ffdshow and ffdshow_tryout Builds Listing

[clsid]

Try installing this http://www.microsoft.com/downloads/details.aspx?FamilyId=73BA7BD7-ED06-4F0D-80A4-2A7EEAEE17E2&displaylang=en
ffdshow needs to be linked against unicows.lib, otherwise it won't use unicows.dll on 9x. However, last time I tried that, ffdshow would just crash.

[videomixer9]

Well then nothing that bothers me more than that, no sense fixing up ffdshow for a lonesome being keeping win98 and I sure won't turn off unicode just because of that or do other stunts ... and if your build works it's fine, plenty of builders plenty of choices!

[Inventive Software]

Meh, fair cop! Just gotta wait for XP I suppose...

[issa]

Meh, fair cop! Just gotta wait for XP I suppose...

Here is the sse gcc build without unicode, [MegaUpload] (http://www.megaupload.com/?d=BTXJIP8J)

[videomixer9]

ffdshow always the latest

requires: SSE
compilers: may vary
patches: various
download: here (http://ffdshow.drition.net/current.php)

automatically always redirects to a random mirror with my latest ffdshow build. :D

[Inventive Software]

I don't know what worked, but clsid's build seemed to do the business, but vm9's didn't.

[issa]

I don't know what worked, but clsid's build seemed to do the business, but vm9's didn't.

Is my build work on your system?

[JarrettH]

Can we close this thread. It makes no sense and I have to keep checking both for development. The first post in this thread points to the other thread. PLEASE PLEASE PLEASE KILL THIS THREAD!!:scared:

[Inventive Software]

@issa: No, it didn't. Sorry. :(
@JarrettH: This thread is for builders. The other thread is for developers. I wanted to keep the two separate, and it seems to be working. Live with it for a while. ;)

[SeeMoreDigital]

@JarrettH: This thread is for builders. The other thread is for developers. I wanted to keep the two separate, and it seems to be working. Live with it for a while. ;)Indeed....

That said it might make more sense to newbies if the thread title said something like: "FFDshow Beta Test/Development Builds: Where to get them and who does what!"


Cheers

[Episode]

ffdshow should get it's own forum section ;)

[SeeMoreDigital]

ffdshow should get it's own forum section ;)A "Decoder Setting-up and Help" section might not be a bad idea ;)

[Inventive Software]

Indeed....

That said it might make more sense to newbies if the thread title said something like: "FFDshow Beta Test/Development Builds: Where to get them and who does what!"


Cheers
Bearing that one in mind, I may change the thread title, but it would be more appropriate to change the other thread's title.


ffdshow should get it's own forum section

A "Decoder Setting-up and Help" section might not be a bad idea

Stay tuned. By the end of this week, I hope to have a simple guide to setting up ffdshow PROPERLY. This'll be in it's own thread, and ideally, stickied so newbies know where to turn.

[Liisachan]

After I updated AVG Antivirus' Virus definitions, it started bitching Downloader.Zlob.ANX was found in ffdshow-rev2546-SSE2.exe (http://ffdshow.faireal.net/mirror/ffdshow/ffdshow-rev2546-SSE2.exe) .

I already msg'ed to celtic_druid, and am 99.9% sure this is just a false positive (i.e. AVG Free is sillily bitching when there's nothing wrong there), but I'd like to re-confirm what I'm assuming.
Does anyone has any better, professional virii scanner to test this thing? Or any comment? TYIA

I already did http://virusscan.jotti.org/ and AVG is the only one who bitches.

http://ffdshow.faireal.net/tmp/avg.png

[wyrd]

@liisachan.
"am 99.9% sure this is just a false positive" me too.
http://tirnanog.fate.jp/tmp/scan/scan.png

[JarrettH]

That's very cool...VirusTotal:eek:

[issa]

@issa: No, it didn't. Sorry. :(
@JarrettH: This thread is for builders. The other thread is for developers. I wanted to keep the two separate, and it seems to be working. Live with it for a while. ;)

Could not install or the ffdshow.ax crash?

