Looking for music with meaningful ultrasonic content

[vintologi]

I plan to do compare sample-rates blind later (when i have better speakers) but i am already looking for good test tracks.

I want to see how high in frequency i can detect a lowpass filter once i have good equipment.

I am especially interested in testing if ultrasonics above say 24 Khz can make a difference (when there is basically zero chance of me hearing them directly).

 

[Blumlein 88]

Tell you a better way to do this than music. If you can get a high bandwidth microphone, then record some jangling keys. 3 metal keys on a metal keyring is about optimum. You can hear it, but the ultrasonics are far louder than what you can hear. If you can't do that, then maybe any decent microphone recording jangling keys then speed it up 2x normal. Maybe someone with an Earthworks wideband mic can do a recording for you.

 

[antcollinet]

Good luck with that - both finding test tracks with meaningful ultrasonic content, AND hearing a difference if you do.

 

[nerdemoji]

Honestly, first id go into tone generator and see if you can even hear that high. I can only really perceive up to 18khz

also i can’t imagine anyone has meaningful content up there given that almost nobody can hear it

 

[vintologi]

Tell you a better way to do this than music. If you can get a high bandwidth microphone, then record some jangling keys. 3 metal keys on a metal keyring is about optimum. You can hear it, but the ultrasonics are far louder than what you can hear. If you can't do that, then maybe any decent microphone recording jangling keys then speed it up 2x normal. Maybe someone with an Earthworks wideband mic can do a recording for you.

Yea i don't have a good mic yet but i might get earthworks M50 + 1021.

 

[vintologi]

Honestly, first id go into tone generator and see if you can even hear that high. I can only really perceive up to 18khz

My hearing starts to drop off significantly at 16K now and after that it depends on volume.

People have been able to hear pure tones as high as 28K but that was at over 100 dB.

Hearing thresholds for pure tones above 16kHz

Hearing thresholds for pure tones between 16 and 30kHz were measured by an adaptive method. The maximum presentation level at the entrance of the outer ear

pubs.aip.org

also i can’t imagine anyone has meaningful content up there given that almost nobody can hear it

What i am interesting in finding out is if tones that are clearly inaudible by themselves due to being say 30Khz+ can make a difference when it's mixed with lower frequency sounds.

 

[antcollinet]

My hearing starts to drop off significantly at 16K now and after that it depends on volume.

People have been able to hear pure tones as high as 28K but that was at over 100 dB.

Hearing thresholds for pure tones above 16kHz

Hearing thresholds for pure tones between 16 and 30kHz were measured by an adaptive method. The maximum presentation level at the entrance of the outer ear

pubs.aip.org

What i am interesting in finding out is if tones that are clearly inaudible by themselves due to being say 30Khz+ can make a difference when it's mixed with lower frequency sounds.

It can't. The ear has no mechanism to detect or respond to the high frequencies. The ear itself acts as a low pass filter.


Some might say that the high frequencies in a speaker that can reproduce them might intermodulate in the speakers to cause IMD products within the audible spectrum. However given that the high frequencies will already be at a very low level, it is pretty unlikely that any IMD products will be at an audible level. Then, even if they were, IMD sounds bad - you are not going to get an improvement in sound going down this road.

 

[Philbo King]

Don't forget Nyquist... Record at 96K or higher sample rate. Even then there is a good chance the audio interface you record with will low-pass the audio at 24 to 30Khz. But you could test for this ahead of time by using a loopback cable & sending a swept sine ending at maybe 48 KHz to see what frequency range it'll actually record.

 

[vintologi]

Don't forget Nyquist... Record at 96K or higher sample rate. Even then there is a good chance the audio interface you record with will low-pass the audio at 24 to 30Khz. But you could test for this ahead of time by using a loopback cable & sending a swept sine ending at maybe 48 KHz to see what frequency range it'll actually record.

I don't need an audio interface, i can just use my old asus xonar STX and an mic pre-amp.

