View Full Version : AviSynth 2.5.8 [Dec 30th] - Release! (was RC-5 [Dec 27th])
Fizick
4th December 2008, 17:49
Sorry, a missed (wrongly interpret) "Standalone" name.
By the way, may be rename it in installer? (I translated it as "partially portable" in Russian installer).
Or may be add some explaining comment to installer.
IanB
4th December 2008, 22:28
So the registry entries are always needed.Yes if you want to use the standard Windows AviFile interface to access your script.
Various people have been trying package a semi useful Avisynth for use on crippled SOE machines where they cannot do a standard Admin install. :search: Use search to find the various discussions about this and programs that can directly use Avisynth without any registry access.
Zach
14th December 2008, 07:36
The Standalone installation is not intended to be functional out of the box.
Could you expound upon that statement?
Case in point, I'm a fresh install of WinXP right now and was about to install AviSynth. Are you saying that if I go to SourceForge.net (http://sourceforge.net/project/showfiles.php?group_id=57023&package_id=105994&release_id=625750) and download and install AviSynth_080912.exe, it won't work?
Would installing 2.5.7 first and then installing this latest released beta over it be the correct install procedure then?
Leak
14th December 2008, 11:50
Are you saying that if I go to SourceForge.net (http://sourceforge.net/project/showfiles.php?group_id=57023&package_id=105994&release_id=625750) and download and install AviSynth_080912.exe, it won't work?
No, he isn't.
You can now choose between "Standard Installation" (or whatever the default is called) and "Standalone Installation" in the installer, and the latter just unpacks everything you choose to the destination folder without trying to register anything, in case you're on a locked-down user account that's not allowed to install new software.
Zach
14th December 2008, 18:20
No, he isn't.
You can now choose between "Standard Installation" (or whatever the default is called) and "Standalone Installation" in the installer, and the latter just unpacks everything you choose to the destination folder without trying to register anything, in case you're on a locked-down user account that's not allowed to install new software.
Ah, okay. It's a choice specified during the install process. I get it now. Thanks for clearing that up for me! :)
Nightshiver
15th December 2008, 02:23
Are there plans to make 2.5.8 "official" and get it out of RC?
IanB
15th December 2008, 07:28
Yes, but I have 2 very minor fixes (YV12 ShowFiveVersions, modified rounding in DSS ConvertFPS=True) and the Czech documentation pending. I just have not had any time to work on it, life is getting in the way of doing it. I hope to do a big push over the Xmas break.
If you don't need the Czech documentation, you might as well be using it. RC4 to RC4a was just some Russian Doc updates. And at this point the Final will have the Czech doc plus a few other general doc updates, with the 2 code fixes.
Gavino
15th December 2008, 12:14
Any plans to address the issue of BlankClip and implicit last (actually a general parser issue), reported here?
IanB
15th December 2008, 13:54
Yes, but not in this release.
asarian
20th December 2008, 11:13
Recently, the matter of AviSynth 2.5.8's downwards compatibility with 2.5.7 came up in this thread:
http://forum.doom9.org/showthread.php?p=1226105#post1226105
Instead of just speculating, or go by hearsay, I figured I might as well just ask the author directly. :)
Thanks
IanB
21st December 2008, 00:01
@asarian,
I have culled the changes list to things maybe related to your thread. Item 2 could be of most interest as early 2.5.8 releases allocated significantly more cache and this did cause problems, hence returning to 2.5.7 values for the final versions.
Most of the other changes should actually improve problem issues over 2.5.7.
There could of course be latent bugs that people are not reporting here.
Changes:
* Initial Memory Max value clamped to 512MB.
* Default Memory Max value restored to quarter of Free memory. Minimum 16Mb. As per 2.5.7.
* Added critical section to CAVIFileSynth class.
* CoUninitialize() done immediately for S_FALSE CoInitialize() returns, i.e we do not hold a COM use count.
* DirectShowSource() now attempts to use partial graph from failing RenderFile call.
* Relax YV12 resizer width restriction, now mod 2 was mod 4.
* SetCacheHints(CACHE_RANGE, n) will now surrender frames to satisfy SetMemoryMax().
* CoInitialize()/CoUninitialize() now done as part of ScriptEnvironment creation/deletion.
asarian
21st December 2008, 11:49
@asarian,
I have culled the changes list to things maybe related to your thread. Item 2 could be of most interest as early 2.5.8 releases allocated significantly more cache and this did cause problems, hence returning to 2.5.7 values for the final versions.
Most of the other changes should actually improve problem issues over 2.5.7.
There could of course be latent bugs that people are not reporting here.
Changes:
* Initial Memory Max value clamped to 512MB.
* Default Memory Max value restored to quarter of Free memory. Minimum 16Mb. As per 2.5.7.
* Added critical section to CAVIFileSynth class.
* CoUninitialize() done immediately for S_FALSE CoInitialize() returns, i.e we do not hold a COM use count.
* DirectShowSource() now attempts to use partial graph from failing RenderFile call.
* Relax YV12 resizer width restriction, now mod 2 was mod 4.
