AviSynth 2.5.8 [Dec 30th] - Release! (was RC-5 [Dec 27th])

[Wilbert]

The changes are fine with me.

[Fizick]

Adding show="true" will display the actual values on the screen.)

What is usefulness of "show" parameter in ScriptClip?

[Gavino]

Originally Posted by Fizick

What is usefulness of "show" parameter in ScriptClip?

I imagine it's for debugging. Particularly if the filter string is constructed dynamically (eg by string concatenation which is a common approach inside functions).

It's curious that 'show' in ScriptClip (and FrameEvaluate) disables the actual filtering (it just returns the input clip) whereas in ConditionalFilter it acts in addition to the normal function.

[Fizick]

OK, I uploaded the updated text to CVS.
I hope that new text is fine not only with Wilbert, but with Avisynth code too

[IanB]

Changelist with respect to 2.5.8 RC4 [080831] :

Changes:

* Updated Russian documentation (Fizick).

Note : Other contents are the same as AviSynth_080831.exe


Changelist with respect to 2.5.8 RC3: See 1st post for full change list.

Additions:

* Updated Japanese documentation [20080813] (Niiyan).

Bugfixes:

* Fixed DirectShowSource() grey frames after end of stream + 30 frames.
* Fixed Histogram() Classic mode restore graph brightness. Regression.
* Fixed Compare() graph pixel values exceeding YUV limits.
* Fixed AddBorders() args negative value clamping.

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.

As usual download from Sourceforge.
Enjoy!


If we do not find any show stopper problems this could become the official 2.5.8 release.

[Quark.Fusion]

TCPsource/TCPserver sometime lost connection and broke my encode, also AviSynth crash a lot when I run script with TCPsource second time. (not tested with today's build)
Why TCPsource not tries to reconnect?

[IanB]

@Quark.Fusion,

TCPsource/TCPserver are unchanged from 2.5.7

[Fizick]

IanB,
sorry, but I made litle changes of doc again
(more compact look of overview.htm, english and russian)

[Fizick]

and again

[tacman1123]

Not sure if this is the right place to report possible bugs for RC4, if not, please let me know where is a better place.

I installed RC4, and the sound came out slower, a much lower frequency. I replaced DirectShowSource.dll with an older one

12/24/2006  05:12 AM            24,576 DirectShowSource.dll

and it started working again. Script was fairly trivial, but I'd be happy to post it (and my avi) if someone would like it. My AVI files have been somewhat problematic, especially in regard to sound (I have another thread here detailing that issue), so it's certainly possible that it's a bad avi, but since it is working with an older DirectShowSource.dll, I thought I'd mention it.

Tac

[Fizick]

tacman1123, of course, please provide link to your previous thread (post), and may be some Avi fragment.

[IanB]

@tacman1123,

Post the log from DirectShowSource(... , logfile="tac.log", logmask=-1) for both d..s..s...dll version!

[tacman1123]

cool, I'll do that! (The avi file is very big). Of course, now I'm having a hard time replicating the problem. Question: if I rename the dll in my plugins directory, is it definitely using the new version when the script runs? Or might it be caching a version? Is there a way I can force my script to use one over the other? I tried simply renaming the dll and calling it with that name, but that didn't work.

[IanB]

Use the {dllname}_{functionname} form of access the functions. i.e. DllA_DirectShowSource() and DllB_DirectShowSource()

And the logfile has a version stamp as the 1st line.

[Fizick]

small doc update again
(SetMemoryMax according to http://forum.doom9.org/showthread.ph...19#post1177819 )

[Wilbert]

Someone reported:

My AviSynth is version 2.58, build date Aug 30, 2008. I have tried "ntsc_video" several times & I always get the same error message.

... being "ntsc_video" not recognized.

[Gavino]

Originally Posted by Wilbert

Someone reported ... "ntsc_video" not recognized.

No, the error message he reported was "Video framerate doesn’t match" when importing into DVD slideshow GUI, which makes me suspect the fault lies with that program.

Further evidence is that he also got the same error for
AssumeFPS(30000,1001,true).

