Werbung: SecurityConsole.de verwaltet Ihre Computer mit Security Essentails aus der Cloud!
30 Tage kostenfrei testen und 20% Rabatt für Ihre Bestellung mit Promocode: WBF2685582
(Promocode gültig bis 31.12.2011)

Group:  English: Entertainment » microsoft.public.windowsmedia.sdk
Thread: Flushing or clearing/resetting push sink of publishing point

HTVi
TV Discussion Newsgroups

Flushing or clearing/resetting push sink of publishing point
Stas 4/15/2007 9:56:01 AM
Hi.

I have a pushing sink sending video to publishing point, that as the time
passes takes more and more memory, and also causing significant time delay in
stream. It's as this sink can't push the video to the publish point fast
enough.

I'm looking to periodically to flush or reset this sink, without
disconnecting the users connected to the publish point. Any idea how to do
this?

Thanks,
Stas.
RE: Flushing or clearing/resetting push sink of publishing point
KBoek 4/16/2007 5:34:04 AM
Hi Stas,

Is it really the push sink that takes the memory? It sounds more logical to
me if it's the writer, since the writer can allocate buffers. Unfortunately,
I don't know about a method to clear the buffers other than calling
IWMWriter::EndWriting; the Flush method is no longer used. I'll search some
more on this one, since I've actually run in to it myself a few times.

By the way, did you manage to properly implement authentication on your push
sink by use of the IWMCredentialCallback interface? If so, I'd be eager to
know, since I've been struggling with that for the last week :S

Regards,
Karel

"Stas" wrote:

[Quoted Text]
> Hi.
>
> I have a pushing sink sending video to publishing point, that as the time
> passes takes more and more memory, and also causing significant time delay in
> stream. It's as this sink can't push the video to the publish point fast
> enough.
>
> I'm looking to periodically to flush or reset this sink, without
> disconnecting the users connected to the publish point. Any idea how to do
> this?
>
> Thanks,
> Stas.
RE: Flushing or clearing/resetting push sink of publishing point
Stas 4/16/2007 5:56:11 PM
Hi Karel.

Actually, in my case all what was needed is to switch the writer to live
mode, via the SetLiveSource() function. When the writer in this mode it
automatically discards packets the moment there is a delay of 3 seconds.

About the authorization, I'm wrestling this as well. I was not even able to
reach the "AcquireCredentials" stage, as the application crashes on me.

Can you please post an example how to implement this?

Thanks,
Stas.

"KBoek" wrote:

[Quoted Text]
> Hi Stas,
>
> Is it really the push sink that takes the memory? It sounds more logical to
> me if it's the writer, since the writer can allocate buffers. Unfortunately,
> I don't know about a method to clear the buffers other than calling
> IWMWriter::EndWriting; the Flush method is no longer used. I'll search some
> more on this one, since I've actually run in to it myself a few times.
>
> By the way, did you manage to properly implement authentication on your push
> sink by use of the IWMCredentialCallback interface? If so, I'd be eager to
> know, since I've been struggling with that for the last week :S
>
> Regards,
> Karel
>
RE: Flushing or clearing/resetting push sink of publishing point
KBoek 4/16/2007 7:26:01 PM
My pleasure, what programming language do you prefer (C++ / C#)?

"Stas" wrote:

[Quoted Text]
> Hi Karel.
>
> Actually, in my case all what was needed is to switch the writer to live
> mode, via the SetLiveSource() function. When the writer in this mode it
> automatically discards packets the moment there is a delay of 3 seconds.
>
> About the authorization, I'm wrestling this as well. I was not even able to
> reach the "AcquireCredentials" stage, as the application crashes on me.
>
> Can you please post an example how to implement this?
>
> Thanks,
> Stas.
>
> "KBoek" wrote:
>
> > Hi Stas,
> >
> > Is it really the push sink that takes the memory? It sounds more logical to
> > me if it's the writer, since the writer can allocate buffers. Unfortunately,
> > I don't know about a method to clear the buffers other than calling
> > IWMWriter::EndWriting; the Flush method is no longer used. I'll search some
> > more on this one, since I've actually run in to it myself a few times.
> >
> > By the way, did you manage to properly implement authentication on your push
> > sink by use of the IWMCredentialCallback interface? If so, I'd be eager to
> > know, since I've been struggling with that for the last week :S
> >
> > Regards,
> > Karel
> >
RE: Flushing or clearing/resetting push sink of publishing point
Stas 4/16/2007 7:36:02 PM
C++ will be great, thanks.

Stas.

"KBoek" wrote:

[Quoted Text]
> My pleasure, what programming language do you prefer (C++ / C#)?
>
> "Stas" wrote:
>
> > Hi Karel.
> >
> > Actually, in my case all what was needed is to switch the writer to live
> > mode, via the SetLiveSource() function. When the writer in this mode it
> > automatically discards packets the moment there is a delay of 3 seconds.
> >
> > About the authorization, I'm wrestling this as well. I was not even able to
> > reach the "AcquireCredentials" stage, as the application crashes on me.
> >
> > Can you please post an example how to implement this?
> >
> > Thanks,
> > Stas.
> >
> > "KBoek" wrote:
> >
> > > Hi Stas,
> > >
> > > Is it really the push sink that takes the memory? It sounds more logical to
> > > me if it's the writer, since the writer can allocate buffers. Unfortunately,
> > > I don't know about a method to clear the buffers other than calling
> > > IWMWriter::EndWriting; the Flush method is no longer used. I'll search some
> > > more on this one, since I've actually run in to it myself a few times.
> > >
> > > By the way, did you manage to properly implement authentication on your push
> > > sink by use of the IWMCredentialCallback interface? If so, I'd be eager to
> > > know, since I've been struggling with that for the last week :S
> > >
> > > Regards,
> > > Karel
> > >
RE: Flushing or clearing/resetting push sink of publishing point
KBoek 4/16/2007 7:40:02 PM
OK, see your other post about this

"Stas" wrote:

[Quoted Text]
> C++ will be great, thanks.
>
> Stas.
>
> "KBoek" wrote:
>
> > My pleasure, what programming language do you prefer (C++ / C#)?
> >
> > "Stas" wrote:
> >
> > > Hi Karel.
> > >
> > > Actually, in my case all what was needed is to switch the writer to live
> > > mode, via the SetLiveSource() function. When the writer in this mode it
> > > automatically discards packets the moment there is a delay of 3 seconds.
> > >
> > > About the authorization, I'm wrestling this as well. I was not even able to
> > > reach the "AcquireCredentials" stage, as the application crashes on me.
> > >
> > > Can you please post an example how to implement this?
> > >
> > > Thanks,
> > > Stas.
> > >
> > > "KBoek" wrote:
> > >
> > > > Hi Stas,
> > > >
> > > > Is it really the push sink that takes the memory? It sounds more logical to
> > > > me if it's the writer, since the writer can allocate buffers. Unfortunately,
> > > > I don't know about a method to clear the buffers other than calling
> > > > IWMWriter::EndWriting; the Flush method is no longer used. I'll search some
> > > > more on this one, since I've actually run in to it myself a few times.
> > > >
> > > > By the way, did you manage to properly implement authentication on your push
> > > > sink by use of the IWMCredentialCallback interface? If so, I'd be eager to
> > > > know, since I've been struggling with that for the last week :S
> > > >
> > > > Regards,
> > > > Karel
> > > >

Home | Search | Terms | Imprint Contact
Newsgroups Reader - provided by WiredBox.Net
Suche nach Orten, Städten, Postleitzahlen, Vorwahlen, Kfz-Kennzeichen