foo_dsp_fsurround, a new surround processor for foobar2000 0.9.x Options

[pro_optimizer]

Thanks for the reply. I tried turning center image to 1.0 and it my foobar use 50% of my cpu. What's a good recommendation setting for center image? So with center it plays more of the "center" sounds of the sound field. And since theres a center one to play between 2 and 20 more if the sounds of like speakers 9-14 get played on the center only. Also for the center image whats the difference between 0.2 or say 0.7 is it volume or what? What setting would be good lets say with the 20 speaker example have left speaker play 1-7 center play 8-13 and right play 14-20. Does having the cetner image at 1.0 make center play 2-19?

No. In the 2-speaker setting there is a constant factor associated with every position (different, for left and right speakers), e.g. for the 20 imagined speakers:
left plays #1@100%, #2@95%, ... #10@50%, #11@45%, ..., #20@0%
center plays 0% always
right plays #1@0%, #2@5%, #3@10%, ..., #19@95%, #20@100%

In the 3-speaker setting, there is again a constant factor for each position (different for left,center,right):
left plays #1@100%, #2@90%, #3@80%, ..., #9@10%, #10@0%, #11@0%, ..., #20@0%
center plays #1@0%, #2@10%, #3@20%, ..., #9@90%, #10@100%, #11@90%, ..., #20@0%
right plays #1@0%, #2@0%, ..., #10@0%, #11@10%, #12@20%, ... #19@90%, #20@100%

So what does center image do?
It blends linearly between the factors of the 2-speaker setting and the 3-speaker setting.
This means, you can gradually blend between both settings.
Note that the positions of the sound sources do not change, i.e. an instrument 30° to the left will be 30° to the left in any of the settings (even the intermediate ones).
The only thing that differs with center enabled is that you can have a more concrete image of the close-to-center instruments, because they are played by a speaker that is actually at that position and not simulated by two speakers which are at completely different locations.

Btw: The center image setting should not influence how much CPU you foobar consumes (except there's some magic involved like "if the buffer is all 0, then don't do any calculations on it." but it's not very likely that anyone would do this since that slows down the normal processing).

Failed to load DLL: foo_dsp_fsurround.dll
Reason: This component is missing a required dependency, or was made for different version of foobar2000.

any reason as to why this would happen?

The most likely reason is that you forgot to download the libfftw3 library (it should go into windows\system32 or your foobar2000 folder)...

[mkeroppi]

Is there any way to reverse the front and rear? I'm using it with Dolby Headphone with the ER-4B and the front and rear are reversed for some reason.

[pro_optimizer]

Is there any way to reverse the front and rear? I'm using it with Dolby Headphone with the ER-4B and the front and rear are reversed for some reason.

Not really, but you could try to eliminate the reason.
I don't think that the output channels got mixed up (assuming you have recent (and non-hacked) versions of foobar, FS and DH). However, you could verify this with the channeltest.mp3.
To be absolutely sure that FS's output mapping is ok, you could put ATSurround-Headphone after it and selectively mute some of its 5 input channels.

I think its more likely that somewhere in your input chain (e.g. the CD ripper or input plugin) the polarity of one of your stereo channels got inverted. This would result in more or less switched front and rear in all common surround decoders. The fix would be to flip the channel's sign back.

If none of the above is the problem, I can only imagine that DH may be not very compatible to your ears/ears-headphone combination.

[mkeroppi]

Is there any way to reverse the front and rear? I'm using it with Dolby Headphone with the ER-4B and the front and rear are reversed for some reason.

Not really, but you could try to eliminate the reason.
I don't think that the output channels got mixed up (assuming you have recent (and non-hacked) versions of foobar, FS and DH). However, you could verify this with the channeltest.mp3.
To be absolutely sure that FS's output mapping is ok, you could put ATSurround-Headphone after it and selectively mute some of its 5 input channels.

