ffdshow and ffdshow_tryout Builds Listing

[LotharZ]

SVN: 2006-04-08 (rev. 2504)
Compilers used:
- MSVC8 (ffdshow.ax)
- GCC 4.1.1 (libavcodec.dll, mplayer.dll, ...)
Minimum CPU requirement: SSE
Download: click here (http://rapidshare.de/files/17494132/ffdshow-20060408_msvc8_gcc41_sse_videomixer9.exe.html)

Also unicode, not much changes, just other compilers and slighty other settings and SSE requirement.

Error during the instalation: "Error while registering ffdshow.ax"


Output folder: C:\Programs\ffdshow
Extract: ffdshow.ax... 100%
Extract: ffdshow.ax.manifest... 100%
Extract: libavcodec.dll... 100%
Extract: TomsMoComp_ff.dll... 100%
Extract: libmplayer.dll... 100%
Extract: libmpeg2_ff.dll... 100%
Extract: ff_liba52.dll... 100%
Extract: ff_wmv9.dll... 100%
Extract: ff_tremor.dll... 100%
Extract: ff_theora.dll... 100%
Extract: ff_libmad.dll... 100%
Extract: ff_libdts.dll... 100%
Extract: ff_libfaad2.dll... 100%
Extract: ff_realaac.dll... 100%
Extract: ff_samplerate.dll... 100%
Extract: ff_unrar.dll... 100%
Extract: ff_x264.dll... 100%
Extract: ff_kernelDeint.dll... 100%
Could not load: C:\Programs\ffdshow\ffdshow.ax

[videomixer9]

Missing runtime libs for Microsoft VC8?

MSVCR80.dll and MSVCP80.dll iirc ...
http://rapidshare.de/files/17503992/msvcr8runtime.7z.html
extract and put in windows or windows system32 dir

[LotharZ]

No, even with your dlls the problem isnt solved :/

[videomixer9]

edit: you don't need .NET Framework ... you just need vc8_redist ... a 2.5 MB download. I assumed it needed .NET as VC8 files also come with it and I tried on my fresh installed notebook XP via installing .NET Framework.

you may try this here:
vcredist_x86.exe (http://faux.warwickcompsoc.co.uk/vcredist/vcredist_x86.exe)

[LotharZ]

Then thats the problem :)
thx

[videomixer9]

SVN: rev. 2504
Minimum CPU requirement: x64 compatible CPU
OS: Win64
Download: click here (http://rapidshare.de/files/17577748/ffdshow64-20060409.exe.html) or here (http://www.megaupload.com/?d=GN7EQAR5) or get a better ISP

x64 build, beware, may not work at all or perform horrible, I don't have a x64 CPU so I couldn't test it.

[SeeMoreDigital]

Could you use a different "files host" please!

It's practically impossible for NTL (and probably TeleWest) broadband users in the UK to access RapidShare's free service.


Cheers

[videomixer9]

seriously go and get a better ISP ... can you call an ISP an ISP if they don't provide proper internet access?

[Lemonzest]

well NTL has just bought out Telewest in the UK, and there the only cable company's around most of the UK, and all the other ISP, are still ADSL on British Telecom lines running at about 2Mb/s tops, UK is home to monopolies, High prices and crap service no point changing ISP's as there all as bad as each other.

[GmorG McRoth]

It is legal to host custom bulids? If yes anyone could use pages.google.com for it, 100MB of free hosting space. You just need gmail account. uplaoding and page editing is thru browser, easy and fun.

[namchik]

videomixer9
i also have error registering ffdshow.ax in Windows x64

will this download help?
vcredist_x64.exe from http://www.microsoft.com/downloads/details.aspx?FamilyID=25ae0cd6-783b-4968-a841-38a2743307d9&DisplayLang=en#filelist

[videomixer9]

I guess so, it is MSVC8 again. So you need the runtime libs on Win64 too of course. It should complain about the missing libs when you try to register the filter by hand. Also I'm not to sure about the nsis installer and the installer script, I used /DX64 but i guess the nsis installer is still 32bit and prolly runs under wow32.

