digitalFAQ.com Forums [Archives] - KVCD: CQ of 47 (with max 2500, min 300)

digitalFAQ.com Forums [Archives] (http://www.digitalfaq.com/archives/)

- Video Encoding and Conversion (http://www.digitalfaq.com/archives/encode/)

- - KVCD: CQ of 47 (with max 2500, min 300) - Is this a bad idea? (http://www.digitalfaq.com/archives/encode/4526-kvcd-cq-47-a.html)

KVCD: CQ of 47 (with max 2500, min 300) - Is this a bad idea?

Hi guys, I was wondering if anyone could give me their opinion on the following:

I have a movie that is 4:3 (fullscreen), 89 minutes in length. I want to make it a KVCDx3a (544x576) and fit it on to a 90min CD-R, however, in order to do this, I am going to have to use a CQ of ~47 (with max 2500, min 300). Is this a bad idea? If so, can anyone give me their opinions on my options? Thanks, forgive my newbieness :)

Well somebody alse might just go for a lower resolution but with the MA Script i would try chainging either the max 2500 to abuot 2300kbs that should gain you some CQ and maybe even lowing the min. bitrate to 150 if your DVD can handle it. :)

Re: Opinions needed :/

Quote:

Originally Posted by Cng

Is this a bad idea? If so, can anyone give me their opinions on my options? Thanks, forgive my newbieness :)

Bad idea? Well, not necessarily (different sources will look better/worse than others at some CQ levels). But generally speaking, I (personally) find that anything lower than ~54CQ with the x3 resolution starts to look "unacceptable", to me (everyone has a different opinion though on what is or isn't "good" or "acceptable" qualitywise...also keep in mind that CQ levels aren't set in stone). Probably the best thing you can do is judge for yourself whether it is good enough by creating a small sample and viewing it on your tv (or whatever device you plan to watch it on). I think most people here would agree however that it probably would not result in a very good quality at the CQ you listed...probably low "acceptable" quality at best.

Showing us your script and giving some more details would probably help. What audio bitrate are you using? Are you using Blocks Overscan? You could also try what ovg64 suggested about lowering your min/max bitrates a little. There are also a couple other "tips" that might help listed in this thread (be forewarned though! I was pretty long-winded :lol: ...even moreso than usual!): http://www.kvcd.net/forum/viewtopic.php?t=4984 Good luck!

-d&c

Hi Cng.
I don't make PAL movies, than my resolution is 544x480. I don't know how much it affects on the final CQ, but I just finished a 85min (not 89), with 4:3 fullscreen, 112Kb audio, on a 80minCD (not 90) and my final CQ was 62! I really don't know if our differences (NTSC not PAL) and few minutes less, could give me so much CQ leverage. Unfortunatelly I don't have my script anymore, I remember to use letterbox(24,24,24,24) to give more compression. Are you using the latest optimal script to AVS 2.5?
Let us see your script :wink:

LaBomba

Hi guys, sorry for the late reply. First, thank you for giving me your opinions/suggestions/comments.

@ovg64:
Thanks for the suggestion(s), and I followed your advice, but had to lower my average bitrate to 2000, and minimum bitrate to 150 (thankfully my DVD player is able to play it). I kept the resolution at 544x576, with a CQ of 54. The quality was... so so :) I'll tell you the reason for such low values in a second :)

@dazed&confused :
You wouldn't believe how many comments that I have read of yours on this forum, they are extremely helpful. The day I posted this topic, I was in sort of a hurry, so I didn't get time to give details on my script, and other settings, such as audio - sorry about that. I'll do a better 're-post' below, stating my full problem.

@labomba:
Well, technically I should have been able to get the same, or approximately the same CQ as you. But, I'll stop babbling on, and get on with a more detailed post :)

-----------------------------------------------

First of all, my AVISynth 2.52 script is as follows:

Code:

## DLL Section ## 

# 

LoadPlugin("D:\AVISynth\MPEG2Dec.dll")

LoadPlugin("D:\AVISynth\STMedianFilter.dll") 

LoadPlugin("D:\AVISynth\Asharp.dll") 

LoadPlugin("D:\AVISynth\Plugins\Unfilter.dll") 

LoadPlugin("D:\AVISynth\Undot.dll") 

LoadPlugin("D:\AVISynth\Decomb.dll")

LoadPlugin("D:\AVISynth\Convolution3DYV12.dll") 

LoadPlugin("D:\AVISynth\Sampler-2.5.dll")

LoadPlugin("D:\AVISynth\TemporalCleaner.dll")

# 

#### 



## Defined Variables and Constants ## 

# 

MaxTreshold = 1.50 

nf =  0 # Current frame. 

