audacity-users

 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: David Walsh - 2004-08-30 23:34:21 ```Gale wants a graphic equalizer, but David is building a parametric = equalizer, a different thing.=20 A 3 band equalizer woul be great. Centres at 100, 1k, 10k. David Walsh -----Original Message----- From: David R. Sky [mailto:sky@...] Sent: Tuesday, 31 August 2004 5:11 AM To: audacity-users@... Subject: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? Gale: I think to be useful there would need to be 10 bands at least (ie = similar to most equalisers in players). You could make equal steps = between the limits of human hearing for the Kh and perhaps +12dB to = -12dB gain? I am afraid I have not understood the purpose of the octave switch, = other than if you choose 2 octaves the effect is greater than with 1 = octave. I do not understand because diffferent frequencies naturally = will tend to be concentrated in different pitches, so if you boost lower = frequencies you will naturally be boosting lower pitches? David: I'll use an example to illustrate the width control. If the = center frequency you choose is 1000 Hz and the width is 2 octaves, the = band of frequencies affected will be between 1 octave below to 1 octave = above 1000 Hz. Each octave of frequency is based on a factor of 2, so 1 = octave below 1000 Hz is 500 Hz (1000 divided by 2), 1 octave above 1000 = Hz is 2000 Hz (1000 _times) 2). So if you choose center frequency of = 1000 Hz with 2 octave width, the effect will boost or dampen frequencies = between 500 and 2000 Hz. And of course, if you apply the effect with = those settings a second time, you will have additional boosting or = dampening of those same frequencies. Human hearing ranges generally from between 20 and 20,000 Hz (20 khz), = approximately 10 octaves. So your request for a ten-band equalizer makes = it sound like each ban covers an octave range. I haven't used equipment = like that before (maximum five bands on my ghetto blaster), so thanks = for the info. If there are frequency numbers listed on your equalizer, could you send = those, or at least the first (lowest) frequency. Does that help? David --=20 ------------------------------------------------------- This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools! Get your free copy of BEA WebLogic Workshop 8.1 today. http://ads.osdn.com/?ad_id=3D5047&alloc_id=3D10808&op=3Dclick --=20 Mailing list: Audacity-users@... To UNSUBSCRIBE, use the form at the bottom of this web page: https://lists.sourceforge.net/lists/listinfo/audacity-users ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: David Walsh - 2004-08-31 23:36:43 Attachments: Message as HTML ```=20 Try this for a definition: =20 ; = http://www.harmony-central.com/Effects/Articles/Equalization/ =20 DW =20 =20 -----Original Message----- From: mail@... [mailto:mail@...] Sent: Tuesday, 31 August 2004 9:03 PM To: audacity-users@... Subject: Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: = what do you want to see in it? =20 >>>>Gale wants a graphic equalizer, but David is building a parametric = equalizer, a different thing.=20 =20 I am afraid I am a bit ignorant here as I am only familiar with graphic = equalisers in player apps...... and thought each knob thereon applied = positive or negative gain only at the stated frequency not that it was = centred at that frequency with a range above and below. Could we have an = explanation of the difference between a graphic and parametric = equaliser? How would you define Audacity's native FFT filter which as = you know I use a lot but wish it could be made to stay enlarged.=20 ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: SMITH Linton - 2004-09-01 00:03:17 ```Guys, (puts electronic engineeer hat on ;) A little filter theory for you... Gain: a multiplication factor. Gain > 1 =3D amplification. Gain < 1 =3D attenuation. Gain =3D 0 =3D no signal gets though at all. Graphic Equalizers and Parametric Equalizers (also known as Parametric Amplifiers) are two forms of the same beast. What I mean is that they allow the user to change the level of a signal at a particular frequency. All of these devices involve an amplification stage that has the ability to affect selected frequency bands in the input signal. A graphic equalizer uses many amplifiers in parallel and each is designed to affect the signal at a specific frequency (centre frequency) and with a specific bandwidth (ie band of frequencies around the centre frequency). Think of a line of overlapping hills or valleys where each hill or valley is a centre frequency and the width of the hills/valleys is the bandwidth. Now if you overlapp them closely enough and set the gain to unity (ie no gain - no attenuation) you get a flat frequency response. Basically the same as a plain piece of wire ;) Bass and Treble tone controls on a stereo is nothing more than a two band Grahic EQ. On the other hand a parametric amplifier at its simplest is the equivalent of a single band amplifier from the graphic equalizer. The difference being that it has controls that allow the user to set the centre frequency, the bandwidth, and the amount of gain or attentuation of the centre frequency and band frequencies. An additional