[Inventive Software]

When I try to configure the decoders the dialog box doesn't appear. Happens in the Start Menu and in the "configure" tab in VirtualDub. It installed fine!

[videomixer9]

I dunno if Win9x uses manifests but maybe deleting the external manifest may help, it defines the use of Common Controls 6.0 which may be not available on Win9x or need to be updated seperately. Though if e.g. milan version works this would be odd. But if it doesn't crash this might cause that the dialogs cannot be created. Though if I think about it you said clsid worked ... hm.

[Inventive Software]

Forgive my n00bness, but what does the manifest actually do?

[Liisachan]

AVG, that annoying antivirus that made a false accusation against celtic_druid's ffdshow binary, now updated its virus definitions, and it's now a good boy again.

This was nuts!
Virus base: 268.8.2/356
Release date: 2006-06-05 14:30:00

This is ok :)
Virus base: 268.8.2/357
Release date: 2006-06-06 12:20:00

[therealjoeblow]

ffdshow should get it's own forum section ;)

I agree, it's difficult to determine where to post comments/requests because the info is scattered all over.

Case in point, I have a request for the Keys & remote activation keys, and posted it at the following location. I'm just posting a pointer here in case that was not the correct location:

http://forum.doom9.org/showthread.php?p=843482#post843482

[videomixer9]

I was bored a bit (in fact I needed one for my old Thunderbird Athlon :P) and compiled a generic working ICL 9.1 build, found on the website (http://ffdshow.da.cx) as usual. Only requires MMX as basic.

Other than that, some people may have noticed I did some random builds, those usually only vary in compilers and don't have any real changes, there is no real gain always getting the latest one if I put on up, if your current build is working fine, please just stick to it if it's the same revision anyways and you see no new patches listed, one build e.g. has broken HuffYUV and other codecs as I tried for fun to use GCCs tree vectorizer again which is much more stable by now, as it seems it wasn't stable enough to not break at least some things, unlike before it didn't break everything though so it seemed fine to me during regular tests.

Latest build that should be completly stable should be 20060604. I know there's some people that like fiddling with new versions but don't stick with builds if they are broken for you, no extra features you can expect, only varying speeds :P If you wait for new toys and features get new versions once there's new code added by milan. That's currently not the case, new builds are just there to keep people interested in ffdshow and have sth. they can play with or randomly benchmark :P

[Liisachan]

Hm your "applied patches" list says nothing about lavc Vorbis multi-chan fix?

[videomixer9]

it is applied though ... just forgot about it after I had it listed together with other ffmpeg stuff, though I doubt anyone would regularly even notice if it were missing :P

[clsid]

SVN revision 2543
Compilers used:
- MSVC71 (ffdshow.ax, etc)
- GCC 3.4.5 (libavcodec.dll, mplayer.dll, kerneldeint.dll)
Patches:
- accuracy.diff (high accuracy tremor and libmad)
- ogg-5.1ch.diff (libavcodec 6ch vorbis fix)
- TsampleFormat.patch
- inttypes.diff
Minimum CPU requirement: MMX
Download: click here (http://www.mytempdir.com/785868)

This build is based on the latest revision before the MT patches by haruhiko.

Can someone verify that DTS works correctly? I haven't applied the DTS patch since DTS decoding works ok for me. But I only tried one sample file. So some feedback is appreciated.

[NULUSIOS]

2543?

what about 2546?

I am confused...

[Liisachan]

2546 is the newest revision as of writing this.

Yamagata Haruhiko's multithreading (MT) patch was officially applied in Rev. 2544. So 2543 is the last revision that does not include the above patch. Simply put, if you suspect "tricky" multithreading makes ffdshow less stable, you might want to try 2543 instead of 2546.