I think its more likely that somewhere in your input chain (e.g. the CD ripper or input plugin) the polarity of one of your stereo channels got inverted. This would result in more or less switched front and rear in all common surround decoders. The fix would be to flip the channel's sign back.

If none of the above is the problem, I can only imagine that DH may be not very compatible to your ears/ears-headphone combination.

I actually tried the channeltest.mp3 before the post. I've tried ATSurround and it seems the channels are mapped correctly for the plugin, so it seems to be a problem with DH. the LRC channels seems to be behind me, as well as moving the "dimension" would seem to move the sound linearly in reverse (negative moves the sound forward while positive moves it backward).

So it's not simply a matter of flipping the front and rear channels I guess.

[narcosiss]

Heya, thanks alot for this DSP! It works much better than the standard channel mixer.

However, whenever I have it on, all my music gets a kinda 'jangly' chorusy effect on it like a low quality mp3.
I'm sure you know the type I mean, like when you stream a low quality video on a website and the audio has a jingling noise to it.

I was wondering, is there a way to prevent this or fix it?

I'm not the biggest audiophile buff so I don't know most of the terms, I do have a fairly good understanding though.

Thanks

This post has been edited by narcosiss: Apr 23 2007, 15:26

[Acropolis]

I found that this component (as well as downmix I think) has a very low performance, especially when starting a track, even I only put this in the DSP chain, it will always interrupt the plackback (I can achieve uninterrupted when manually change the track.

I think the cause is because it is using std:vector.
I've made a modification so that a lot faster and reinitialize when only needed and no interruption, if you want my modified source, please pm me.

This post has been edited by Acropolis: May 3 2007, 07:44

[poisas]

Any update will come up ? we are waiting... anyway good job

[metal_termite]

It was mentioned to use foo_channel_mixer to redirect bass to the subwoofer to work with this component, yes? Could someone help me with that, because when I put foo_chanel_mixer into the DSP chain it seems to nullify this component. How do I set them up?

Also I am just curious if someone could answer this in layman's terms. Why doesn't this component allow redirection of bass to the subwoofer? Does that mean my subwoofer does not get used?

Sorry for the dumb questions, but I'm totally clueless regarding DSPs.

[poisas]

anyone found other alternative type of mixers like fsurround ?
as we know there are atsurround but this one is #2 - no rear channel steering,

[pro_optimizer]

Hello folks,

sorry for my low post frequency lately, reason is that I have a huge load of inevitable work to do at work (and university (and elsewhere... )).
There will be progress once I chewed through all this (although I don't know if there is actually an open feature request in the queue, apart from speed optimizations).

Concerning the bass redirection, you can do this by putting Channel Mixer afterwards and redirect all bass from the 5 satelites to the subwoofer. It goes like this (at least in the version I have):

General tab: "Output channels" = 6
Upmix tab: "Mode" = off
Downmix tab: "Mode" = off
Subwoofer tab:
"Use subwoofer" = checked
"Send to subwoofer from 5.1 sources:" = "All channels"
"Bass redirection" = checked
"Type" = "Full redirection"
Delays tab: "Use delay" = unchecked

I don't know how the quality is, though (since I have no subwoofer).
You could also try to get your soundcard to do it (it should be possible, e.g. by setting output channels in channel mixer to 5 and deactivate bass redirection -- then you card has to infer the bass channel from somewhere).

@Acropolis:
Thanks for the hint. I didn't expect that the whole DSP chain is being destroyed and recreated when one changes the track. I fixed it by retaining the buffers and the processor inside an object pool. Now it takes less than 0.1 seconds to switch the tracks (including harddisk seek time).

@posias:
I hope you are writing this because you want to "bring me back" ...

Edit: One person downloaded a premature version (with a glitch when switching tracks) of the file, 2 hours ago or so. If your are the one, please redownload...

This post has been edited by pro_optimizer: Jun 14 2007, 00:44

[Rswave2k2]

I tried turning center image to 1.0 and it my foobar use 50% of my cpu. ?