# 

#### 



## Main section and static filters ### 

# 

AVISource("D:\Capture.avi")

AssumeFPS(24.991)

FieldDeinterlace()

TemporalCleaner()

ConvertToYV12()

Convolution3D(preset="VHSBQ")

# 

undot() 

Limiter() 

asharp(1, 4) 

STMedianFilter(8, 32, 0, 0 ) 

MergeChroma(blur(MaxTreshold)) 

MergeLuma(blur(0.1))

BicubicResize(544,576) 

# 

# 



## Linear Motion Adaptive Filtering ## 

# 

# ( Portions from AviSynth's manual ) - This will apply temporalsoften to 

# very static scenes, and apply variable blur on moving scenes. 

# We also assign a variable - and this is why a line break is inserted: 



SwitchThreshold = (Width<=544) ? 4 : (Width<=576) ? 3 : 2 

ScriptClip("nf = YDifferenceToNext()"+chr(13)+ "nf >= SwitchThreshold ? \ 

unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ) : \ 

TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 1) ") 



# 

# 

# 





LetterBox( 16, 16 ) # Depends on situation. Use MovieStacker! 

Limiter() 



# 

# 

## Functions ### 



function fmin( int f1, int f2) { 

  return ( f1<f2 ) ? f1 : f2 

} 



# 

####

#Sampler(length=24)

## MPEG size = ((Total frames/MovieTimeInMinutes)/24) * MPEG sample file size * .98 ##

As you can see, since I'm a major newbie at KVCD conversions (and AVISynth), I used the current optimal script recommended by Kwag, and modified it a bit. I'm not even sure if I modified it right and/or chose to use the right script :)

Anyway, the reason why I did not use GripCrop/GripBorders, is because my source is actually captured from the television (30gig file - PICVideo MJPEG codec). It's a nice clean capture @ 544x576, with some noise, that's why I chose to use the above script, and add in: Convolution3D("VHSBQ") and TemporalCleaner(). The reason for removing GripCrop/Borders, was because when I tried to encode it with TMPG, it added major borders, and I mean major :) on each side of the video - this decreased the initial width by quite a bit, and replaced it with black borders. I found out the reason why it was happening - the line AssumeFPS(24.991) caused it to happen, for some reason GripCrop/Borders add huge bars on the side when I change the source's FPS to this. My reason for changing the FPS, is due to re-syncing the audio & video properly, then I encode it back to 25FPS with TMPG (then it syncs fine). Therefore, I used the BicubicResize function instead, with letterboxing enabled.

Before my first attempt at encoding, I did a sample (with Sampler.dll), and it came out to be 19.1MB @ 544x576, MPEG-1, Max: 2500, Min: 300, CQ: 60. My target for an 80min CD-R would have to be about ~12MB per sample. But unfortunately, my captured AVI does not seem to compress very well (even after lowering bitrates and CQ), or am I doing something terribly wrong here? :) Before I forget, my audio was encoded in HeadAC3e @ 112kbps stereo.

Any further suggestions on the script I'm using, and your suggestions on what you would do in my current situation, etc would be helpful :) Because I'm planning on staying with KVCD, it's definitely the greatest thing since sliced bread :P Thanks again guys.

Cng

PS - Movie details: 132934 frames, ~89 minutes, 4:3 - PAL. I want to fit this onto 1 CD-R, preferably 80 minutes, but a 90 minute CD-R is fine.

If you're capturing from VHS, there's probably no reason to encode to a resolution higher than 352x480 (due to resolution limitations of vhs). This would give you a raise in CQ.

Since you're working on a fullscreen capture, when you tried using Gripfit, did you use the source_anamorphic=false parameter? This is necessary for fullscreen, non-anamorphic sources, just in case you didn't konw.

Code:

GripCrop(width=544, height=576, overscan=2, source_anamorphic=false)

Instead of using TemporalCleaner and C3D, you might want to try using FaeryDust(temporal filter) or Pixiedust(temporal+spatial, stronger). You can see ak47's post here for an idea of how/where he used it: http://www.kvcd.net/forum/viewtopic.php?t=5067 . Dust is very slow though, so keep that in mind...it's also a 2.0x filter, so colorspace conversion is required.

Good luck.
-d&c