[NULUSIOS]

ah ok

[asasadad_1]

SVN revision 2543
Compilers used:
- MSVC71 (ffdshow.ax, etc)
- GCC 3.4.5 (libavcodec.dll, mplayer.dll, kerneldeint.dll)
Patches:
- accuracy.diff (high accuracy tremor and libmad)
- ogg-5.1ch.diff (libavcodec 6ch vorbis fix)
- TsampleFormat.patch
- inttypes.diff
Minimum CPU requirement: MMX
Download: click here (http://www.mytempdir.com/785868)

This build is based on the latest revision before the MT patches by haruhiko.

Can someone verify that DTS works correctly? I haven't applied the DTS patch since DTS decoding works ok for me. But I only tried one sample file. So some feedback is appreciated.

here (http://rapidshare.de/files/25070098/dts_avc1_sample.mkv.html) is a dts sample, audio is bad via ffdshow-rev2543-20060705.exe,audio is good via ffdshow-20060604-rev2546.exe (by videomixer9).

[clsid]

SVN revision: 2543
Build date: 2006-07-06
Compilers used:
- MSVC71 (ffdshow.ax, etc)
- GCC 3.4.5 (libavcodec.dll, mplayer.dll, kerneldeint.dll)
Patches:
- accuracy.diff (high accuracy tremor and libmad)
- ogg-5.1ch.diff (libavcodec 6ch vorbis fix)
- TsampleFormat.patch
- inttypes.diff
- dts.patch
Minimum CPU requirement: MMX
Download: click here (http://www.mytempdir.com/787331)

This build is based on the latest revision before the MT patches by haruhiko.

[Liisachan]

kurosu's new builds:
http://kurosu.free.fr/ffdshow.htm

Simple rebuild based on source from 2006-07-06 20:49
No SVN/CVS updating, hence no log.

* Build Athlon XP (has 3DNow, SSE1 but no SSE2), 20060706-20H07 (3207 KB)
* Build Pentium4 (has SSE1 and SSE2), 20060706-20H07 (3078 KB)
* Build Pentium 3 (only has SSE1), 20060706-20H07 (3337 KB)
* Build Prescott (has from SSE1 up to SSE3), 20060706-20H07 (3088 KB)
* Build Athlon 64 (has 3DNow, SSE1 and SSE2), 20060706-20H07 (3219 KB)

[Jubijub]

SVN revision: 2543
Build date: 2006-07-06
Compilers used:
- MSVC71 (ffdshow.ax, etc)
- GCC 3.4.5 (libavcodec.dll, mplayer.dll, kerneldeint.dll)
Patches:
- accuracy.diff (high accuracy tremor and libmad)
- ogg-5.1ch.diff (libavcodec 6ch vorbis fix)
- TsampleFormat.patch
- inttypes.diff
- dts.patch
Minimum CPU requirement: MMX
Download: click here (http://www.mytempdir.com/787331)

This build is based on the latest revision before the MT patches by haruhiko.


Absolutely wonderful : this is the first ffdshow I test where I don't have random hang when I scroll the video timeline, I can also manage to play h264 matroska without problems...

I just tried before a build I believe was build against SVN epository and I still had the above mentionned bugs...

Adopted :)

[midiboy]

Hi everyone !

Does anyone have an idea why AC3 passthrough (decoding AAC and encoding to AC3, then passing this via SPDIF) does not work anymore with some of the latest 6 channel AAC encoded Apple quicktime 7 trailers for instance this (http://images.apple.com/movies/lionsgate/theusvsjohnlennon/theusvsjohnlennon_h720p.mov) one or this (http://movies.apple.com/movies/wb/ladyinthewater/ladyinthewater-tlr3_h720p.mov) one ?

It still works fine with older clips like this (http://images.apple.com/movies/lionsgate/the_descent/the_descent_h720p.mov) one.

I tried all the latest ffdshow builds. The only way I can decode this properly is via the 3ivx Audio decoder, however encoding and passing via SPDIF then does not work ...

Thanks for your help !
Alex

[SeeMoreDigital]

Most probably because Apple elected to associate extra data for those particular 6Ch AAC-LC streams within the MOV container.

The MOV container is quite different to the MP4 container - despite what some people might like to think.

