Digital Sine Wave Generator

[arvind]

Hi Guys,

Could anyone help me with a link to a website from where I can download a programme to run digital sine waves of any frequency through the NOS 1/amps/speakers.

Best regards,

Arvind

[PeterSt]

Hi Arvind,

Answer is easy or difficult, depending on what you want to do with it. So, please be exact on the purpose you need it for.

Peter

[CoenP]

If you are in need of simple testsignals, sweeps, noises and the like take a look at ARTA.

Regards, Coen

[arvind]

Hi Arvind,

Answer is easy or difficult, depending on what you want to do with it. So, please be exact on the purpose you need it for.

Peter

Hi Peter,

I need to test the room for frequencies which create standing waves.

Best regards,

Arvind

[PeterSt]

Arvind, then I assume what Coen mentioned will work fine.

Remember the conveniency of A-B playback for this ...
(buttons at the bottom)

Regards and good luck,
Peter

[arvind]

Hi Peter/Coen,

Thanks for the link, however would the NOS 1 soundcard be supported by the ARTA software?

Best regards,

Arvind

[PeterSt]

Arvind,

Hoping that you talk about D/A (and not A/D to capture the measured result (by microphone)), ... why not.

All you need to do is generate those signals (a length of 2 minutes really is enough) and play that through XXHighEnd and the NOS1 just like you play everything else.

Regards,
Peter

[CoenP]

Switch the nos mode swith to normal dac mode first. Likely you will have to set the nosdriver to a mode that will support the testsignal.

Afaik arta supports higher bitrates en bitlength.

If this all too cumbersome: I used my mobo soundcard and a behringer mic without any problems.

Regards, Coen

[PeterSt]

Hi Coen,

With a thank you for the help, I now don't understand what you are getting at. No any switches or driver modes etc. need to be changed.

Or do you think a test signal is something else than music ? (apart from that it is no music )

Regards,
Peter

[CoenP]

No, not really. Only a testsignal is very boring!

Since the NOS1 is NOS, you will have to use/create a testsignal at a samplefrequency as high as possible to avoid sampling noise contamination of your testsignal.

The regular SDM motherboardstuff delivers a nice testsignal with 44.1k input becaus of the on-chip digitalfilters. Since you don't listen to it for enjoyment (if ever possible ) the mobo chips are fine signal generators for general purpose analysis (not the -90dB, 0,0x% distortion ones).

If you intend to measure the effect of the NOS1/XXHE on standing waves, obviously you will need testsignal files and play with arc-predict.

Regards, Coen

[PeterSt]

Hey Coen,

I am sorry, but you make a veeeery odd story of this all ...

1.
If you think you should use testsignals with "other" quality than XXHighEnd/NOS1 produces normally, you better listen to another combination in the future.

2.
How do you think I test myself ? through testsignals which don't represent (your) reality ? Up to playing them through another sound device ?

3.
Since this is about (Arvind's) standing waves, and since it should be well known that the best to eliminate that is XXHighEnd/NOS1, we should be using something else to measure through in the first place ? So :

If you intend to measure the effect of the NOS1/XXHE on standing waves, obviously you will need testsignal files and play with arc-predict.

*That* makes completely sense. It's only that you should replace that "If" into "Since". Or better : "Since you will be tuning your system for XXHighEnd/NOS1 usage, you should be using a test signal passing the same chain, and representing what you normally play.".

So, just generate a 16/44.1 signal assumed this is about Redbook CD we're after, and test that for its merits through XXHighEnd/NOS1.

I hope you will agree with this ...
Peter

[CoenP]

1.
If you think you should use testsignals with "other" quality than XXHighEnd/NOS1 produces normally, you better listen to another combination in the future.

2.
How do you think I test myself ? through testsignals which don't represent (your) reality ? Up to playing them through another sound device ?

Horses for courses of course. With all measurements you've got to know what you are doing, all generators and measuring devices are able to provide some usefull information. You don't need a sophisticated setup to detect 1dB anomalies or 0,x% distortion of a loudspeaker or get qualitative information about the setup. You can match capacitors even with a capacitace meter that is 5% off, just don't trust their absolute value.
I've been able to minimise the distortion of my compressiondrivers by adjusting the lockingscrews of the diaphragm housing (!) while looking at the fft of a testsignal on my sound card. This saved a lot of time, otherwise I would have been running from and to my listening position to assess the sound quality, likely to end up somewhere unconclusive. The effect was a 6dB lowering of dominant distortions.

I am shurely not claiming that you can use such an unrepresentative measurement setup for very low level effects. Especially when you are going to use the most sensitive measuring device: your ears.

So I disagree that such quick and dirty soundcardmeasurements have no place in audio, you've  just got to know when they stop giving you usefull information. Assessing a high end setup on anomalies by ear is not one of them.