Remember when I post this a while back? Well whatever that guy told you about the fix in switching songs it worked. Now when I put the center image to .70 it no longer uses 50% of the cpu. So thanks for the fix! I did reinstall my drivers so that may be it I'm not sure fixed it.

This post has been edited by Rswave2k2: Jun 19 2007, 05:11

[poisas]

Thats right pro_optimizer i hate silence,
i want you back, because youre doing great job here
waitin here for next update

[tech]

I have to same problem

Failed to load DLL: foo_dsp_fsurround.dll
Reason: This component is missing a required dependency, or was made for different version of foobar2000.

any reason as to why this would happen?

This post has been edited by tech: Jul 12 2007, 21:11

[GeSomeone]

I have to same problem

As who?
Did you read the first post that says you need also libfftw3f-3.dll? Put that in your foobar2000 directory if you haven't done so.

edit: I misspelled foobar2000

This post has been edited by GeSomeone: Aug 13 2008, 17:49

[blahdy]

apologies for rehasing an old thread here...

I was wondering, for next updated version (hopefully there will be one ), could you add a control to set the Rear speaker volumes? This would be different than Dimension slider; A simple volume slider for each 5 speakers (L, C, R, RL, RR) like how Channel Mixer has would be nice

The matrix decoding on your plugin is awesome though, great job!

[Jorgo]

I am amazed at the quality this plugin retains when put in a DSP chain, I'd like to use it as a replacement for binaural listening on headphones.

Center Image, Dimension and Stereo Separation all seem to go into this but I'm not 100% what each one does exactly:
Stereo Separation moves the speakers closer together?
Center Image moves the listener back and forth between front and rear?
Dimension makes the room smaller/bigger?

Any suggestions as to what good values to use?

This post has been edited by Jorgo: Aug 8 2008, 20:21

[pro_optimizer]

I am amazed at the quality this plugin retains when put in a DSP chain, I'd like to use it as a replacement for binaural listening on headphones.

Center Image, Dimension and Stereo Separation all seem to go into this but I'm not 100% what each one does exactly:
Stereo Separation moves the speakers closer together?
Center Image moves the listener back and forth between front and rear?
Dimension makes the room smaller/bigger?

Any suggestions as to what good values to use?

Here is the settings which I am currently using, and have been using for about a year, both with Dolby Headphone and my home surround receiver:


As you see, most values are at their defaults. This means that a vanilla stereo mix should sound as if played on a stereo system (except for echoes, which usually come from behind), and a surround-encoded mix should sound like the artist wanted it to sound.
You use the sliders if
a) your (possibly virtual) speakers are inadequately placed or chosen
b) you are listening to music which was mixed by less technically adept people
c) you want to get surround-like effects even with good old stereo music

Here is the meaning of the sliders:

High values (>1) of Stereo Separation stretch the sound sources further to the left/right (this can give you a wider soundfield), lower values (<1) give you a narrower soundfield. 1.0 means that the original (i.e. artist intended) stereo separations are retained.

Center Image does not change the location of any perceived sound. However, it controls to what degree the center speaker is allowed to reproduce center sound sources.
E.g. if you have an excellent center speaker (e.g. same as your front left/right speakers), you may want to use a high value here (e.g. 1.0). Otherwise, you use a smaller value (in the extreme case 0.0), which makes your left/right front speakers do all the work of the center, and the center does less or nothing.
Apart from the quality of your speakers, there is of course a difference between 3 same speakers in front of you versus just 2 of them. The difference is exactly that with 3 speakers it's easier for your brain to locate the center sources, which can give you a more concrete sound.

Dimension moves the sound sources towards the front (negative values), or towards the back (positive values). Slightly positive values may therefore result in more 'immersive' sound, negative values result in a more frontal, stereo-like sound.
For good surround mixes, it should not be necessary to tinker with this slider, and also if you want your vanilla stereo music to sound like stereo, you should not play with it. But, if you want to get lots of surround, even with not-so-surroundish music, you can use e.g. 0.5 here.

