Subject: Re: lopsided tones. From: David Smith <smithd@xxxxxxxx> Date: Wed, 19 Aug 2009 13:47:09 -0500 List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>This is a multi-part message in MIME format. --_----------=_1250707629159690 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="iso-8859-1" If there is a zero-referenced (non-differential) gain stage after the offset is added, the zero crossings will still have been moved relative to the peaks and some of the added harmonics will remain even after the offset is removed. A lot depends on the type of gain stage. "Flutter" sounds like you're generating subharmonics.=A0 Try raising the frequency of the whole signal (record a few seconds and change the sample rate) to get the subharmonics up into the range of the FFT processor. D Smith ----- Original Message ----- From: "Bob Masta" To: AUDITORY@xxxxxxxx Subject: Re: [AUDITORY] lopsided tones. Date: Wed, 19 Aug 2009 09:08:27 -0400 On 17 Aug 2009 at 15:41, Ranjit Randhawa wrote: > Dear List, > I have been experimenting with a 100hz tone, where the positive > half sinusoid of the period is larger than the negative, the > phase is however is not changed. Speech seems to have this > profile of larger positive pulses as compared to the negative, > hence my interest. Applying fft to such a signal, I get an > increase in magnitude for the 100hz component, and an increase in > the dc component. What I hear however is the basic 100hz tone, > and a flutter on top of it, not what fft seems to indicate. My > assumption was that the increased dc component would not be > heard, and I would hear an increase in loudness of the 100 hz. > However, the base 100hz loudness does not seem to change as I > increase the area under the positive sinusoid, but the flutter > does. Any history or explanation would be most welcome. > Thanks and regards, > Randy Randhawa Please go into more detail about your waveform and setup. If you have simply added DC to the baseline, as Matt mentions, you would see this in the FFT of the raw internal signal, but the DC would normally be blocked from reaching your headphones or speakers so would not be audible. (Not that it would likely have been audible anyway, since speakers and headphones won't go down to DC without exotic treatments, like perfect air-tight seals to your head.) If you had made the generated wave too positive, you would get clipping that would be audible as harmonics and easily visible in the spectrum. So I'm wondering if you are using some sort of tone burst or envelope system. If so, that might explain what you are finding, since the DC would be modulated by the envelope and would result in a component at the burst rate, which would probably be described as flutter. You might not see this flutter component in the FFT since it could be lost in the lowest FFT bin with the DC. The flutter should vanish if you get rid of the envelope and just leave the tone on. Best regards, Bob Masta D A Q A R T A Data AcQuisition And Real-Time Analysis www.daqarta.com Scope, Spectrum, Spectrogram, Signal Generator Science with your sound card! --=20 Be Yourself @xxxxxxxx mail.com! Choose From 200+ Email Addresses Get a Free Account at www.mail.com! --_----------=_1250707629159690 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset="iso-8859-1" <div> <br>If there is a zero-referenced (non-differential) gain stage after the o= ffset is added, <br>the zero crossings will still have been moved relative = to the peaks and some of the<br>added harmonics will remain even after the = offset is removed.<br>A lot depends on the type of gain stage.<br><br>"Flut= ter" sounds like you're generating subharmonics. Try raising the freq= uency <br>of the whole signal (record a few seconds and change the sample r= ate) to get the <br>subharmonics up into the range of the FFT processor.<br= >D Smith<br> <br> <blockquote style=3D"border-left: 2px solid rgb(16, 16, 255); margin-left: = 5px; padding-left: 5px;">----- Original Message -----<br> From: "Bob Masta" <audio@xxxxxxxx><br> To: AUDITORY@xxxxxxxx<br> Subject: Re: [AUDITORY] lopsided tones.<br> Date: Wed, 19 Aug 2009 09:08:27 -0400<br> <br> <br> On 17 Aug 2009 at 15:41, Ranjit Randhawa wrote:<br> <br> > Dear List,<br> > I have been experimenting with a 100hz tone, where the positive <br> > half sinusoid of the period is larger than the negative, the <br> > phase is however is not changed. Speech seems to have this <br> > profile of larger positive pulses as compared to the negative, <br> > hence my interest. Applying fft to such a signal, I get an <br> > increase in magnitude for the 100hz component, and an increase in <br> > the dc component. What I hear however is the basic 100hz tone, <br> > and a flutter on top of it, not what fft seems to indicate. My <br> > assumption was that the increased dc component would not be <br> > heard, and I would hear an increase in loudness of the 100 hz. <br> > However, the base 100hz loudness does not seem to change as I <br> > increase the area under the positive sinusoid, but the flutter <br> > does. Any history or explanation would be most welcome.<br> > Thanks and regards,<br> > Randy Randhawa<br> <br> Please go into more detail about your waveform and setup.<br> If you have simply added DC to the baseline, as Matt<br> mentions, you would see this in the FFT of the raw internal<br> signal, but the DC would normally be blocked from reaching<br> your headphones or speakers so would not be audible. (Not<br> that it would likely have been audible anyway, since<br> speakers and headphones won't go down to DC without exotic<br> treatments, like perfect air-tight seals to your head.)<br> <br> If you had made the generated wave too positive, you would<br> get clipping that would be audible as harmonics and easily<br> visible in the spectrum.<br> <br> So I'm wondering if you are using some sort of tone burst<br> or envelope system. If so, that might explain what you are<br> finding, since the DC would be modulated by the envelope<br> and would result in a component at the burst rate, which<br> would probably be described as flutter. You might not see<br> this flutter component in the FFT since it could be lost in<br> the lowest FFT bin with the DC. The flutter should vanish<br> if you get rid of the envelope and just leave the tone on.<br> <br> Best regards,<br> <br> Bob Masta<br> <br> D A Q A R T A<br> Data AcQuisition And Real-Time Analysis<br> www.daqarta.com<br> Scope, Spectrum, Spectrogram, Signal Generator<br> Science with your sound card!<br> </audio@xxxxxxxx></blockquote> </div> <BR> --=20 <div>Be Yourself @xxxxxxxx mail.com <br> Choose From 200+ Email Addresses<br> Get a <b>Free</b> Account at <a href=3D"http://www.mail.com" target=3D"_bla= nk">www.mail.com</a>!</div> --_----------=_1250707629159690--