[namchik]

ok... but 10 megs seems too much for my dial-up :(

by the way about the broken libmpeg2 in the latest k8 build on athlon64 cpus... Disabling SSE2 in ffdshow settings solves the problem

[bob0r]

I will try some new tests, when:

1: svn co https://svn.sourceforge.net/svnroot/ffdshow/trunk ffdshow
might work, cus for me it does not work, any tips?

svn co https://svn.sourceforge.net/svnroot/ffdshow/trunk ffdshow
Error validating server certificate for 'https://svn.sourceforge.net:443':
- The certificate is not issued by a trusted authority. Use the
fingerprint to validate the certificate manually!
Certificate information:
- Hostname: *.sourceforge.net
- Valid: from Dec 8 13:40:07 2005 GMT until Feb 7 13:40:07 2007 GMT
- Issuer: Equifax Secure Certificate Authority, Equifax, US
- Fingerprint: 49:b8:cb:87:04:8c:49:39:45:83:dd:4c:cf:c7:54:57:b0:9e:84:5d
(R)eject, accept (t)emporarily or accept (p)ermanently? p

Then nothing happens.....

2: i will test gcc 3.4.5 4.0.3 4.1.0, bug report on all if compiling fails and ask Milan which version(s) he will support

If i make a new build it will be gcc only, so it depends on SF and Milan to make this possible.

[Liisachan]

I don't think this will be of any help, but just in case... Since I'd like to see/test your new builds again :)
http://ffdshow.faireal.net/tmp/svn.png
I permanently accepted that thing, thinking to myself like 'SSL certificates are too expensive. no way sourceforge could be hacked'

[Egh]

I will try some new tests, when:

1: svn co https://svn.sourceforge.net/svnroot/ffdshow/trunk ffdshow
might work, cus for me it does not work, any tips?



best thing is to ask Kuroshu. He uses gcc for his AthlonXP and Athlon/w/SSE2 builds.

But iirc he used his own builds of gcc for that. He used 2501 svn revision for his last (atm) builds.

From his website:

April 3
* Not sure which gcc version is used, as gcc 4.1.0 autovectorization hardly works and can lead to faulty binaries
* Using SVN branch (oh I am such a l33t dud3 n0w!)

April 6
* Files are first built with CC as defined in update.cfg, then the hardcoded 4.0.2.

[videomixer9]

GCC 4.1.0 prolly won't work, GCC 3.4.5 will and 4.0.3 i'm not sure about ... I only tried a 4.1.1 and that one failed, mingw gcc 3.4.4 works fine.

[LoRd_MuldeR]

kurosu's "experimental" gcc 4.2 build (AthlonXP) seems to work fine here.

[Egh]

kurosu's "experimental" gcc 4.2 build (AthlonXP) seems to work fine here.

Yeah, since 20060127 kuroshu's builds were practically best.

http://kurosu.free.fr/ffdshow-20060407-19H04-athlon.exe (sse1 only build, sauce -- svn 2504)

@Bob0r:
Now what kuroshu wrote:

April 10
* Experimental builds: gcc-4.2 is used to compile (by hand), then gcc 4.1.0 when 4.2 fails (mostly simd.h asm stuff), then 4.0.2 (baseclasses and related stuff).

[Lemonzest]

tryied the latest gcc k8 build from kurosu's site and libmpeg still crashes MPC when playing back MP1 files, works fine when libavcodec is decoding mp1/2 dont know about other builds.

[namchik]

disable SSE2 in settings and it will not crash

Did anyone try prescott builds (http://kurosu.free.fr/ffdshow-20060407-19H04-sse3.exe) on AMD 64 CPUs? Coz afaik latest processors since Venice core (if i'm not mistaken) support sse1, sse2 and sse3.
How fast are they compared to k8 buidls?

[Carpo]

Could you use a different "files host" please!

It's practically impossible for NTL (and probably TeleWest) broadband users in the UK to access RapidShare's free service.


Cheers

im on telewest and have never had a prob getting onto there - except when dns when tits up :)

[cristox]

Since a couple of months the gradual denoise is broken (at least on my system).
It is blurring, colouring and adding a kind of smear...slo-mo effekt to the picture.
The last build from the developers homepage is working.

Has anyone the same experience?

[DrNein]

I have an odd and likely very rare problem here I hope someone can help with. Perhaps consider it a challenge ;)

I am using mpc2kxp6490 to play everything with all internal source filters enabled and all internal transform filters disabled, with ffdshow-20060202-icc-sse2 handling decoding of most common video types and a just a few audio types.

Audio is output solely via S/PDIF to a HT receiver and sometimes headphones are plugged into it for isolated listening of movies, in which case Cyberlink's Audio Decoder (PowerDVD6) Speaker Environment is toggled within MPC from S/PDIF to Headphone in order to down-mix AC3 and DTS into PCM 2ch surround in one of various mock surround modes (Dolby, SRS, Cyberlink).