Earthworks 1021 should be good enough but that feels a bit wasteful in terms of $.

Ill record at 192 unless/until i get a better dac (which i will probably not need).

 

[radix]

You should make sure the ADC and mic-preamp are not applying low pass filters. Some device might also have a lot of noise above 30 kHz or so. I guess I'd summarize this as first test your test equipment to see how clean it is.

 

[vintologi]

You should make sure the ADC and mic-preamp are not applying low pass filters. Some device might also have a lot of noise above 30 kHz or so. I guess I'd summarize this as first test your test equipment to see how clean it is.

Oh yea i kinda forgot.

Getting an external ADC that can be used with a laptop might be a good idea.

I could also take measurements in another room and use a really long cable.

My computer is fairly quit during idle but not fully obviously.

 

[DVDdoug]

What's "meaningful"?

What i am interesting in finding out is if tones that are clearly inaudible by themselves due to being say 30Khz+ can make a difference when it's mixed with lower frequency sounds.

Not unless there is intermodulation distortion. Or, if you're an "audiophile" who "hears things."

OK... Do you have Audacity? (It's free.) I just tried a quick experiment... I didn't actually listening, but you can... I just wanted to confirm it was going to work.

1. Open a file in Audacity and then Effects -> Amplify with a negative value (attenuation) of -3 or -6dB to leave room for some added information without clipping

2. If it's not already at a high sample rate, go to Tracks and Resample to 96kHz. Save this file as your new "original".

3. You can optionally go to Analyze -> Plot Spectrum to see what you're starting with. (It will only analyze a short section so if the beginning is quiet or silent you may want to select some audio in the middle.)

3. Go to Effects -> EQ and Filters and apply a high-pass filter at around 10kHz to kill-reduce the low & mid frequencies.

4. Go to Effects -> Pitch and Tempo -> Change Pitch and change it by 200%. (That's 3X... A 100% change is double... 100% more than the original, not 100% of the original.) Nothing should be audible.

5. You may want to look at the spectrum again and it's a good idea to high-pass again to make sure there is nothing (or very little) below 20kHz. Note that the cutoff frequency of any filter, no matter how sharp the filter is, is defined as the -3dB point. So a 20kHz high pass filter is only down -3dB at 20kHz and you'll have to set the frequency above that if you want to "kill" everything 20kHz and below.

6. File -> Import the "original" file. (Don't use File -> Open this time.) The "original" file will show in the same Audacity Project below the one you've been playing with.

7. File -> Export will create a mix of the regular and ultrasonic information.

8. If you want to do a valid, scientific, listening comparison, do a Blind ABX Test. That's only necessary if you hear a difference or if you think you hear a difference.

 

[krabapple]

I plan to do compare sample-rates blind later (when i have better speakers) but i am already looking for good test tracks.

I want to see how high in frequency i can detect a lowpass filter once i have good equipment.

I am especially interested in testing if ultrasonics above say 24 Khz can make a difference (when there is basically zero chance of me hearing them directly).

the heroes of science who tried this before and gave us the 'hypersonic effect', used hi rez recordings of Javanese gamelan music. Lots of bells and chimes and tintinabulation. You'll also need supertweeters

 

[vintologi]

What's "meaningful"?


Not unless there is intermodulation distortion. Or, if you're an "audiophile" who "hears things."

Ruling out IMD in the speaker/amp being the reason for someone passing a blind-text is a bit messy, you could use a separate speaker/amp just for the ultrasonics but that will result in nasty interference so you should ideally have your ears positioned accurately.

But logically you would not expect IMD from ultrasonics in typical music to be audible when it comes to decent speakers and amplifiers, not sure about the human body itself.

Another possibility is some psychoacoustic effect resulting in high frequency sounds being more noticeable when mixed with lower frequencies. That's relevant for people with very good hearing.

OK... Do you have Audacity? (It's free.) I just tried a quick experiment... I didn't actually listening, but you can... I just wanted to confirm it was going to work.

