On Thu, 22 Jan 2004 [EMAIL PROTECTED] wrote:
> > What about yuvmedianfilter? Is it a spatial lowpass filter?
>
> It's a nonlinear spatial filter that generally gives an overall
> lowpass effect. Ideally it removes noise without attenuating high
> signal frequencies (thus preserving edges), but
On Thu, 22 Jan 2004, George Kola wrote:
> find that the generated MPEG-1 and MPEG-2 have perfect A/V sync (Thanks to
> people here for helping me with it). I find that the MPEG-4 (two pass) has
> A/V sync issue. The audio slowly drifts behind the video. In the beginning,
The alternate
> On Wed, 21 Jan 2004, George Kola wrote:
> > Command? create quicktime playable MP4). I find that MPEG-4 has A/V...
> > Command? towards the end (one stream is behind the other), but MPEG...
> > Command? fine. Any suggestions ?.
>
> Something doesn't scan right in that paragraph - seems as
On Thu, 22 Jan 2004, Nicolas Boos wrote:
> > That's what yuvdenoise is for :)
>
> For "big" noise only... :-)
As in "VHS" ;)
> Yes, but a _clean_ source is rare... If source is noisy (even a
DV (from miniDV or Digital8) is very clean. DVDs are clean, at least
for
> > But it would be nice to have a spatial lowpass filter in the
> > toolbox also to reduce noise that way.
>
> What about yuvmedianfilter? Is it a spatial lowpass filter?
It's a nonlinear spatial filter that generally gives an overall
lowpass effect. Ideally it removes noise without attenuati
Le Thu, 22 Jan 2004 13:06:13 -0800 (PST)
"Steven M. Schultz" <[EMAIL PROTECTED]> écrivait :
> > The sinc* kernels are very bad at removing noise. IMHO, the
> > cubic* kernels are better for this job...
>
> That's what yuvdenoise is for :)
For "big" noise only... :-)
> For clean sour
On Thu, Jan 22, 2004 at 05:30:55PM -0500, [EMAIL PROTECTED] wrote:
>
> But it would be nice to have a spatial lowpass filter in the
> toolbox also to reduce noise that way.
What about yuvmedianfilter? Is it a spatial lowpass filter?
---
The
> > The sinc* kernels are very bad at removing noise. IMHO, the
> > cubic* kernels are better for this job...
>
> That's what yuvdenoise is for :)
Well, sort of. It removes noise temporally. But it would be nice to
have a spatial lowpass filter in the toolbox also to reduce noise that
wa
On Thu, 22 Jan 2004, Richard Ellis wrote:
> Is there any way to concatenate plural sets of mp2/m2v files for
> multiplexing together into a single stream.
Not in a 100% reliable way. If you simply cat the files together
you'll find that some software players ignore the disconti
On Thu, 22 Jan 2004, Nicolas Boos wrote:
> The sinc* kernels are very bad at removing noise. IMHO, the
> cubic* kernels are better for this job...
That's what yuvdenoise is for :)
For clean sources the sinc kernels are better at preserving
detail (it's an open question h
Le Thu, 22 Jan 2004 10:21:41 -0800 (PST)
"Steven M. Schultz" <[EMAIL PROTECTED]> écrivait :
Hello Steven,
[...]
> "If I use -q 2 or -q 3, I get pulsating (every second or so) block
> artifacts in the background, which I get not with -q 4."
>
> I wonder if those artifacts are also due to t
Is there any way to concatenate plural sets of mp2/m2v files for
multiplexing together into a single stream.
The m2v files are created in mpeg format (-f 8) to mpeg2enc.
I tried:
cat f1.mp2 f2.mp2 f3.mp2 > all.mp2
cat f1.m2v f2.m2v f3.m2v > all.m2v
mplex -f 8 -r 0 -V -b 230 -S 0 -o all.%0d.mpeg
On Thu, 22 Jan 2004, [ISO-8859-1] Thomas Börkel wrote:
> > "If I use -q 2 or -q 3, I get pulsating (every second or so) block
> > artifacts in the background, which I get not with -q 4."
> I tried without scaling and still have the artifacts. Maybe it's the
> source material.
That mig
HI!
Steven M. Schultz wrote:
I can see light vertical stripes in bright scenes in my encoded result.
I seem to recall transcode's -Z scaling causing problems for someone
else in the past.
A (much) better way to perform scaling is to use 'y4mscaler'. You can
find y4mscaler
On Thu, 22 Jan 2004, [ISO-8859-1] Thomas Börkel wrote:
> Bernhard Praschinger wrote:
> >
> >>I can see light vertical stripes in bright scenes in my encoded result.
> >
> > How large is the stripe ?
> > Do you scale (guessed from the -Z option) the picture ?
> > If yes does it also happen when
Hallo
> > Can you encode you fils on the command line without using the lav2mpeg
> > script.
> > You command looks like that than:
> > lav2wav capture.avi | mp2enc -b 224 -r 44100 -o capture.mp2
> > lav2yuv capture.avi | yuvscaler -o VCD | mpeg2enc -f 1 -4 1 -2 1 -o
> > capture.m1v
> > mplex -f 1
HI!
Bernhard Praschinger wrote:
Hallo
I can see light vertical stripes in bright scenes in my encoded result.
I can't see them in the original AVI (Xvid).
How large ist the stripe ?
Do you scale (guessed from the -Z option) the picture ?
If yest does it also happen when you don't scale it ?
That
Hallo
> I can see light vertical stripes in bright scenes in my encoded result.
> I can't see them in the original AVI (Xvid).
How large ist the stripe ?
Do you scale (guessed from the -Z option) the picture ?
If yest does it also happen when you don't scale it ?
That it would be from scaling.
HI!
I have an example to download and would be glad to hear comments:
http://www.boerkel.de/q.zip
Forgot to mention: File size is 1.5 MB.
Thomas
---
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Developmen
HI!
I can see light vertical stripes in bright scenes in my encoded result.
I can't see them in the original AVI (Xvid).
Using transcode 0.6.12, mjpegtools .93 (RC 4), xvidcore 1.0b3 like this:
transcode -i "$1" -Z 720x352 -Y-64
-w 5000 --pulldown --export_fps 0,4 --export_asr 2 -V -M 4 -y
mpeg2
HI!
I have made some tests encoding an NTSC AVI (Xvid) to MPEG2 with
transcode/mpeg2enc.
If I use -q 2 or -q 3, I get pulsating (every second or so) block
artefacts in the background, which I get not with -q 4.
I have an example to download and would be glad to hear comments:
http://www.boerke
HI!
Steven M. Schultz wrote:
A lower -q (2 or 3) gets me higher file sizes, so shouldn't it also be
potentially better quality?
Yes, it will give better quality. That may or may not be visible :)
I have a problem with -q 2 and 3. See seperate thread.
Ah, it's letterboxed so you have
22 matches
Mail list logo