Naturally, the receiver can only decode AC3, DTS and PCM so for AAC 6ch I use ffdshow's output to AC3 feature to maintain true surround to 5.1 speakers rather than down-mixing to PCM.

Everything is working fine up to there. However, I started wondering how I could get WMA3 6ch to work as well (for WMV HD). ffdshow does not seem to offer WMA3 decoding (proprietary?) and so Microsoft's WMAudio Decoder DMO must be used. But I was able to figure out that enabling ffdshow's Uncompressed format decoder does the trick.

However, and here's the problem bit... when enabled it for some reason sets Cyberlink Audio Decoder's Speaker Environment to 2-Speaker and prevents it from being changed in MPC until Uncompressed is disabled again. From within PowerDVD it remains able to be changed since it of course does not use ffdshow so there is no conflict.

Any idea what is going on here and how to fix it or alternatively another way to get WMA3 output as AC3 without having to change any settings (same as AAC)?

-
Hmm, maybe the WMAudio Decoder DMO is panicking at S/PDIF and somehow setting the Cyberlink Audio Decoder to 2-Speaker in an attempt to deny 6ch digital and only allow 2ch PCM or up to 6ch analog? But these do not appear to be protected files (if that matters) and in any case CAD is not even being used as proven with GraphEdit.

[videomixer9]

just don't waste cpu cycles and more by using the AC3 out function? The soundcard drivers should bother with getting the multichannel audio properly to your receiver. The AC3 out in ffdshow decodes to PCM and then reencodes to AC3 ... my receiver gets 6ch audio fine without this from my soundcard though too, so I don't bother.

[DrNein]

CPU usage is negligible. Most PC audio has no trouble passing AC3/DTS/PCM via S/PDIF to a receiver or decoding AC3/DTS/AAC/WMA3/&c. and sending analog to a receiver however as said I only use digital out because it is not near the PC and I otherwise have no use for analog. I am not aware of any sound card driver or decoders available to pass a 6ch AAC or WMA3 stream via S/PDIF as with AC3 and DTS, so fddshow's output seems the only way around that. How do you manage it with yours? Or do you mean it has DTS Connect and/or Dolby Live to re-encode AAC and WMA3 streams which I guess is pretty much the same thing but done by a C-Media DSP instead of CPU?

[celtic_druid]

There are Dolby Live and DTS connect sound cards that do hardware AC3 and DTS encoding.

[videomixer9]

The cheaper solution usually is to use Creative's propiertary multichannel cables, I dunno how exactly it works but I used it a long time with my DTT 3500 soundsystem from creative. In the digital out of your soundcard you put an adapter for a special cable that can be connected to a special multichannel connector on the soundsystem. It's one cable and outputs anything to the soundsystem, I still wonder how this really works. After all it's the usual SPDIF output plug, no matter what kind of multichannel audio, it works through this singe cable. Question is if it is multichannel PCM or if they just send some kind of analog signal that's split up into 6 channels, after all, the decoding of things is done before. Negative thing, only Creative uses this. Also, I dunno if a second adapter in front of the receiver might magically convert it back to sth. usual receivers accept.

[celtic_druid]

http://www.hidiaudio.com/

[bob0r]

Does
svn co https://svn.sourceforge.net/svnroot/ffdshow/trunk ffdshow
work for anyone? Or is SF having some problems, if so, how do you guys get the sources?

[videomixer9]

works fine, check your routing to sf svn server or get a better ISP(tm)

[bob0r]

works fine, check your routing to sf svn server or get a better ISP(tm)

isp: amx-is best and biggest provider in the world
routing: see isp

What software do you guys use? svn.exe for mingw is what i use.

Edit:
Updated to svn, version 1.3.1 (r19032)
compiled Mar 30 2006, 02:41:28
Same result:
svn co https://svn.sourceforge.net/svnroot/ffdshow/trunk ffdshow
Error validating server certificate for 'https://svn.sourceforge.net:443':
- The certificate is not issued by a trusted authority. Use the
fingerprint to validate the certificate manually!
Certificate information:
- Hostname: *.sourceforge.net
- Valid: from Dec 8 13:40:07 2005 GMT until Feb 7 13:40:07 2007 GMT
- Issuer: Equifax Secure Certificate Authority, Equifax, US
- Fingerprint: 49:b8:cb:87:04:8c:49:39:45:83:dd:4c:cf:c7:54:57:b0:9e:84:5d
(R)eject, accept (t)emporarily or accept (p)ermanently?

