Bauer stereophonic-to-binaural DSP plugin, new plugin Options

[AstralStorm]

Apparently LADSPA plugin feedback control is inversed... or it doesn't control the amount of crosfeed.
It'd be nice to have a "constant" crossfeed tunable as well, or other tuning of crosfeed EQ curve shape change option.

This post has been edited by AstralStorm: Oct 14 2009, 07:39

[O8h7w]

Using 900 Hz/1.0 dB.
Seems that higher feedback dB levels induce "sound behind head" effect for me - negate crossfeed.
Note: this is with Sennheiser IE-7 IEMs. I suspect normal headphones need less crossfeed.

As far as I know, not having tested it out much yet, you should want a way lower cutoff frequency when using In-Ear monitors. Try in the near of 400 Hz. And maybe you want to add an EQ after this plugin to tidy things up a bit.

+1 for changing on the fly. Since these settings are such that you need to listen to them to get it right, you want to do that...

And now please recompile the plugin with the SDK for foobar v 1.0 !

[lvqcl]

And now please recompile the plugin with the SDK for foobar v 1.0 !

Reasons?

[garik]

Is there way to use this plugin in Mac OS X, as mac really lacks good crossfeed filters?

[Northpack]

As far as I know, not having tested it out much yet, you should want a way lower cutoff frequency when using In-Ear monitors.

Could you explain on that?

[lvqcl]

My settings: level = 6.5 dB, frequency = 450 Hz. (Sennheiser CX-500)

[Northpack]

But why a lower cutoff frequency for IEMs?

This post has been edited by Northpack: Apr 11 2010, 21:30

[boris_mikhaylov]

Is there way to use this plugin in Mac OS X, as mac really lacks good crossfeed filters?

1) http://www.mplayerhq.hu/
2) VST plugin sources should be compileable for Mac OS, but I have not...

[Skye]

What are the odds of you making it work for RTAS so I can use it in Pro Tools for headphone mixing? Pro Tools doesn't accept VST plugins.

I'm loving it for listening, by the way. I just hope to use it for mixing.

[Northpack]

Finding the optimal cutoff frequency shouldn't be affected by the type of headphones at all. It's a matter of your head's proportions, nothing else. If your head is broad, you'll need a lower cutoff frequency. If you got a small head, you'll need a higher one.

Low frequency non-directional characteristics

At half wavelengths longer than the spacing spacing between the ears, directional characteristics of speakers are reduced, and the sound waves will diffract around the head; stereo characteristics can not be reproduced, and each ear will hear about the same sound intensity from either speaker.

The strategy to make headphones sound like speakers in a spatial environment at low frequencies is to mix both channels, equally, for frequencies below Fl, (e.g., combine low frequencies from each channel in a monophonic fashion.)

The wavelength of a sound signal, w, is:

w = v / f

where v is the velocity of sound, which is about 1100 feet per second.

The lower limit for directionality, Fl, is:

f = v / w'

were w' is twice the distance between the ears, or about a foot, or Fl = 1100 Hz.

This means that the low frequency listening environment for speakers can be approximated with headphones by crossfeeding about a factor of unity of the opposite channel's signal into the other channel-about doubling the sound intensity below Fl.

Although this is an approximation, it is reasonably close to the value used in other designs-Jan Meier used 650 Hz. in An Enhanced-Bass Natural Crossfeed Filter, and Chu Moy used 700 Hz. in An Acoustic Simulator for Headphone Amplifiers. The original Linkwitz paper used 700 Hz., also.

You can use this formula to calculate your own Fl, mine is 950 hz (I've got a bullhead ), so I'd need a rather low cutoff frequency in order to obtain the most natural sound.

This post has been edited by Northpack: May 7 2010, 13:34

[Northpack]

I figured out that there's another thing to consider for the cutoff frequency: it makes a difference for the spatial quality of a singers voice whether the common ɛ-formant (F5/700hz) is included in the crossfeed. This does suggest that you should use 700hz as a minimum. I guess that could be the reason why about all crossfeed designs came down to this frequency.

This post has been edited by Northpack: May 9 2010, 19:45

[boris_mikhaylov]

What are the odds of you making it work for RTAS so I can use it in Pro Tools for headphone mixing? Pro Tools doesn't accept VST plugins.
I'm loving it for listening, by the way. I just hope to use it for mixing.

Check up a VST to RTAS adapters like this http://www.fxpansion.com/index.php?page=15&tab=43
I think, after reading of http://www.avid.com/us/partners/audio-plugin-dev-program , a developing of RTAS plugin is not for me...

