|
|
| i4004 |
| Posted: Aug 8 2004, 03:50 PM |
 |
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
F_U_C_K!
F_U_C_K!
F_U_C_K!
i used avisource here
http://neuron2.net/ipw-web/bulletin/bb/vie...=normalize#3151
, but this
| CODE | | #clip=SegmentedAviSource("D:\Video\rawstuff\Brain Storm_000.avi") |
leaves the vdub in deadlock
while this
| CODE | | clip=AviSource("D:\Video\rawstuff\Brain Storm_.avi") |
works (although, not instantly; if i open it, at first #normalize(show=true) doesn't show anything...then, on "save wav", after few seconds, thing starts to work, and after it i get the numbers on normalization (from show=true) ie it works ok after it)
i better of called the thread "segmentedavisource() strikes again".....
(no, i won't keep the brainstorm, it's a crappy movie...  )
--------------------
my signature:
 |
|
| Wilbert |
| Posted: Aug 8 2004, 05:12 PM |
|
|
Advanced Member
Group: Members
Posts: 132
Member No.: 6270
Joined: 11-September 03
|
1) regarding the linked thread: MonotoStereo works fine for me. Could you recheck it?
2) I will tell Sh0dan, but I would like to know which codec you used?
|
|
| i4004 |
| Posted: Aug 9 2004, 06:10 AM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
| CODE | #clip=SegmentedAviSource("D:\Video\rawstuff\hendrix__000.avi")
clip=AviSource("D:\Video\rawstuff\hendrix__000.00.avi")
#Normalizes each channel seperately.
#left_ch=GetChannel(clip,1).Normalize()
#right_ch=GetChannel(clip,2).Normalize()
left_ch=GetleftChannel(clip).Normalize()
right_ch=GetrightChannel(clip).Normalize()
#audio=mergechannels(left_ch,right_ch)
audio=monotostereo(left_ch,right_ch)
audiodub(clip, audio) |
i've toyed with this......nothing
now it seems to me that even avisource is not a solution either, as it doesn't really make a difference...
(perhaps it starts to work only after numerous attempts ie not a reliable solution.........
{edit} i have now come to the conclusion that normalize actually works, *but only for very short clips* ie, if i take , say, 300-400 frames from my source, and make avi from that (directstreamcopy), then normalize output is OK (that's why i thought avisource was OK: 300-400 frames clip is not segmented anyway... ) ; wilbert, can you check if there's any difference between 2.54 and most recent alpha? may it be some buffering/memory-usage problem? {edit end} )
"normalize:result not yet calculated" and vdub deadlock
is all i get most of the time...
the more i toy with it, the more i think this doesn't work at all....as is said at the start of neuron2 thread....
i just tried the approach of separated wav;
flopped again!
(i tried both monotostereo and mergechnnels approach here too....)
i use old version of picvideo mjpeg..(only what does that have to do with audio?)
in the end i used
| CODE | | SegmentedAviSource("D:\Video\rawstuff\hendrix__000.avi").amplifydb(5.0,0) |
and i still think normalize should have simple syntax, just like amplifydb;
| CODE | | ("D:\Video\rawstuff\hendrix__000.avi").normalize(0.95,0.95) |
or so...
--------------------
my signature:
|
|
| Wilbert |
| Posted: Aug 9 2004, 02:00 PM |
|
|
Advanced Member
Group: Members
Posts: 132
Member No.: 6270
Joined: 11-September 03
|
| QUOTE | | wilbert, can you check if there's any difference between 2.54 and most recent alpha? may it be some buffering/memory-usage problem? {edit end} |
There are indeed some differences regarding audio chaching, but I don't know whether that's the problem. Btw, I'm testing with latest binaries.
| QUOTE | "normalize:result not yet calculated" and vdub deadlock
is all i get most of the time... |
You will get this (when show=true) for a few seconds (at least on my Athlon 3000+ it takes a few seconds) when you start to play the clip. From docs:
| QUOTE | | The calculation of the peak value is done the first time the audio is requested, so there will be some seconds until AviSynth continues. |
I tried with a 6000 frames clip, and it workes fine. How long do your deadlocks take?
I still have to check SegmentedAviSource though ...
|
|
| i4004 |
| Posted: Aug 9 2004, 02:21 PM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
deadlock occurs fairly soon;i think it's instant, only it takes vdub some time to confirm it (or something)
the wav ripping (or avi encoding) never really starts....
and after perhaps 10-20sec deadlock
(but i usually don't wait for deadlock....when i see nothing happening i kill it)
i should check latest binaries as source is still on my hdd, but in 20' it won't be anymore...
--------------------
my signature:
|
|
| i4004 |
| Posted: Aug 10 2004, 04:11 AM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
i tried AviSynth_290704-2.exe
still no luck
on top of it, i tried 8000 frames clip, and it didn't work....
--------------------
my signature:
|
|
| i4004 |
| Posted: Aug 25 2004, 12:09 PM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
more about this avs bug;
-avisource or segmentedavisource doesn't matter; it acts the same
-it took it 50sec (!) to start exporting (via vdub's "save wav") a 10k frames clip: as said avisource or segm.avisource didn't matter; 50 sec in both cases
-it seems as avs is scanning the complete file prior to applying normalization (judging by the hdd lamp)
-this just doesn't work for longer files...(ie for a complete movie)
-255rc2 or 254 version doesn't matter; it acts the same
-tests done on cel1.3ghz+320mb ram machine....
conclusion; avs normalize is working only on shorter .avi clips
and i had to use amplifydb yet again...
it seems as if main problem is the fact that it tries to scan complete file, instead of doing it chunk by chunk....
this loses time and in the end deadlocks the vdub...
untill this is solved, this should be clearly stated in the docs....
--------------------
my signature:
|
|
| phaeron |
| Posted: Aug 25 2004, 06:50 PM |
|
|
Virtualdub Developer
Group: Administrator
Posts: 7773
Member No.: 61
Joined: 30-July 02
|
A delay of 30+ seconds from Avisynth won't, or shouldn't, deadlock VirtualDub. You will get warnings that a deadlock may have occurred, but the interlocks on the pipeline do not have timeouts and the pipeline should restart once Avisynth starts to feed data. You won't be able to stop it while Avisynth is scanning, however.
The next version will report where the threads are currently working when deadlock warnings appear, so you will have an idea of what's going on. (Yes, yes, I know... I'm trying to get the damn thing out the door.)
The need for an entire second pass is one of the reasons I'm not in favor of full normalization; windowed AGC is much faster and would probably work as well.
|
|