[videomixer9]

Well, the clips themselves decode fine for me, the AC3 encoding produces some stuttering more high pitched sounds though.

[Taurus]

SVN revision: 2543
Build date: 2006-07-06
Compilers used:
- MSVC71 (ffdshow.ax, etc)
- GCC 3.4.5 (libavcodec.dll, mplayer.dll, kerneldeint.dll)
Patches:
- accuracy.diff (high accuracy tremor and libmad)
- ogg-5.1ch.diff (libavcodec 6ch vorbis fix)
- TsampleFormat.patch
- inttypes.diff
- dts.patch
Minimum CPU requirement: MMX
Download: click here (http://www.mytempdir.com/787331)

This build is based on the latest revision before the MT patches by haruhiko.

My virus scanner says your build is the "TR/Drop.Zlob.LH.3" trojan.
Can you please investigate this?

Cheers

Taurus

[clsid]

Conclusion: crappy virus scanner

[Taurus]

Conclusion: crappy virus scanner
I will investigate on my own expense ;)

[sillKotscha]

Hi everyone !

Does anyone have an idea why AC3 passthrough (decoding AAC and encoding to AC3, then passing this via SPDIF) does not work anymore with some of the latest 6 channel AAC encoded Apple quicktime 7 trailers

Personally I gave up trying to "fix" it - I've tried a lot of settings I can think of - but the only version that does really work for me is still milan's latest release... and I stick with it.

remember one thing, for other members/ users realtime aac-> ac3 transcoding still works with videomixer's builds, so in our case maybe it is a hardware/ technical issue.

I guess videomixer didn't touch the spdif part, once it worked why should he change it... so, maybe it is like SMD said and apple changed the way they store 6ch aac in mov - who knows?

sill

[videomixer9]

My virus scanner says your build is the "TR/Drop.Zlob.LH.3" trojan.
Can you please investigate this?

Cheers

Taurus

Finally stop with this f*cking Zlob already, it's totally annoying seeing lame companies blacklist the NSIS installers. If this goes on with stupid false claims of viruses in open source products I'll seriously think about sueing those companies for badmouthing and ask them to pay billions of compensation to the authors of all projects that are affected. If that'd happen maybe milan and others wouldn't have to disappear for months randomly :P

[Taurus]

Conclusion: crappy virus scanner
So OK!
You're right.
The binary is clean.
Sorry

Cheers

Taurus

[videomixer9]

Always happy to see people trust clueless companies that much, I bet it was that crap called AntiVir, maybe you should scan that for some unwanted behaviour first, it phones home more than ffdshow or most trojans do. First give out free copies everywhere and then start using it to make open source look like it's infected. Great plot.

[breez]

First give out free copies everywhere and then start using it to make open source look like it's infected. Great plot.

Indeed :rolleyes:

[celtic_druid]

Must be quite a few closed source commercial apps using NSIS though yeah? At least I think I can recall installing non OSS that used NSIS.

[videomixer9]

Well NSIS is open source, I bet the company producing InstallShield bought some virus detections to come out with a new slogan "don't get your products identified as virii" before people have a chance to install them ;P

[Inventive Software]

@celtic_druid: DivX for one.

[videomixer9]

I sure hope anyone got that this is a joke on the paranoia and wild assumptions people create when they got too much time on their hands, so with people that honestly think they have to investigate shit like viruses in the builds available here.

[Taurus]

I sure hope anyone got that this is a joke on the paranoia and wild assumptions people create when they got too much time on their hands, so with people that honestly think they have to investigate shit like viruses in the builds available here.
I don't like your words.
Better get back on topic.
Wanted to be helpful and show you guys something from the outside world....
Using ffdshow right from the start years ago, but for the first time ever I got a virus alarm on it (even so it was a false alarm).
So I thought that it might be interesting to see for the devs, compilers and users.
I duck my head now and dissapear in the vanish.

Cheers

Taurua

[clsid]

VirusTotal (http://www.virustotal.com) is a good site to get a second-opinion on virus alerts.