1. Open a file in Audacity and then Effects -> Amplify with a negative value (attenuation) of -3 or -6dB to leave room for some added information without clipping

2. If it's not already at a high sample rate, go to Tracks and Resample to 96kHz. Save this file as your new "original".

3. You can optionally go to Analyze -> Plot Spectrum to see what you're starting with. (It will only analyze a short section so if the beginning is quiet or silent you may want to select some audio in the middle.)

3. Go to Effects -> EQ and Filters and apply a high-pass filter at around 10kHz to kill-reduce the low & mid frequencies.

4. Go to Effects -> Pitch and Tempo -> Change Pitch and change it by 200%. (That's 3X... A 100% change is double... 100% more than the original, not 100% of the original.) Nothing should be audible.

5. You may want to look at the spectrum again and it's a good idea to high-pass again to make sure there is nothing (or very little) below 20kHz. Note that the cutoff frequency of any filter, no matter how sharp the filter is, is defined as the -3dB point. So a 20kHz high pass filter is only down -3dB at 20kHz and you'll have to set the frequency above that if you want to "kill" everything 20kHz and below.

6. File -> Import the "original" file. (Don't use File -> Open this time.) The "original" file will show in the same Audacity Project below the one you've been playing with.

7. File -> Export will create a mix of the regular and ultrasonic information.

8. If you want to do a valid, scientific, listening comparison, do a Blind ABX Test. That's only necessary if you hear a difference or if you think you hear a difference.

Oh that's an interesting approach, you can speed up an existing track to get ultrasonic content that way.

 

[vintologi]

the heroes of science who tried this before and gave us the 'hypersonic effect', used hi rez recordings of Javanese gamelan music. Lots of bells and chimes and tintinabulation. You'll also need supertweeters

Ill look into that. Ill try to find some of that on RED.

 

[Blumlein 88]

Another thing you can do is take a track with high frequencies, and filter out everything below 20 khz, high pass filter. Then use Audacity or similar to slow it down to half speed. The range between 20-khz and 40 khz now becomes 10 khz to 20 khz. You'll be surprised how little is there. If you cannot hardly hear it like this what chance you'll hear it masked by all the content below 20 khz?

I get wanting to prove things to yourself. I've done it too. I can tell you it is a complete waste of time.

 

[vintologi]

Another thing you can do is take a track with high frequencies, and filter out everything below 20 khz, high pass filter. Then use Audacity or similar to slow it down to half speed. The range between 20-khz and 40 khz now becomes 10 khz to 20 khz. You'll be surprised how little is there. If you cannot hardly hear it like this what chance you'll hear it masked by all the content below 20 khz?

I get wanting to prove things to yourself. I've done it too. I can tell you it is a complete waste of time.

I atually found a track where the difference seemed to be noticeable when i tested it.

MediaFire

MediaFire is a simple to use free service that lets you put all your photos, documents, music, and video in a single place so you can access them anywhere and share them everywhere.

www.mediafire.com

(i preferred the 44.1 version 30 of 43 times).

I could just retest that again once i have good much better speakers to see if the difference disappeared (due to less IMD from tweeters or my hearing becoming worse).

I need a better amplifier too.

 

[krabapple]

Ill look into that. Ill try to find some of that on RED.

It may not have come through in my post but...I think you'll be wasting your time.

 

[danadam]

I atually found a track where the difference seemed to be noticeable when i tested it.

I don't know what you mean exactly by "meaningful ultrasonic content", but I don't see any in this track:

 

[CapMan]

I plan to do compare sample-rates blind later (when i have better speakers) but i am already looking for good test tracks.

I want to see how high in frequency i can detect a lowpass filter once i have good equipment.

I am especially interested in testing if ultrasonics above say 24 Khz can make a difference (when there is basically zero chance of me hearing them directly).

Sorry, but why ??

Are you a bat