George Ridgway

Hi all, Just wondering if its possible to use / create a spectrogram where frequency can extend above 20000 Hertz. The standard one I'm using in max at the moment can't do this and such a shame - Have audio recordings of bats I'm wanting to put through it.

Hi all,  Just wondering if its possible to use / create a spectrogram where frequency can extend above 20000 Hertz. The standard one I'm using in max at the moment can't  do this and such a shame - Have audio recordings of bats I'm wanting to put through it.
Thanks
George

Hi all,  Just wondering if its possible to use / create a spectrogram where frequency can extend above 20000 Hertz. The standard one I'm usi

spectrogram-above-20khz

Heterodyning first then hear and visualize?

But you'll likely need a bat detector/external hardware: the audio coming inside of your computer will never go higher than half your sampling rate. To get the high bat notes, you'll need a sampling rate in the ballpark of... 300 kHz?

Heterodyning first then hear and visualize?
But you'll likely need a bat detector/external hardware: the audio coming inside of your computer will never go higher than half your sampling rate. To get the high bat notes, you'll need a sampling rate in the ballpark of... 300 kHz?

i am pretty sure all board tools which show stuff up to the highest possible frequency. 

but like jfc says, you will never have more than what you IO supports in realtime.

in case i am wrong about the board tools you currently use, you could try changing the header of the orignal file - for example from 384kHz to 48kHz - and play the material as if it would be 48kHz, so that it runs 8 times slower than in realtime.

i am pretty sure all board tools which show stuff up to the highest possible frequency. 
but like jfc says, you will never have more than what you IO supports in realtime.

in case i am wrong about the board tools you currently use, you could try changing the header of the orignal file - for example from 384kHz to 48kHz - and play the material as if it would be 48kHz, so that it runs 8 times slower than in realtime.


The recordings I Have are 192KHz Sample Rate - so I have up to 90Khz I'm trying to image in the spectrogram. recording with an passive acoustic monitoring device.

Could you explain more about changing the header of the original file - I'm not sure what you mean.

I want them played in real time - no running slower and at lower frequencies - as I input multiple audio recordings into the spectrogram at the same time to look at frequencies.

Could you explain more about changing the header of the original file - I'm not sure what you mean.
 I want them played in real time - no running slower and at lower frequencies - as I input multiple audio recordings into the spectrogram at the same time to look at frequencies.
Thanks
George 

1. In your original post, do you mean that [spectroscope~] doesn't show frequencies up to 96 kHz when you set your sampling rate to 192 kHz (in Audio Status)?

2. Without changing the header of the file, you might load the file in a [buffer~]. I would suggest you create a [buffer~], and kind of [play~] object, etc. Then:

- give your buffer the message 'sr 192000'

- use the 'replace' message to import your sound file in the buffer~

- play at normal speed (for instance with the [play~] object). If your Max sampling rate is 96k, you'll hear the sound at half speed, one octave lower. If it's 48k, you'll hear 4 times slower, 2 octaves lower. 

When you do that, you are using the fact that the [buffer~] sampling rate is not necessarily Max sampling rate. I think Roman was suggesting to do the same thing another way, in any case to read all samples in the file, regardless of your actual Max sampling rate.

2. Without changing the header of the file, you might load the file in a [buffer~]. I would suggest you create a [buffer~], and kind of [play~] object, etc. Then:
- give your buffer the message 'sr 192000'
- use the 'replace' message to import your sound file in the buffer~
- play at normal speed (for instance with the [play~] object). If your Max sampling rate is 96k, you'll hear the sound at half speed, one octave lower. If it's 48k, you'll hear 4 times slower, 2 octaves lower. 
When you do that, you are using the fact that the [buffer~] sampling rate is not necessarily Max sampling rate. I think Roman was suggesting to do the same thing another way, in any case to read all samples in the file, regardless of your actual Max sampling rate.

Spectrogram above 20kHz