I press "p", but nothing happens..... Any tips?

[videomixer9]

using this http://subversion.tigris.org/servlets/ProjectDocumentList?folderID=91 which also works fine with no mingw installed at all and never saw a request because of the certificate. Prolly you're missing the root certificate and it cannot secure the connection or establish it properly or your openssl libs are broken. Equifax root certificate usually comes with Windows. You prolly removed it or fucked it up somehow.

Get the Equifax root certificate and install it: http://www.geotrust.com/resources/root_certificates/index.asp

[bob0r]

Hmm i did checkout via CMD, the cert and checkout worked.
Tried again via mingw and now it works! Oh well.... Thnx for the tips *looks at stphn (irc)

[bob0r]

I tried compiling rev 2504 with 3 different gcc versions.
Unlike in the past, all version are failing at some point, gcc 4.0.3 should be working, but is not, so i reported this as bug (my version(s) will be 1 gcc version only)

If the bug(s) are fixed, ill host a new version again.

The bug report:
http://sourceforge.net/tracker/index.php?func=detail&aid=1469519&group_id=53761&atid=471489

[dragongodz]

What software do you guys use?
its pretty hard to go wrong with tortoisesvn aswell.
http://tortoisesvn.sourceforge.net/

[Lys]

These are all GCC compiles, any possibility of having an ICL instead?

[videomixer9]

ICL compiles most parts of this revision correct for me, except for one funny error in Tdll::loadFunction in the subtitle part. However, for me it always fails on linking the .ax file. I guess I'll try again later today, currently I'm at university. Milan seems to have planned ICL only for 64bit compiling as the makefiles have icl parameters set to win64 and amd64 target machine, so that needed fixing. Maybe something else is still wrong and that's why linking fails, didn't have a closer look yet.

[DrNein]

The cheaper solution usually is to use Creative's propiertary multichannel cables, I dunno how exactly it works but I used it a long time with my DTT 3500 soundsystem from creative. In the digital out of your soundcard you put an adapter for a special cable that can be connected to a special multichannel connector on the soundsystem. It's one cable and outputs anything to the soundsystem, I still wonder how this really works. After all it's the usual SPDIF output plug, no matter what kind of multichannel audio, it works through this singe cable. Question is if it is multichannel PCM or if they just send some kind of analog signal that's split up into 6 channels, after all, the decoding of things is done before. Negative thing, only Creative uses this. Also, I dunno if a second adapter in front of the receiver might magically convert it back to sth. usual receivers accept.

Yeah, that's a proprietary DIN for 6ch PCM. Too bad something like this did not become standard on HT receivers. There have been half-arsed/obscur efforts like the proprietary DenonLink and also a FireWire audio spec. If there were combiners available for 3 S/PDIF to HDMI then maybe any SoundBlaster that has the 4-pole digital jack could work with HDMI receivers now.

In any case, it seems we are finally on the verge of everthing coming together with HDMI. With the receivers available it is only a matter of waiting for the PC audio. Sapphire has announced an HDMI video card with combined audio output via SiL 1390 chip.

It will be interesting to see how this schtuff develops. Real-time encoders and analog will soon be passe and we won't have to compromise. Indeed, DSP soundcards should become cheaper -or at least digital-only models could be without the necessity of analog convertors or connectors etc. Then again, it will be one less reason to own a good soundcard if on-board software audio already has HDMI output (but that may be awhile).

[DrNein]

http://www.hidiaudio.com/

Yeah, as implied in post #126 I am aware of those cards but they don't do anything the software ain't doing for me now (plus, I get OpenAl and full EAX support from reliable drivers). Indeed, I suspect one of those cards would prolly necessitate toggling settings frequently as most sources are preferably streamed as-is. Right now, I just find that I have to toggle one setting in ffdshow if I want to encode WMA3 to AC3 and only so that it does not conflict with the Cyberlink decoder setting -and admittedly WMA3 will remain a seldom used novelty.

But I am still curious about this. Isn't the AC3 output supposed to work with any audio playback? I do not understand the difference between "any filter, DirectSound and WaveOut" as none seem to allow connecting to WMAudio Decoder DMO, or indeed any codec not activated from within ffdshow. Also, how is LPCM used with the Cyberlink decoder? Activating LPCM just makes noise.