[Donunus]

Finding the optimal cutoff frequency shouldn't be affected by the type of headphones at all. It's a matter of your head's proportions, nothing else. If your head is broad, you'll need a lower cutoff frequency. If you got a small head, you'll need a higher one.

Low frequency non-directional characteristics

At half wavelengths longer than the spacing spacing between the ears, directional characteristics of speakers are reduced, and the sound waves will diffract around the head; stereo characteristics can not be reproduced, and each ear will hear about the same sound intensity from either speaker.

The strategy to make headphones sound like speakers in a spatial environment at low frequencies is to mix both channels, equally, for frequencies below Fl, (e.g., combine low frequencies from each channel in a monophonic fashion.)

The wavelength of a sound signal, w, is:

w = v / f

where v is the velocity of sound, which is about 1100 feet per second.

The lower limit for directionality, Fl, is:

f = v / w'

were w' is twice the distance between the ears, or about a foot, or Fl = 1100 Hz.

This means that the low frequency listening environment for speakers can be approximated with headphones by crossfeeding about a factor of unity of the opposite channel's signal into the other channel-about doubling the sound intensity below Fl.

Although this is an approximation, it is reasonably close to the value used in other designs-Jan Meier used 650 Hz. in An Enhanced-Bass Natural Crossfeed Filter, and Chu Moy used 700 Hz. in An Acoustic Simulator for Headphone Amplifiers. The original Linkwitz paper used 700 Hz., also.

You can use this formula to calculate your own Fl, mine is 950 hz (I've got a bullhead ), so I'd need a rather low cutoff frequency in order to obtain the most natural sound.

I don't get it. Youre saying 950hz is low? I thought If one had a big head it would go under 700hz?

[boris_mikhaylov]

w = v / f; f = v / w'
were w' is twice the distance between the ears, or about a foot, or Fl = 1100 Hz.
Although this is an approximation, it is
reasonably close
to the value used in other designs-Jan Meier used 650 Hz. in An Enhanced-Bass Natural Crossfeed Filter, and Chu Moy used 700 Hz. The original Linkwitz paper used 700 Hz., also.

The cutoff frequency of singleRC filters, like of IIR digital filter of bs2b, is a value at a half of pressure/voltage (-3dB) or at quarter of power (-6dB). The responce of these filters is half per octave. Remember this. Things above is just a very simple mathematical theory - see underlined comment. And read a references @ http://bs2b.sf.net/
http://gilmore2.chem.northwestern.edu/tech/sshd_tech.htm
http://gilmore2.chem.northwestern.edu/tech/headrm1_tech.htm
http://www.headphone.com/learning-center/f...electronics.php
Also simple but more complex note: Not only size of head, also a distance to a virtual speakers and an azimuth of one is a significant parameters for your pleasure.
My pleasure is "Def"
Pupular - "Chu Moy"
For audiophiles - who don't like big difference from original - "Jan Meier - low"
Another words, a size of head more together with distance and azimuth will leads to your favorite value of cutoff frequency, unfortunately, the delay or phase response is depend of Fcut for these filters. Additionally, the 'mix' value may help you for your audiophile prefferences, not only for virtual distance or hair on your head. On the other hand, if you are audiophile, you should love big distances and furry hair ;-)
PS
A type of record has a value, or rather - a mixing work of sound engineer...

This post has been edited by boris_mikhaylov: May 26 2010, 11:48

[Axon]

I don't get it. Youre saying 950hz is low? I thought If one had a big head it would go under 700hz?

Jesus. My bs2b is configured for 615hz based on my head size.

[hm9393]

i'm using dbpoweramp to encode from FLAC to AAC for my ipod with your VST plugin.

i'm not sure if samples passing to VST plugin are always typecasted from 16-bit audio to 32 bit float. Is it necessary to use the 'Bit depth' effect in dbpoweramp before (to 32 bit float) and after (back to 16 bit) the bs2b VST plugin ?

Many thanks !

[boris_mikhaylov]

i'm using dbpoweramp to encode from FLAC to AAC for my ipod with your VST plugin.

i'm not sure if samples passing to VST plugin are always typecasted from 16-bit audio to 32 bit float. Is it necessary to use the 'Bit depth' effect in dbpoweramp before (to 32 bit float) and after (back to 16 bit) the bs2b VST plugin ?

Many thanks !

VST interface has a flot (32bit) and a double float (64bit) *only* types for input/output buffers.
See bs2b_vst.cpp or VST SDK.
So, any VST host must pass a float type samples to VST plugin, and optionally - a double float.
Usually, a double float transfer are preffered by VST hosts if plugin have.

