LSP Project

Discuss anything new and newsworthy! See http://planet.linuxaudio.org and https://libreav.org/news for more Linux Audio News!

Announcements of proprietary software may fit better in the Marketplace.


Moderators: raboof, MattKingUSA, khz

User avatar
Capoeira
Established Member
Posts: 1321
Joined: Tue May 12, 2009 1:01 pm
Location: Brazil
Has thanked: 3 times
Been thanked: 2 times

Re: LSP Project

Post by Capoeira »

Capoeira wrote:a output peak meter for plugins like EQs and convolvers would be very usefull
how come when I feed a white noise through the convolver I don't get a constant peak (feeding it into Mixbus, for example)? Is the software producing those flutuations or is this a nature or IR convolution?
User avatar
Capoeira
Established Member
Posts: 1321
Joined: Tue May 12, 2009 1:01 pm
Location: Brazil
Has thanked: 3 times
Been thanked: 2 times

Re: LSP Project

Post by Capoeira »

Hi LSP devs,
as you guys seam to be the most active Linux plugin devs, is there any intrest to dev a plugin version of this software?
https://hydrogenaud.io/index.php?topic= ... #msg979219

I would sponsor this or try to crowdfund it (depending on price)
User avatar
Capoeira
Established Member
Posts: 1321
Joined: Tue May 12, 2009 1:01 pm
Location: Brazil
Has thanked: 3 times
Been thanked: 2 times

Re: LSP Project

Post by Capoeira »

User avatar
lilith
Established Member
Posts: 1698
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Has thanked: 117 times
Been thanked: 57 times
Contact:

Re: LSP Project

Post by lilith »

I'm using the LSP Spectrum Analyzer very often and I have a problem with its scalings.

It seems that e.g. brown noise scales the spectrum by +6dB per octave and purple noise (which is violet noise?) by -6dB / octave.
Other DAWs like Renoise or plugins like Fab Filter use a scaling of +4.3 or 4.5dB per octave which results in quite a flat spectrum.

Would it be possible to implement this in the analyzer? Would be much easier to compare it to these programs / plugins.
User avatar
sadko4u
Established Member
Posts: 986
Joined: Mon Sep 28, 2015 9:03 pm
Has thanked: 2 times
Been thanked: 359 times

Re: LSP Project

Post by sadko4u »

lilith wrote:It seems that e.g. brown noise scales the spectrum by +6dB per octave and purple noise (which is violet noise?) by -6dB / octave.
Other DAWs like Renoise or plugins like Fab Filter use a scaling of +4.3 or 4.5dB per octave which results in quite a flat spectrum.
Yes, 'purple noise' is in fact 'violet noise' (will rename it in new release).
What does mean 'flat spectrum'?
AFAIR, the best envelope for mastering/mixing is the pink noise because our ears are tuned to comfortable percieve the noise with such envelope.
It's not hard to add 4.3 db/oct or 4.5db/oct compensation for the analyzer but I want to know reason first, why Renoise or Fab Filter use such scaling.
LSP (Linux Studio Plugins) Developer and Maintainer.
User avatar
sadko4u
Established Member
Posts: 986
Joined: Mon Sep 28, 2015 9:03 pm
Has thanked: 2 times
Been thanked: 359 times

Re: LSP Project

Post by sadko4u »

OK, here's the answer:
The Tilt setting tilts the measured spectrum around 1 kHz with a specified slope, expressed in dB per octave. The default setting of 4.5 dB/oct results in a natural looking spectrum, resembling best how loudness is perceived by the human ear.
But:
The frequency spectrum of pink noise is linear in logarithmic scale; it has equal power in bands that are proportionally wide.[4] This means that pink noise would have equal power in the frequency range from 40 to 60 Hz as in the band from 4000 to 6000 Hz. Since humans hear in such a proportional space, where a doubling of frequency (an octave) is perceived the same regardless of actual frequency (40–60 Hz is heard as the same interval and distance as 4000–6000 Hz), every octave contains the same amount of energy and thus pink noise is often used as a reference signal in audio engineering. The spectral power density, compared with white noise, decreases by 3 dB per octave (density proportional to 1/f ). For this reason, pink noise is often called "1/f noise".
LSP (Linux Studio Plugins) Developer and Maintainer.
User avatar
lilith
Established Member
Posts: 1698
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Has thanked: 117 times
Been thanked: 57 times
Contact:

Re: LSP Project

Post by lilith »

sadko4u wrote:OK, here's the answer:
The Tilt setting tilts the measured spectrum around 1 kHz with a specified slope, expressed in dB per octave. The default setting of 4.5 dB/oct results in a natural looking spectrum, resembling best how loudness is perceived by the human ear.
But:
The frequency spectrum of pink noise is linear in logarithmic scale; it has equal power in bands that are proportionally wide.[4] This means that pink noise would have equal power in the frequency range from 40 to 60 Hz as in the band from 4000 to 6000 Hz. Since humans hear in such a proportional space, where a doubling of frequency (an octave) is perceived the same regardless of actual frequency (40–60 Hz is heard as the same interval and distance as 4000–6000 Hz), every octave contains the same amount of energy and thus pink noise is often used as a reference signal in audio engineering. The spectral power density, compared with white noise, decreases by 3 dB per octave (density proportional to 1/f ). For this reason, pink noise is often called "1/f noise".
Thanks for your answers and sorry for replying so late. I try to get the difference and do some tests here. Anyway, as Fabfilter, etc. are usually using 4.3 or 4.5 dB / octave it would be great to have such an option. When using the same scaling all the time one is getting used to it.
User avatar
lilith
Established Member
Posts: 1698
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Has thanked: 117 times
Been thanked: 57 times
Contact:

Re: LSP Project

Post by lilith »

sadko4u wrote: Mon Feb 10, 2020 10:05 pm OK, here's the answer:
The Tilt setting tilts the measured spectrum around 1 kHz with a specified slope, expressed in dB per octave. The default setting of 4.5 dB/oct results in a natural looking spectrum, resembling best how loudness is perceived by the human ear.
But:
The frequency spectrum of pink noise is linear in logarithmic scale; it has equal power in bands that are proportionally wide.[4] This means that pink noise would have equal power in the frequency range from 40 to 60 Hz as in the band from 4000 to 6000 Hz. Since humans hear in such a proportional space, where a doubling of frequency (an octave) is perceived the same regardless of actual frequency (40–60 Hz is heard as the same interval and distance as 4000–6000 Hz), every octave contains the same amount of energy and thus pink noise is often used as a reference signal in audio engineering. The spectral power density, compared with white noise, decreases by 3 dB per octave (density proportional to 1/f ). For this reason, pink noise is often called "1/f noise".
Ok, so it makes more sense to use 3dB per octave. I will test this tomorrow.
User avatar
sadko4u
Established Member
Posts: 986
Joined: Mon Sep 28, 2015 9:03 pm
Has thanked: 2 times
Been thanked: 359 times

Re: LSP Project

Post by sadko4u »

lilith wrote: Wed Mar 04, 2020 8:43 pm Ok, so it makes more sense to use 3dB per octave. I will test this tomorrow.
In 1.1.14 release of LSP Plugins I've added +4.5 dB/oct and -4.5 dB/oct envelope compensation to spectrum analyzer.
LSP (Linux Studio Plugins) Developer and Maintainer.
User avatar
lilith
Established Member
Posts: 1698
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Has thanked: 117 times
Been thanked: 57 times
Contact:

Re: LSP Project

Post by lilith »

Thanks you, that's great :D
Post Reply