Hi,

I'm trying to get a steady output of A=440 c.p.s. from my speakers to calibrate a device. The only command I'm aware of in Linux is 'speaker-test'. I've been using the command:

speaker-test -f 440 -t sine

When the tone initiates, it is 0.08 cents low, then after a moment comes up to exactly A440, then there is a quick 'pop' or 'tick', and the the tone resets itself and is again beginning at -0.08 cents, then come up again to being correct to within 0.00 cts of A440.

I think that if the tone would be allowed to just continue without apparently resetting itself, I could get a stable A440 output.

Is there a way I could have the tone go on without resetting, or, is there another way to get this steady output from my speakers?

Also, I've also tried:


speaker-test -f 440.5 -t sine
speaker-test -f 440.1 -t sine

and so on without any variation in the output. However,


speaker-test -f 441 -t sine

definitely does give me a different output.

Does anyone know if 'speaker-test' will allow for cps to be output with accuracy beyond the other side of the decimal? ( -f 440.00 )

Thanks,
-Joe

Btw, the device I'm measuring with does not have the ability to output sound.

I don't know if it's any use to you but Audacity has a tone generator built in. There's a Slackbuild here. The package has a lot more besides the tone generator and if you are doing audio stuff is definitely worth looking at.

Thank you bgeddy!

I installed Audacity as you suggested. In order to get an exact A=440 tone, I have Audacity set at 440.376Hz. This gives the output of A=440 with a fluctuation of +/- 0.01 cents.

Thanks for the advice!

-Joe

You could precalculate a pcm file with the tone and then just play it

I read somewhere that PC hardware audio playback accuracy is poor, especially when poor quality interpolation is used to transcode from 44.1 KHz .wav to the 48 KHz or 96 KHz hardware operation of most soundcard hardware

Sox is the best tool for precision conversion

David in KL

Thank you David!

I will investigate sox.

Thanks,
-Joe

I get a strange warning on one of my Ubuntu machines:

pulseaudio[3271]: alsa-util.c: Device front:0 doesn't support 44100 Hz, changed to 44099 Hz

44100 Hz is a bodge on most modern hardware, dvd players are also fundamentally nx48 KHz machines and are known to not play audio cds particularly well

If you want high precision you should experiment with native 48 KHz encoding
I don't have perfect pitch and would never hear 0.1 cents off

I was able to do what I wanted to do with Audacity.

David, you could probably hear it with a little practice and a reference pitch sounding at the same time. The higher the frequency, the more Hz per cent. It would be easier to hear a heterodyne or beat in the higher range.

In my case, I used an extremely accurate device to measure and tweak the output of the speakers. I then used the speaker output to calibrate another device +/- 0.01 cents (or 1/10,000th of a half-step) while still verifying the output with the first device.

Hi joegumbo - just out of interest - what is it your doing with all this ? (Being very nosey I know). Tell me to naff off if you don't feel like saying

Sure.... No secrets here.

I've been tuning and repairing pianos for a living. I was originally taught by ear just over 30 years ago, but over the last 10 years, I've been using a device called a Verituner <www.veritune.com> along with my ear. We check each other. (The VT is a dedicated tuning device.)

Lately, I decided that I wanted to test drive the trial version of TuneLab software <www.tunelab-world.com>, but it requires calibration.... because of issues relating to soundcards.

The thing that interested me in TuneLab, is that it will allow continuing the 4:2 octaves into the bass the whole way down to A0. (As far as I can tell from the display, the VT measures only partials 6 through 12 at A0.) Most pianos prefer at least 6:3 in the bass, but a few really shine with 4:2 in the entire bass (plus tweaking)... IMO.