[boris_mikhaylov]

i'm using dbpoweramp to encode from FLAC to AAC for my ipod with your VST plugin.

i'm not sure if samples passing to VST plugin are always typecasted from 16-bit audio to 32 bit float. Is it necessary to use the 'Bit depth' effect in dbpoweramp before (to 32 bit float) and after (back to 16 bit) the bs2b VST plugin ?

Many thanks !

VST interface has a flot (32bit) and a double float (64bit) *only* types for input/output buffers.
See bs2b_vst.cpp or VST SDK.
So, any VST host must pass a float type samples to VST plugin, and optionally - a double float.
Usually, a double float transfer are preffered by VST hosts if plugin have.

Sorry, I have forgot about point of question.
I think, a conversion like 16bit integer -> 32bit integer -> 64bit float -> 16bit integer would not add any enhancement of sound than 16bit integer -> 64bit float -> 16bit integer conversion. May be, the first one would be more wrong.
PS
Floating point types of audio samples is a values from -1 to 1.

[grrr!]

I cannot manage to use the ladspa plugin on linux.

Build goes fine (libbs2b-3.1.0 then ladspa-bs2b-0.9.1) but it fails to load in any app.

jack-rack :

plugin_mgr_get_object_file_plugins: error opening shared object file '/usr/local/lib/ladspa/bs2b.la': /usr/local/lib/ladspa/bs2b.la: invalid ELF header
plugin_mgr_get_object_file_plugins: error opening shared object file '/usr/local/lib/ladspa/bs2b.so': libbs2b.so.0: cannot open shared object file: No such file or directory

ardour :

ardour: [ERROR]: LADSPA: cannot load module "/usr/local/lib/ladspa/bs2b.so" (libbs2b.so.0: cannot open shared object file: No such file or directory)

I'm on ubuntu linux 64 bits.

Thanks for your help...

[grrr!]

I cannot manage to use the ladspa plugin on linux.

Now working, I answer to myself - it may helps somebody else :
I installed libbs2b from here http://ppa.launchpad.net/stiff.ru/qmmp-rel...n/libb/libbs2b/
and then compiled ladspa-bs2b-0.9.1

This post has been edited by grrr!: Nov 27 2010, 15:28

[slecoanet]

Hi everyone,

I hope I won't sound too noob... I'd like to use bs2b using Audio Hijack under Mac Os X.
If I've understood correctly I'll need a VST or LADSPA version of bs2b but can't find any info on how to compile and install it properly in my case, the tuto I found were about Linux, and win32 precompiled version are straightforward, but not for me.

Can anyone give me a clue on how to compile a VST or LADSPA version which I can install in Audio Hijack?

Thank you in advance, I hope you didn't laugh too much as you read my post (-:

Best regards

Stephane

[boris_mikhaylov]

... If I've understood correctly I'll need a VST or LADSPA version of bs2b...

About VST:
I don't know, how. But, I know, at first, you need to get a copy of VST SDK 2.4. Docs and samples are included in the SDK.
The dependences of sources of libbs2b and VSTSDK may be seen at SDK's samples and at bs2bvst MSVC project.
Next - google (http://www.google.ru/search?client=opera&rls=ru&q=compiling+vst+OSX&sourceid=opera&ie=utf-8&oe=utf-8&channel=suggest)
Sorry, I can't help you more with OSX, but I will be glad to put a link to your success at bs2b.sf.net.

[slecoanet]

... If I've understood correctly I'll need a VST or LADSPA version of bs2b...

About VST:
I don't know, how. But, I know, at first, you need to get a copy of VST SDK 2.4. Docs and samples are included in the SDK.
The dependences of sources of libbs2b and VSTSDK may be seen at SDK's samples and at bs2bvst MSVC project.
Next - google (http://www.google.ru/search?client=opera&rls=ru&q=compiling+vst+OSX&sourceid=opera&ie=utf-8&oe=utf-8&channel=suggest)
Sorry, I can't help you more with OSX, but I will be glad to put a link to your success at bs2b.sf.net.

Thanks Boris for this quick and accurate answer.

That's what I thought, I need to check out the dependences to get the right environment to compile. I'll try my best, I'll let you know what the results are (-:

Thanks again for your help and all the efforts you put into this software.

Best regards

Stephane

[boris_mikhaylov]

bs2b audio DSP plugin for Windows Media Player released.
http://bs2b.sourceforge.net/download.html

[jaro1]

I'm simply speechless, I though it will never be done. Super, super, super, thank you very much.
Yet flawless in W7/WMP12, perfect.
Some questions, what is the purpose of that proxy/stub modul bundled in installator? Which compiler did you use, MSVC2010?
Thanks for your hard work.