# Rough frequency analysis **URL:** \ **Category:** FMOD Engine\ **Created:** [June 13, 2024, 1:38pm UTC](https://qa.fmod.com/t/rough-frequency-analysis/21760 "2024-06-13T13:38:14Z")\ **Posts on this page:** 3\ **Page:** 1
**Author:** ![jbat100](https://yyz2.discourse-cdn.com/flex036/user_avatar/qa.fmod.com/jbat100/32/6843_2.png) [@jbat100](https://qa.fmod.com/u/jbat100)\ **Post date:** [June 13, 2024, 1:38pm UTC](https://qa.fmod.com/t/rough-frequency-analysis/21760/1 "2024-06-13T13:38:14Z")
Hello. I want to drive visuals using energy in different frequency bands for a given unity event instance. I only need very rough granularity (lows, mids, highs). Typically, frequency analysis would be done using FFT, but this overkill and not really feasible given performance constraints. I’m looking into using a set of three simple (biquad or similar) DSP filters, then use getMeteringInfo to have an estimate for each frequency band. A few questions: - Is there a better way of doing this? - Is routing the signal to analysis DSP nodes as well as keeping the normal playback route (to keep playback unchanged) reasonable? - If I have N sounds I want to analyse, does this mean creating N\*3 channels (three bands for each) - The simple high/low pass filters are marked as obsolete [https://www.fmod.com/docs/2.00/api/effects-reference.html#high-pass-simple](https://www.fmod.com/docs/2.00/api/effects-reference.html#high-pass-simple) [https://www.fmod.com/docs/2.00/api/effects-reference.html#low-pass-simple](https://www.fmod.com/docs/2.00/api/effects-reference.html#low-pass-simple) is there a way to get separate signals for the different filters is a 3EQ Thanks for any help or guidance ---
**Author:** ![Leah\_FMOD](https://yyz2.discourse-cdn.com/flex036/user_avatar/qa.fmod.com/leah_fmod/32/3685_2.png) [@Leah\_FMOD](https://qa.fmod.com/u/Leah_FMOD)\ **Post date:** [June 19, 2024, 1:42am UTC](https://qa.fmod.com/t/rough-frequency-analysis/21760/2 "2024-06-19T01:42:52Z")
> [@jbat100](#): > > Is there a better way of doing this? If you only need low granularity, filtering the signal and obtaining metering info for each band would probably be the best option. > [@jbat100](#): > > Is routing the signal to analysis DSP nodes as well as keeping the normal playback route (to keep playback unchanged) reasonable? Yes, you can do this by either by using [Send DSPs](https://fmod.com/docs/2.02/api/effects-reference.html#send), or by outputting to multiple DSPs using [`DSPConnection`](https://fmod.com/docs/2.02/api/core-api-dspconnection.html)s. You can then connect the filter DSP for each band back to the signal chain at a later point, but mute their output with [`DSP::setWetDryMix`](https://fmod.com/docs/2.02/api/core-api-dsp.html#dsp_setwetdrymix) to ensure they don’t contribute to the mix. The `dsp_effect_per_speaker` Core API example included with the FMOD Engine installer should serve as an example of how to go about outputting to multiple DSPs like this. > [@jbat100](#): > > If I have N sounds I want to analyse, does this mean creating N\*3 channels (three bands for each) You won’t need to create 3 channels for each sound, but you will need to create the 3 filtering DSPs, and set each of them up to receive a signal, filter, and so on for a given sound. > [@jbat100](#): > > The simple high/low pass filters are marked as obsolete > [https://www.fmod.com/docs/2.00/api/effects-reference.html#high-pass-simple](https://www.fmod.com/docs/2.00/api/effects-reference.html#high-pass-simple) > [https://www.fmod.com/docs/2.00/api/effects-reference.html#low-pass-simple](https://www.fmod.com/docs/2.00/api/effects-reference.html#low-pass-simple) > is there a way to get separate signals for the different filters is a 3EQ 3EQ or Multiband EQ should both work for your purposes - 3EQ is probably a little simpler to interact with and match crossovers, while Multiband is more flexible. ---
**Author:** ![jbat100](https://yyz2.discourse-cdn.com/flex036/user_avatar/qa.fmod.com/jbat100/32/6843_2.png) [@jbat100](https://qa.fmod.com/u/jbat100)\ **Post date:** [June 19, 2024, 3:59pm UTC](https://qa.fmod.com/t/rough-frequency-analysis/21760/3 "2024-06-19T15:59:13Z")
Thanks for all this info, really useful. Got it working with the send/return and using wet/dry mix to avoid the filtered signals contributing the the mix.