Re: 60 Hz hum causing problems with speech recognition (Brian Gygi )


Subject: Re: 60 Hz hum causing problems with speech recognition
From:    Brian Gygi  <bgygi@xxxxxxxx>
Date:    Tue, 16 Jun 2009 18:59:24 +0000
List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>

----=_vm_0011_W7873316111_6526_1245178764 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable It figures, the Hz don't go as far in Europe as they do in the US. Like t= he dollar... -----Original Message----- From: Piotr Majdak [mailto:piotr@xxxxxxxx Sent: Tuesday, June 16, 2009 11:38 AM To: AUDITORY@xxxxxxxx Subject: Re: 60 Hz hum causing problems with speech recognition In Europe, we call it "the 50-Hz hum" :-)Tarun, I also vote for a ground = loop. You could look if you have more than one device grounded. If you ha= ve no grounded devices try to ground the mic-preamp. Usually this helps i= n our lab.Good luck,PiotrPierre Divenyi wrote:> Tarun, you have two optio= ns: (1) get rid of the ground loop that is > quite probably the culprit, = or (2) move to Europe where the 60-Hz hum > is unknown.>> -Pierre>>> On 6= /16/09 7:43 AM, "Tarun Pruthi" wrote:>> Hi all:>> I have been having a l= ot of problems lately with 60 Hz humming> noise in my speech recordings (= I will confess that most of my> tests have been on windows/linux laptops = - and my guess is that> they probably have much worse recording quality t= han desktops on> average - but we live in a mobile world now and I have t= o make it> work on laptops). Whenever it is there, it causes a serious dr= op> in speech recognition accuracy, especially since it comes in very> ra= ndomly at times giving for example clean recordings during> speaker enrol= lment and noisy recordings in live testing. The> problem is even worse on= linux which has given me all kinds of> artifacts in the recordings besid= es the 60 Hz hum - my guess is> that this is most likely because of drive= r problems.>> So, I have 3 questions:> 1. What is the source of this 60 H= z noise? Powerline? RF> interference? I have tried disconnecting the powe= r cord - it works> sometimes, but not always.> 2. Is there a standard way= to get rid of the 60 Hz hum and ensure> clean recordings for the purpose= s of speech recognition? The only> way I can think of is to use notch fil= ters at 60 Hz, and some of> the harmonics, say 120 Hz and 180 Hz, since m= ost of the times it> doesn't manifest as a single tone at 60 Hz, but has = harmonics all> over the frequency range.> 3. If this is a well-known prob= lem with sound cards, then why> aren't we, speech recognition community, = trying to come up with a> mandatory compliance standard for sound card ma= nufacterers which> ensures that the microphone input is not corrupted by = 60/50 Hz> noise? That should provide a big boost to speech recognition> a= ccuracies, and a wider acceptance of speech recognition.>> Tarun> Senior = Research Engineer> Think A Move, Ltd>> ----=_vm_0011_W7873316111_6526_1245178764 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <html>It figures, the Hz don't go as far in Europe as they do in the US.&= nbsp; Like the dollar...<br><div><font color=3D"#0000ff" face=3D"Verdana"= size=3D"2"></font>&nbsp;</div> <blockquote style=3D"border-left: 2px solid rgb(0, 0, 255); padding-left:= 5px; margin-left: 5px; margin-right: 0px;"><font face=3D"Tahoma" size=3D= "2">-----Original Message-----<br><b>From:</b> Piotr Majdak [mailto:piotr= @xxxxxxxx<br><b>Sent:</b> Tuesday, June 16, 2009 11:38 AM<br><b>To:</b= > AUDITORY@xxxxxxxx<br><b>Subject:</b> Re: 60 Hz hum causing probl= ems with speech recognition<br><br></font>In Europe, we call it "the 50-H= z hum" :-) Tarun, I also vote for a ground loop. You could look if you have more than one device grounded. If you have no grounded devices try to ground the mic-preamp. Usually this helps in our lab. Good luck, Piotr Pierre Divenyi wrote: &gt; Tarun, you have two options: (1) get rid of the ground loop that is = &gt; quite probably the culprit, or (2) move to Europe where the 60-Hz hu= m &gt; is unknown. &gt; &gt; -Pierre &gt; &gt; &gt; On 6/16/09 7:43 AM, "Tarun Pruthi" <t.pruthi@xxxxxxxx> wrote: &gt; &gt; Hi all: &gt; &gt; I have been having a lot of problems lately with 60 Hz humming &gt; noise in my speech recordings (I will confess that most of my &gt; tests have been on windows/linux laptops - and my guess is that &gt; they probably have much worse recording quality than desktops on= &gt; average - but we live in a mobile world now and I have to make i= t &gt; work on laptops). Whenever it is there, it causes a serious drop= &gt; in speech recognition accuracy, especially since it comes in ver= y &gt; randomly at times giving for example clean recordings during &gt; speaker enrollment and noisy recordings in live testing. The &gt; problem is even worse on linux which has given me all kinds of &gt; artifacts in the recordings besides the 60 Hz hum - my guess is &gt; that this is most likely because of driver problems. &gt; &gt; So, I have 3 questions: &gt; 1. What is the source of this 60 Hz noise? Powerline? RF &gt; interference? I have tried disconnecting the power cord - it wor= ks &gt; sometimes, but not always. &gt; 2. Is there a standard way to get rid of the 60 Hz hum and ensur= e &gt; clean recordings for the purposes of speech recognition? The onl= y &gt; way I can think of is to use notch filters at 60 Hz, and some of= &gt; the harmonics, say 120 Hz and 180 Hz, since most of the times it= &gt; doesn't manifest as a single tone at 60 Hz, but has harmonics al= l &gt; over the frequency range. &gt; 3. If this is a well-known problem with sound cards, then why &gt; aren't we, speech recognition community, trying to come up with = a &gt; mandatory compliance standard for sound card manufacterers which= &gt; ensures that the microphone input is not corrupted by 60/50 Hz &gt; noise? That should provide a big boost to speech recognition &gt; accuracies, and a wider acceptance of speech recognition. &gt; &gt; Tarun &gt; Senior Research Engineer &gt; Think A Move, Ltd &gt; &gt; </t.pruthi@xxxxxxxx></blockquote></html> ----=_vm_0011_W7873316111_6526_1245178764--


This message came from the mail archive
http://www.auditory.org/postings/2009/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University