3.
Since this is about (Arvind's) standing waves, and since it should be well known that the best to eliminate that is XXHighEnd/NOS1, we should be using something else to measure through in the first place ? So :

If you intend to measure the effect of the NOS1/XXHE on standing waves, obviously you will need testsignal files and play with arc-predict.

*That* makes completely sense. It's only that you should replace that "If" into "Since". Or better : "Since you will be tuning your system for XXHighEnd/NOS1 usage, you should be using a test signal passing the same chain, and representing what you normally play.".

So, just generate a 16/44.1 signal assumed this is about Redbook CD we're after, and test that for its merits through XXHighEnd/NOS1.

I hope you will agree with this ...
Peter

I completely agree!
Make shure you testsignal is of very high quality so you are not judging the testsignal  . If anything comes out of this, it would be interesting to know what difference a file with a different sample rate or wordlength would make...

Regards, Coen

[PeterSt]

Hi Coen,

I still think you don't get what this is about ... or oughts to be about ...

Look at the title of this topic. Look at the question :

Could anyone help me with a link to a website from where I can download a programme to run digital sine waves of any frequency through the NOS 1/amps/speakers.

Admitted, this can be interpreted as some program which generates a signal and outputs it to the selected device (NOS1 in this case). But generatiing a signal through a random program is a moot thing, once we know what XXHighEnd does to the sound. Or maybe look at it like this :

We have readily available test signals. In WAV files that is (or FLAC if you want per se). Now we are going to LISTEN to the result on room behavior. I say "listen" on purpose, because it is about that. So, nobody asks for impulse response or impedance behavior or everything we would not be able to listen to anyway (but measure through microphones, yes). Standing waves we CAN observe easily. The most easily in fact. Now ... we can choose between Foobar and XXHighEnd to play the files through;
What should we choose ?

If your answer is "doesn't matter" then something didn't come through quite. So ...

I completely agree!

So admitted that. But :

Horses for courses of course.

No. Nobody asks for this (as far as I am aware of) and when I now look into your Arta (which I just did), it is not suitable for the job at all. So, it merely seems that you came up with Arta and now try to justify it (with the first part of your last post). This is nothing Arvind would be able to use. And you know what ? I am not even able to use it for the purpose intended. But that is me and maybe I have a bad day of understanding things.

So ...
For the last two hours I have been looking for a good signal generator (which can save the generated signal to a WAV file), and you know what ? they ALL s*ck. Freakin' unbelievable. I knew this already, but I never looked that long. And the real problem ? Well, I once had one before a PC died, generated all my testsignals from it (which I still use), so at least one exists. But I can't recall or find which one it is.

In the end Arvind's question is not a bad one at all. So, if people may have an idea or even a hunch about a good program, please speak up.I can test it here.

Look below. All 1KHz generated signals. The first screenshot is what I use. The second one is AudaCity. It isn't even capable of generating 1000Hz when I ask for it. But look how bad it is. The third one is NHC - one of the most fameous and also annoying programs because it installs all kinds of stuff on your PC you don't ask for (and all not free). But look at it ...

And so I tried all I could find. The one worse than the other. It's almost like I have to make it myself (a program for it).


PS: And looking at all the variations I found, I now even wonder whether what I use is the best out there. Why would it ? It's just one of the random programs I found back at the time. I recall though that the author of it said that generating good test signals is not an easy job at all. I appreciated that when I found it and used NHC (free at the time) till then (again, compare). Today I should appreciate it even more.

[PeterSt]

PPS: Also look at the picture below. A phase difference between the channels of a wopping 8 degrees. This should be 0.00(0000).
Quite important when looking for standing waves ...

[PeterSt]

Ha, found it ! I couldn't stand that I had lost it.
Examined 100s and 100s of folders on various PCs and this is the one :

http://www.tropicalcoder.com/AudioTestFileGen.htm

Just select 16 bits for reality (CD Redbook playback assumed as your daily business) and a sample rate of 44100. Leave the 3dBFS attenuation default be.
Don't forget the .wav extension in the box for the Output File.

Play back through XXHighEnd how you normally play music.

Use slow sweeps (like 2 minutes) to discover problematic areas, and notice that the speed of the sweep is linear to the frequency. So, when you'd generate a 2 minute sweep which spans 0-22000Hz, after 1 minute of playback the ferquency will be 11000Hz. When you found a problematic area (it buzzes in the room), e.g. at around 7000 Hz, you can generate another slow sweep from 6500Hz till 7500Hz and narrow down the frequency of the problem.

When all is narrowed down to a specific range, you can or let that continuously play through A-B with a more fast sweep (like 10 seconds) or generate a fixed frequency and let that A-B all the time.


Peter