[DeathTheSheep]

Ever since 2.5.8 alpha 2, directshowsource with convertfps is terrible. The resultant video is very choppy. All I'm doing is converting an mkv's framerate from 24000/1001 to 23.976. I've tried every release up until RC4a, and all have the same glaring bug. Going back to alpha 1 [070518] works perfectly. I've reported this bug before and was told simply that directshowsource was hopelessly b0rked. Well, over a year's worth of releases, and getting to a late 'release candidate' stage... the bug's still here.

Great work nonetheless!

[IanB]

@DeathTheSheep,

Seems you have made 3 posts alluding to the subject, 12th November 2007, 03:38, 28th March 2008, 07:50 and 28th March 2008, 09:15. None really form a concise bug report with any evidence, logs or supporting material.

CVS history of directshow_source.cpp

----------------------------
revision 1.28
date: 2008/06/12 04:17:26;  author: ianb1957;  state: Exp;  lines: +46 -7
Remember Allocator given in NotifyAllocator(), return it for GetAllocator(); 
GetAllocatorRequirements() E_NOTIMPL.
----------------------------
revision 1.27
date: 2008/05/26 22:22:17;  author: ianb1957;  state: Exp;  lines: +106 -47
Add TID + PrintState to log, nuke PulseEvent + Use IGraphBuilder (Avery Lee),
Fix QueryId() need CoTaskMemAlloc (Dean Pavlekovic),
Fix QueryAccept change VideoInfo (Haali).
----------------------------
revision 1.26
date: 2007/10/08 23:09:39;  author: ianb1957;  state: Exp;  lines: +14 -6
get convertfps=false logic right, fix Flush task interlock.
----------------------------
revision 1.25
date: 2007/09/11 06:43:04;  author: ianb1957;  state: Exp;  lines: +122 -18
Refactor GetFrame, ConvertFPS now only relies on get_sample.GetSampleStartTime,
fix jitter problems with recent M$ asf spliter giving bogus SampleEndTime values.
----------------------------
revision 1.24
date: 2007/05/01 05:49:58;  author: ianb1957;  state: Exp;  lines: +6 -1
support WAVE_FORMAT_EXTENSIBLE
----------------------------

I guess you are complaining vaguely about the changes at version 1.25/26 2007 Sep 11 & Oct 08. I completely rewrote the GetFrame code to only use the DirectShow sample start times and not use the stop times at all. The previous code used the DirectShow sample stop time and failed miserably when it was not correctly provided. A few recent splitters seem to be not setting this value correctly.

The algorithm I implemented is to calculate the start time of the requested frame number. Compare this value with the DirectShow sample start time of the next sample, for as long as the value is strictly less than, keep returning copies of the current frame. I have extensively tested the code and it does exactly as intended.

I have to assume the algorithm is not what you require, but you have not offered any suggestion for an alternate algorithm.

One issue I am aware of that may be affecting your results is timecode rounding that a few splitters seem to do. For a CFR 23.976 fps stream, we would expect timecodes to be :-

--- 0, 4170837, 8341675, 12512512, 16683350, 20854187, 25025025, ..., 95929262, 100100100, 104270937, ...

but a splitter might give something like this :-

--- 0, 4200000, 8300000, 12500000, 16700000, 20900000, 25000000, ..., 95900000, 100100000, 104300000, ...

This results in returned samples being 0, 0, 2, 3, 3, 4, 6, ..., 23, 24, 24, ... not exactly what one might find useful.

You can limp around this particular issue by doubling the request framerate and selecting the odd samples, i.e.

DirectShowSource(... , fps=23.967*2, convertfps=true).SelectOdd()

But the bottom line is you need to understand your material and the Directshow software you are using and script accordingly.

[Fizick]

Directshow si never will be finished complitely
In any case, Directshowsource is a plugin and may be updated later (if needed), as an small independent upgrade to Avisynth 2.5.8.
Also for same reason DeathTheSheep can use older (alpha 1) DirectshowSource if he do not want provide a log .