@blahdy: Sorry, currently, I don't plan to add such volume controls. The reason is that any surround (and even stereo) plugin would have to have these controls. But this is not what you want: A plugin is a module, which does a certain job. Mine is to decode the surround information within your stereo files, and deforming it in some ways. Controlling the volume of each speaker is actually the job of foobar2000, but since there is no such option, you need another plugin which does exactly this, and put it behind fsurround in the DSP chain.
Channel Mixer comes closest to that other plugin.

This post has been edited by pro_optimizer: Aug 9 2008, 00:31

[pro_optimizer]

Brought the initial post up to date, included some often requested information from throughout the thread...

[Walrusbonzo]

Been getting sick of having to use Creative's nasty bloated software to get everything sounding right.

Found this plugin, absolutely loving it.

Excellent work. CPU usage almost never goes over 1% even with PPHS Resampler to 96kHz in Ultra Mode.

Thanks!

[TheBlakester]

This means that plain Stereo Music which was not encoded this way sounds very similar to normal Stereo, except that echoes or very distorted sounds usually come over the surround speakers, too. Encoded music sounds much better, of course.

I'm currently using the Dolby Headphone DSP and I'm trying to determine whether it should be preceded by either the free surround dsp or the channel mixer (for upscaling to 5.1). You say the free surround dsp only has a significant effect when the music is encoded for surround playback, but I'm confused. Don't these kind of releases already include a 5.1 mix?
Thanks.

This post has been edited by TheBlakester: Sep 8 2008, 17:08

[pro_optimizer]

I'm currently using the Dolby Headphone DSP and I'm trying to determine whether it should be preceded by either the free surround dsp or the channel mixer (for upscaling to 5.1). You say the free surround dsp only has a significant effect when the music is encoded for surround playback, but I'm confused. Don't these kind of releases already include a 5.1 mix?
Thanks.

I think you're assuming that a release is either plain stereo or full-blown SACD/DVD-Audio.
But since it's easy to add surround effects here and there in a stereo mix, and since most people with surround setups have a stereo decoder, it is not that black-and-white.
There is a spectrum going from vanilla stereo music, over somewhat-enriched music, to multichannel-mastered and downmixed music (including the CD/mp3/(internet) radio versions of a 5.1 mix). The big problem is that no one ever cares to state on a CD what you will find on it, surround-wise (except for the occasional classical music label). So you can do no more than use it and see what unfolds and what doesn't, or use a channel replicator right away.

[TheBlakester]

I'm currently using the Dolby Headphone DSP and I'm trying to determine whether it should be preceded by either the free surround dsp or the channel mixer (for upscaling to 5.1). You say the free surround dsp only has a significant effect when the music is encoded for surround playback, but I'm confused. Don't these kind of releases already include a 5.1 mix?
Thanks.

I think you're assuming that a release is either plain stereo or full-blown SACD/DVD-Audio.
But since it's easy to add surround effects here and there in a stereo mix, and since most people with surround setups have a stereo decoder, it is not that black-and-white.
There is a spectrum going from vanilla stereo music, over somewhat-enriched music, to multichannel-mastered and downmixed music (including the CD/mp3/(internet) radio versions of a 5.1 mix). The big problem is that no one ever cares to state on a CD what you will find on it, surround-wise (except for the occasional classical music label). So you can do no more than use it and see what unfolds and what doesn't, or use a channel replicator right away.

Ah, I see. Thanks a lot for clearing that up.

[pro_optimizer]

Ah, I see. Thanks a lot for clearing that up.

Actually, you made me start this little public rating board for surround music, where anyone can rate and add his or her albums.
Congrats!

It's now linked in the initial post, let's see how this pans out.

This post has been edited by pro_optimizer: Sep 10 2008, 01:20

[Sounds Blast]

Nice idea pro_optimizer will be on there soon

[NullString]

Quick question: Is the public rating board for surround albums? Like Jean Michel Jarre's Aero, or for stereo albums that sound good with this plugin?