Smoothest possible noise at a given slope

[Annabanna]

I'm conducting some experiments in creating synthetic reverbs via noise generated IR's, and would like to create a 3 second sample of the smoothest possible noise with a 4.2 dB / octave slope.

My first inclination is to grab a pink noise generator in mSoundFactory, then tilt it via the DynamicEQ module and check the slope in mAnalyzer. Simple enough, but being noise, it's a bit chaotic, and I'd like to smooth out that randomness as much as possible.

My next thought is to stack instances so that the random peaks or dips at any given frequency at any given time are averaged out across the instances leading to a combined waveform that hovers closer to the aggregate value for all frequencies.

It doesn't seem to be working, though. With 8 of them stacked, the aggregate waveform seems every bit as chaotic as any of those contained within it. I'm wondering if this is because of some facet of how the RNG works, like even if I start them at slightly different spots, are they all just spitting out the same RNG at any given sample? If so, is there a way around this?

Also interested in any alternate smoothing methods like perhaps using mSpectralDynamics. Even at it's highest resolution, though, is it reacting fast enough to make a difference? Will check that out in a moment, but interested in anyone else's thoughts on how to approach this.

[Hexspa]

idk but natural and physical reverbs have resonances so a few irregularities should be fine.

[smadger]

I'm conducting some experiments in creating synthetic reverbs via noise generated IR's, and would like to create a 3 second sample of the smoothest possible noise with a 4.2 dB / octave slope.

My first inclination is to grab a pink noise generator in mSoundFactory, then tilt it via the DynamicEQ module and check the slope in mAnalyzer. Simple enough, but being noise, it's a bit chaotic, and I'd like to smooth out that randomness as much as possible.

Not entirely sure what you mean by this, what do you mean by "smoother" and "chaotic", and at what point in the process - on the raw noise signal, on the resulting convoluted signal after the impulse? I'm struggling to make sense of "smoothest possible noise".

[Ogundimu]

I don't know if i understand you correctly. Instead of the noise generator module in msf you can use an oscillator set to noise output. This will help to get rid of the randomness, because it's basically a white noise sample looping. To make it kind of a dense noise and to eleminate the hearable looping point, you can enable unison and set it to maybe 8 voices, enable high detuning and make sure you use optimized phase instead of random phase. Now you pretty much have a white noise without any kind of randomness. I confirmed it with a phase cancellation test.
In the FX section you can use MfreeformEq to create the db/octave slope you want.
And play with the smooth button in the oscillator shape section to reduce the harshness, if you want.

You can try my attempt with this code in msf.

Code: Select all

$eNrVWFtTIjsQfudXpMLrunPjZtXMbCEX11pZOcKq+xhnGkgZktkko+KvP5UJA4jKAdmX80KR5OsvnS+dTnrCb89zhh5BKip4hL2vLkbAE5FSPo1wricnLfwtroSDkch52ieJFnKhQGvKpwpVI+xhVG13u5DpWYTdr3WMqudM3BOmxuS+GO6lVAs5oi8QYa9Wa35pNhubvT2iFhH2gkZgR77-8ytTZJ6xwgUXo8vrUcao1lDAPYN1-e1uv5hrQJ4v+1dnTCQPdsLAx2hw0b0YSjGVZN6ZET6FHif3DAruIZGEMWD0hWirAEavsB2iYSokfQFZGNyUUp22gnrdx3E4yu+HLJ9SvlQnDm1z9Fqlw13b5cYIGCQaUjtVhN0vLi4nVuUfdEb4Q0fkXBce9HPGfpI5RPhKJZQxooVEhWfq4YIXtNvDGN0V-b-NbxwaPtf5YIHVq0STR1ARPmliVB1SncwGIoUIdwRXmnBtem8Htu8aklwqMPNXf3GqBC+Xvt3TEYyRTEFaeLIc+k6nsy7onL8yGM6IMv5nms7pC6SrgRtBE+NZC6PqjWC5WaX7te56XqvpB0HDb-7lDYrD0YxkgDq5lMD1UIICfdGNcBcmJGcao0fC8uJEGKyGDF1omLsRPvFw8dezf82QaS630cdOHDoFdxw6r-dh1bH6o97DvInXykYI3-0PorcvASZCznt-csIMpwmi7SGMrnKd5fqCp-C81PKAWB7QMphM69osIsI1H6Oqmgthc53-12OmLYEoQ1qE77V4UmZ+Jw6nkmQzNOIkG4tzSdO7wq91+7dpx+FQUK5Rn5GpOjOh5WH0bJJ64Hquf3p6Wm80MFrYPO18BK+9gZ8s8U7hRhy2OWGLF5BegEaFGhyUWipyS3kqnsaLDCL8nXBuvJIwoUyDpHy6n497+mZcek3eHrrIqp3aVA7kQVmtZuKpI5iQN+bclbo666XEYTvXYh1TI5HLBIpxRVUXJpQvU9CYyCnod0bisJ2SzNwjfQl-cuDJYkC58143eXbKCXv8EZjIAPV+3gyve6Pe2OQpJw5fe1DC9t-lE999T0m3gPsfDwUHxUbNtSnpA2edV6oen7Pu4krY1pokDx66BD41R9HcwisV20rR6frAbYhqc+fSeEXjH0Pjr2iCY2iCFU3tGJraiqZ+DE19RdM4hqaxomkeQ9Nc0bSOoWnFFWPixeFYwlwwYe-n0bg3LOE2kAtQgfX3wfoWG+yDDSy2tg+2ZrH1fbB1i23sg21YbHMfbNNiW-tgjb724d-jj8hqbq4E161jNMqVJpTbS+o-uFYkcSW87F952-u7joE3xoWBf6hBcKhB7VCD+qEGjUMNmocatA4yuCY8FfPi7BCuzAOrqJZ2HL3iAnNKu5LB-ySDv2IIPskQrBhqn2QwB-IHLFRCGByuxNpyzeJ-msXocQNMJFR-1qEt8y0+-zi+Tf+2oZU1s6mKVC5hF8RUkWfA07sdoIFIb2cAbAdkTOf3Epq1HZAzCUTPXH8HpPecGZep4J63A7ZMdzsQP4WGq8nkLeKc0RRsanwz1hFcS8EYSIVGlE8ZmDKCAyuPaTvR9BFsrYnKsnkswW5ZFx5pAgPQINVGTb38QmNz87IiWZiK84JPBBoSpZ6ETCNsdrkcHmmiIQ7PgUtw4tCKax60UtMkZ0XEmCYz1Y2ezZ047APRuQRlYmWLZlkjv-XY1nQbbh0xb7HI8gPDO044Sy9MDWEVLN4OG4oagWWGrokG+wIZEC3p8wZYZuZLWSGwH4e3NNWzQfmMR5uFSnnXlRCnNKqEzntf2lRc+RfDDToX

[MeldaProduction]

Exactly - MNoiseGenerator -> MFreeformEq (with extreme quality ideally)