* SetCacheHints(CACHE_RANGE, n) will now surrender frames to satisfy SetMemoryMax().
* CoInitialize()/CoUninitialize() now done as part of ScriptEnvironment creation/deletion.
Thank you very much for your clear and helpful answer. :)
IanB
27th December 2008, 03:36
Changelist with respect to 2.5.8 RC4 [080831] : See 1st post for full change list.
Additions:
* Added Czech doc translation.
Bugfixes:
* Fixed ShowFiveVersions() YV12 chroma position in bottom half. Regression.
Optimizations:
* DirectShowSource() convertfps=true add 0.5ms rounding to expected sample time. Improves performance with millisecond based DS spliters.
Changes:
* Updated Russian documentation (Fizick).
As usual download from Sourceforge (http://sourceforge.net/project/showfiles.php?group_id=57023&package_id=105994&release_id=649730).
Enjoy! :cool:
I will promote this into the official 2.5.8 release in a few days, unless someone finds a serious issue. ;)
halsboss
27th December 2008, 03:57
Great ! Any news on what that means in relation to an MT version ? http://avisynth.org/tsp/
IanB
30th December 2008, 07:44
Going live, 2.5.8-RC5 [081227] is now 2.5.8 [081230] (http://sourceforge.net/project/showfiles.php?group_id=57023&package_id=72557&release_id=650255) :cool:
Well almost, Fizick, had a last minute translation to the Russian ReduceBy2 help. ;)
Ajax_Undone
30th December 2008, 08:32
Congrats IanB and thank you very much for your effort in bringing this to fruition... Thanks all who helped...
LaTo
30th December 2008, 10:50
Great!!! :thanks:
elguaxo
30th December 2008, 22:25
thanks! :)
Kurth
31st December 2008, 02:21
http://img519.imageshack.us/img519/3908/imagemmw0.png
Thanks for the new version :)
Installed the Dec 30th file but the Version command say Dec 22 :p
buzzqw
31st December 2008, 08:48
yea!
thanks for all the good work!
BHH
IanB
31st December 2008, 09:19
@Kurth,
Yes I did the full build on the 22nd. Niiyan and Fizick had some new texts translated for the installer, so I did a proof cut for them, when they confirmed their texts were okay I put up the RC5 release on the 27th, Fizick fixed his Russian ReduceBy2 help, so I recut the installer on the 30th. The Avisynth contents are the same in all 3 cuts, only the Russian and Japanese texts were corrected slightly.
Ideally the 3 releases would have been identical and all dated the 22nd, but things happened.
Leak
31st December 2008, 13:32
Shouldn't the copyright date have been extended to "2000-2008" as well? (Or just way 'til tomorrow and bump it to 2009... :D)
np: Kid606 - Oakland Highsiding (Pretty Girls Make Raves)
asarian
31st December 2008, 17:37
Outstanding job! Much appreciated! :)
IanB
31st December 2008, 22:15
Shouldn't the copyright date have been extended to "2000-2008" as well? (Or just way 'til tomorrow and bump it to 2009... :D)Damn! Missed one! Ahwell to late now :(
hajj_3
1st January 2009, 18:45
just if we could have a new gordian knot pack with this and latest dgmpgdec in, would be alot better.
leeperry
21st February 2009, 00:35
is it mandatory to have avisynth.dll in C:\Windows\system32\ ?
I've tried to copy it on my ramdisk & change the path of each avisynth.dll occurence in the registry, but if I don't have it there ffdshow refuses to work :confused:
also regsvr32 refused to register it...
squid_80
21st February 2009, 00:48
Why are you trying to store avisynth.dll on a ramdisk?
leeperry
21st February 2009, 01:17
Why are you trying to store avisynth.dll on a ramdisk?
to make ffdshow open movies faster :D
it's all on a ramdisk actually(ffdshow/kmplayer/hms+all my decoders)...only that avisynth.dll refuses to cooperate :rolleyes:
Kurtnoise
1st March 2009, 15:54
May I ask you to add something to the nsis script installer for the next release ?
WriteRegStr HKLM "SOFTWARE\AviSynth" "CurrentVersion" "${VERSION}" (line ~248)
that would be great...
:thanks:
Fizick
1st March 2009, 19:56
Kurtnoise,
what is there are several DLL versions in different folders (e.g. in VitrualDub folder), or DLL is updated by copying?
Kurtnoise
1st March 2009, 20:22
by dll, you mean avisynth.dll or the others ? Dunno frankly.
...but my point was to retrieve the avisynth version # somewhere (not via an avs script of course...;)) So, the registry seems fine to store that.
Fizick
1st March 2009, 20:34
The version info may be also retreaved from versioninfo structure of avisynth.dll ( I do not remember exact windows SDK function name).
Kurtnoise
1st March 2009, 20:41
right...but this implies to know where is located the avisynth lib. :)
leeperry
1st March 2009, 20:54
apparently it's mandatory to have it in %SystemRoot%\system32
Kurtnoise
2nd March 2009, 08:09
or in the SysWOW64 folder...
vBulletin® v3.8.11, Copyright ©2000-2025, vBulletin Solutions Inc.