just my 2 cents
Obrigado
|
All times are GMT -5. The time now is 05:27 PM — vBulletin © Jelsoft Enterprises Ltd
|