1/3 octave classic graphic eq bands

[jonathan-annett]

i'm wondering how closely the bands of an "old school classic graphic equalizer" can be displayed.

https://www.engineeringtoolbox.com/octave-bands-frequency-limits-d_1602.html

i've tried plugging various values into octave-bands-viewer.html in the tools folder, but can't find any settings that work.
is it possible to manually define an array of custom bands?

i'm hoping to make a simple tool to assist in tuning a room, which engineers can load up in their web browser, either on a laptop with a reasonable external usb sound card, or on a phone for less acurate measurement (phone mics are not that great)

[hvianna]

Hello Jonathan!

The analyzer bands are created using the equal temperament scale, based on the 440Hz pitch (A4) standard - see reference frequency table here. Using the equal temperament scale has the advantage to allow the visualization of individual musical notes when using the 1/12th octave bands, where each band corresponds to a piano key.

On the other hand, the standard octave and third-octave bands are based on a central 1000Hz frequency, which lies somewhere between B5 and C6, so the frequencies of all bands are slightly different from those of the equal temperament scale.

That being said, your question made me realize that I'm using what should be the center frequency for the band as its lower edge frequency, and that's why the values reported by the viewer tool seem even more distant from the standard bands.

I tried a different approach to generate the frequency bands and this is what the 1/3rd octave bands should look with the new code:

I plan to work more on this during the weekend, to implement the changes in the analyzer module.

Do you think that would better meet your needs?

[jonathan-annett]

Hi I appreciate that from a purist point of view, the frequencies of a classic graphic eq don't perfectly align with the actual music notes /harmonics However What I'm attempting to do is create a tool that shows those (impure) bands as they actually are. In other words I'd like to display the area of the spectrum that will be directly affected by each physical slider on the device. It appears that its possible to just call getEnergy() with upper and lower band limits. Is that a possible way to go? ( I'd just make my own hardcoded const array and loop around getEnergy() myself. ) Or if you are making changes to allow specifying the centre freq, my array can just define those instead

…

On Fri, 18 Mar 2022, 9:06 am Henrique Vianna, ***@***.***> wrote: Hello Jonathan! The analyzer bands are created using the equal temperament scale, based on the 440Hz pitch (A4) standard - see reference frequency table here <https://pages.mtu.edu/~suits/notefreqs.html>. Using the equal temperament scale has the advantage to allow the visualization of individual musical notes when using the 1/12th octave bands, where each band corresponds to a piano key. On the other hand, the standard octave and third-octave bands are based on a central 1000Hz frequency, which lies somewhere between B5 and C6, so the frequencies of all bands are slightly different from those of the equal temperament scale. That being said, your question made me realize that I'm using what should be the *center* frequency for the band as its lower edge frequency, and that's why the values reported by the viewer tool seem even more distant from the standard bands. I tried a different approach to generate the frequency bands and this is what the 1/3rd octave bands should look with the new code: [image: image] <https://user-images.githubusercontent.com/1033735/158900583-7dd72ad1-43fb-4286-bfd5-0643e1d3dc69.png> I plan to work more on this during the weekend, to implement the changes in the analyzer module. Do you think that would better meet your needs? — Reply to this email directly, view it on GitHub <#28 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADGSBMHNGX4VXAGYHWX4NODVAOUFJANCNFSM5QW7D6JQ> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[hvianna]

@jonathan-annett

In other words I'd like to display the area of the spectrum that will be
directly affected by each physical slider on the device.

Ah, I see!
I believe getEnergy() should indeed work for you, then. Please let me know how it goes.

I'll look into allowing users to specify the bands center frequency in a future release.

Cheers!

[jonathan-annett]

ok, i had a first pass at this codepen here https://codepen.io/sophtwhere/pen/qBpRdOJ cloned copy of codepen debug mode (ie single page full screen) https://audiogeek.1mb.site/ I'd appreciate any comments on how i am using the audioMotion code - is this a efficient way to do it? also, i couldn't seem to stop it calling the callback function even after the device is closed. i got around it by saving a flag, however i suspect this is not the ideal way to do it (search for "running" and you will see what i mean) jon

…

On Sat, Mar 19, 2022 at 4:34 AM Henrique Vianna ***@***.***> wrote: @jonathan-annett <https://github.com/jonathan-annett> In other words I'd like to display the area of the spectrum that will be directly affected by each physical slider on the device. Ah, I see! I believe getEnergy() should indeed work for you, then. Please let me know how it goes. I'll look into allowing users to specify the bands center frequency in a future release. Cheers! — Reply to this email directly, view it on GitHub <#28 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADGSBMGJJRFSD6YKX7E44LDVAS5CRANCNFSM5QW7D6JQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Kind regards Jonathan Annett ---- You can download my PGP public key here <https://keybase.io/jonathanannett/key.asc> (to send me encrypted mail using a tool like mailvelope <https://www.mailvelope.com/>)

[hvianna]

Pretty cool spectrogram!! I think it's as efficient as it can be done.

You can pause/resume the analyzer with toggleAnalyzer() - this will stop triggering the callback.

[hvianna]

So I had a more in-depth look on the standard octave bands specification and I'll definitely try and have it included in the next release of audioMotion-analyzer.

Please keep an eye on this issue for progress information.

[jonathan-annett]

Check out this https://www.mstarlabs.com/docs/tn257.pdf I'm working towards using a multi fft approach to optimise hi,mid,low bands

…

On Sat, 26 Mar 2022, 4:56 am Henrique Vianna, ***@***.***> wrote: So I had a more in-depth look on the standard octave bands specification and I'll definitely try and have it included in the next release of audioMotion-analyzer. Please keep an eye on this issue <hvianna/audioMotion-analyzer#29> for progress information. — Reply to this email directly, view it on GitHub <#28 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADGSBMAB45ALKBAMF7N4TI3VBX44VANCNFSM5QW7D6JQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[borisorlov21]

The sound quality is really on top compared several players while this one plays very well thank you

[hvianna]

Now available in version 24.3 🎉