Swept Sinus is no Sinus ?

[PeterSt]

Hi,

I just wondered whether it's known that the sweeps from sinuses are no real sinuses ?
They have lumps on them, which is very audible in the higher regions, where they contribute to the squarishness of the wave.

Also, do you realize that this creates "sub harmonics" ?

When playing the sweeps over a non-oversampling DAC it goes like WeeeeeWeeeeWeeeWeeWeW (an ongoing higher pitch sweep incurring for ongoing lower sweeps). And no, this is not related to the DAC, because you just can visualize it on the screen.

It doesn't seem right to me for test purposes ... (where coinciding "subharmonics" can reach the exact same dB level of the (higher) fundamental). Not with a nos-DAC which can make this audible to the microphone ...

Thanks,
Peter

[MACCA350]

This has me perplexed, are you talking about Sine Waves like those we use for testing frequency response? If so I don't see any sub harmonics within the tones themselves.
Below is a 10Hz tone, looks pretty clean from 0Hz-20KHz to me

If this is not what your talking about just ignore what I said

cheers

[PeterSt]

MACCA350 wrote:

This has me perplexed, are you talking about Sine Waves like those we use for testing frequency response?

Yes.

It's not so easy to reveil them. And keep in mind, I was talking about "SUB harmonics", not harmonics, so you'd have to start just higher up. Further more, when you just put up one sinus, chance is 99% you won't meet what's actually in there. So you'd real time (on screen) have to follow a sweep. Or just listen (nos DAC required !).

Please note that I'm using 3.29 (and use REW for a long time now; it never changed).


In order to instantly see what I mean, sweap a regular square wave (with some prog); it's the Nyquist (and 0Hz) mirror doing it.

Btw, I'm okay with it, but I just wondered whether this is known or maybe even on purpose (the bump feels to me like "catching" the backloop better). But since you can't hear this with oversampling DACs I thought to mention it ...

[JohnM]

The spectrum of the sine sweep does not have any "sub-harmonics" or other artefacts. There are 2 main sources of the features you describe: distortion due to overdriving the replay chain (your mention of "squarishness" hints of clipping occurring somewhere) and aliasing in a badly designed DAC (a DAC which does not oversample will require extremely tight filtering on its output if it is to correctly reproduce the input, without this filtering it will generate aliasing artefacts. Easing the filtering requirements was one of the drivers for the design of oversmapling DACs). Looking at the individual samples of a swept sine does not show you the true shape of the waveform unless the plotting function applies Nyquist filtering (as the DAC should do), without such filtering (e.g. with a simple join-the-dots representation) the displayed waveform will have an aliased appearance. Basically the same problem a DAC without over-sampling suffers from if it omits the (essential) output filters.

[PeterSt]

There is no need to debate this John, and it's just what I observe.

Possibly it is more complicated. I do understand the aliasing thing, but I do not understand where it comes from. I mean, I can see it on the screen and I can hear it. What I see on the screen cannot come from the DAC (which can't even talk back, in my case it's just an audio DAC). Thus, the DAC certainly could produce the sound (in theory anyway), but what I hear "comes from the screen". Okay, what's on the screen can come from improper filtering, but were that the case, I wouldn't here it.
There's only one tiny chance of how to explain things : the aliasing on the screen is coincidentally the same as produced by the DAC because both do not apply the proper filtering (which the DAC certainly does not, becase it's against the principles of it).

Whatever the cause is, it could (!) influence measuring. To what extend I don't know, and e.g. it could be that you only read back the expected frequencies, so the "subharmonics" just don't disturb.

[JohnM]

PeterSt wrote:

There's only one tiny chance of how to explain things : the aliasing on the screen is coincidentally the same as produced by the DAC because both do not apply the proper filtering (which the DAC certainly does not, becase it's against the principles of it).

That is correct, which is one of the things I was trying to say in my previous post. From an engineering and analytical standpoint that is simply an incorrect thing to do, the aliasing artefacts you hear are the result. It is worth bearing in mind that those same artefacts are present in everything you listen to on such a DAC, not just test tones.

[PeterSt]

Ok and agreed. Thank you.