Due to above limitations and as said previous I can only get WMA3 output as AC3 when the codec "Uncompressed" is activated which I understand to be akin to "Raw" on the video side (a kind of catch-all for any source being played rather than connecting to a specific decoder). So, I even tried creating a ffdshow Preset with "Uncompressed" enabled but only auto load upon a decoder match or DirectShow filter presence (using variously WMAudio Decoder DMO, WMA3, WMA*, wmadmod.dll, etc.) or 6ch match. However the preset never loaded when WMA3 material was played.

How can I either get WMA3 to be encoded as AC3 without enabling "Uncompressed" codec or otherwise only have that done by a Preset Auto Load condition which detects the WMA3 decoder?

[clsid]

SVN: 2006-04-20 (rev. 2523)
Compilers used:
- MSVC71 (ffdshow.ax, etc)
- GCC 3.4.5 (libavcodec.dll, mplayer.dll, kerneldeint.dll)
Minimum CPU requirement: MMX
Download: click here (http://rapidshare.de/files/18585145/ffdshow-svn2523-20060420.exe.html)

[JarrettH]

When I updated to the SSE2 build my image settings would crash the player when I try to use them. Like for Live TV, DVD, DivX/XviD, etc. Crashed Media Player Classic and Got All Media (for Live TV). I installed the unoptimized 2006-04-20 build instead and that seemed to fix things.

[XmSurfer]

SVN: 2006-04-20 (rev. 2523)
Compilers used:
- MSVC71 (ffdshow.ax, etc)
- GCC 3.4.5 (libavcodec.dll, mplayer.dll, kerneldeint.dll)
Minimum CPU requirement: MMX
Download: click here (http://rapidshare.de/files/18585145/ffdshow-svn2523-20060420.exe.html)

Much appreciated.:thanks:

[videomixer9]

ffdshow rev. 2527 (25th April 2006)

requires: a CPU, with SSE preferably
compilers: GCC 4.1.1 + MSVC 8
specials: hopefully high accuracy tremor, maybe libavcodec default vorbis decoder
links: here (http://rapidshare.de/files/18909341/ffdshow-rv2527.exe.html) and here (http://www.megaupload.com/de/?d=B69NO8JR)
dependencies: if registering fails get and install this (http://faux.warwickcompsoc.co.uk/vcredist/vcredist_x86.exe) first

[videomixer9]

ffdshow rev. 2529

requires: a CPU, with SSE preferably
compilers: GCC 4.1.1 + MSVC 8 (ffdshow.ax)
specials: high accuracy tremor
tested: yes, working
links: here (http://rapidshare.de/files/18954480/ffdshow-rv2529.exe.html) or here (http://www.megaupload.com/de/?d=ZVPYYON9) (identical files just other hoster)

needs MSVC8 runtime libs, otherwise fails registering!!!

[videomixer9]

ffdshow rev. 2532

requires: SSE
compilers: GCC 4.1.1 + MSVC 8
specials: high accuracy tremor
download: here (http://rapidshare.de/files/19057679/ffdshow-rev2532.exe.html) or here (http://www.megaupload.com/de/?d=E6EGBBNQ) or here (http://www.filepoint.de/download/2841-dl-ffdshow-rev2532_exe) (no nagging) (2.93 MB)

[namchik]

i guess celtic_druid resumed his builds
http://ffdshow.faireal.net/mirror/ffdshow/

btw i can't understand why ffdshow topic is splitted into 2 parts???

[Shirokuu]

This thread is just for posting new builds accompanied with links to where the binaries can be downloaded. The idea is to keep this thread clean from bugreports, questions and discussion. The other FFDSHOW thread is meant for the latter.

i guess celtic_druid resumed his builds
http://ffdshow.faireal.net/mirror/ffdshow/

btw i can't understand why ffdshow topic is splitted into 2 parts???

[videomixer9]

ffdshow rev. 2533

requires: SSE
compilers: GCC 4.1.1 + MSVC 8
specials: high accuracy tremor
download: here (http://rapidshare.de/files/19084161/ffdshow-rev2533.exe.html) or here (http://www.megaupload.com/de/?d=IZD51S29)

changes: VMR9 color controls

[videomixer9]

ffdshow rev. 2534

requires: SSE
compilers: GCC 4.1.1 + MSVC 8
specials: high accuracy tremor, more experimental gcc switches
download: here (http://www.filepoint.de/download/2870-dl-ffdshow-rev2534_exe) or here (http://rapidshare.de/files/19110555/ffdshow-rev2534.exe.html) or here (http://www.megaupload.com/de/?d=MKML8A61)