difference that may occur is that the Graphic Equalizer amplifiers are configured as BANDPASS filters. ie anyhting outside the pass band is attenuated. Where as a Parametric Amplify is commonly, though not definitely, a LOWPASS filter where the frequecies below the centre frequency a amplified to a fixed level, the frequencies above the centre frequencies are attentuated at a fixed rate (roll off) eg single pole low pass rolloff =3D 6 db / octave. At the centre frequency band the signal is attentuated or amplified by the gain setting. The single tone control on an old style tv or radio is just a parametric eq with the centre frequency being the only control, and the band gain set to unity. I set out to provide some clarity on Grahic EQ vs Parametric EQ, I hope this all helps but I am afraid I may have muddied the water more. cheers Linton -----Original Message----- From: audacity-users-admin@... [mailto:audacity-users-admin@...]On Behalf Of David R. Sky Sent: Wednesday, 1 September 2004 1:16 AM To: audacity-users@... Subject: Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? On Tue, 31 Aug 2004 mail@... wrote: > > >>>>Gale wants a graphic equalizer, but David is building a parametric equalizer, a different thing. > > > I am afraid I am a bit ignorant here as I am only familiar with graphic equalisers in player apps...... and thought each knob thereon applied positive or negative gain only at the stated frequency not that it was centred at that frequency with a range above and below. Could we have an explanation of the difference between a graphic and parametric equaliser? How would you define Audacity's native FFT filter which as you know I use a lot but wish it could be made to stay enlarged. > I don't know the difference either. But hardware graphic equalizer knobs do boost or attenuate (increase or decrease in volume) a certain frequency range. Just like the bass and treble knobs on simple playback devices (those ones would affect a larger band of frequencies than a graphic equalizer with many more knobs) (the number of freqneicies that can be changes for each band tends to be larger the fewer the knobs there are). Also, hardware equalizers would alter the volume of a range of frequencies, and _only_ for those frequency ranges. The neat thing about software (like these plug-ins we've been talking about) is that many more things can be changed. Like with the EQ plug-in I sent a couple days ago. Not only does it amplify or attentuate a frequency range, the bandwidth (how wide the band is, or how many frequencies each band will change) is changeable, as well as the actual center frequency of that range of frequencies. I doubt that you could do these last two alterations using hardware EQ units. David ------------------------------------------------------- This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools! Get your free copy of BEA WebLogic Workshop 8.1 today. http://ads.osdn.com/?ad_id=3D5047&alloc_id=3D10808&op=3Dclick --=20 Mailing list: Audacity-users@... To UNSUBSCRIBE, use the form at the bottom of this web page: https://lists.sourceforge.net/lists/listinfo/audacity-users ```
 [Audacity-users] re: graphic vs. parametric EQ From: David R. Sky - 2004-09-01 00:55:54 ```thanks Linton for the explanation and elucidation of boost and attenuation with graphic vs parametric equalization! David On Wed, 1 Sep 2004, SMITH Linton wrote: > Guys, (puts electronic engineeer hat on ;) A little filter theory for > you... > > Gain: a multiplication factor. Gain > 1 = amplification. Gain < 1 = > attenuation. Gain = 0 = no signal gets though at all. > > Graphic Equalizers and Parametric Equalizers (also known as Parametric > Amplifiers) are two forms of the same beast. What I mean is that they > allow the user to change the level of a signal at a particular > frequency. > > All of these devices involve an amplification stage that has the ability > to affect selected frequency bands in the input signal. > > A graphic equalizer uses many amplifiers in parallel and each is > designed to affect the signal at a specific frequency (centre frequency) > and with a specific bandwidth (ie band of frequencies around the centre > frequency). Think of a line of overlapping hills or valleys where each > hill or valley is a centre frequency and the width of the hills/valleys > is the bandwidth. Now if you overlapp them closely enough and set the > gain to unity (ie no gain - no attenuation) you get a flat frequency > response. Basically the same as a plain piece of wire ;) > > Bass and Treble tone controls on a stereo is nothing more than a two > band Grahic EQ. > > On the other hand a parametric amplifier at its simplest is the > equivalent of a single band amplifier from the graphic equalizer. The > difference being that it has controls that allow the user to set the > centre frequency, the bandwidth, and the amount of gain or attentuation > of the centre frequency and band frequencies. > > An additional difference that may occur is that the Graphic Equalizer > amplifiers are configured as BANDPASS filters. ie anyhting outside the > pass band is attenuated. Where as a Parametric Amplify is commonly, > though not definitely, a LOWPASS filter where the frequecies below the > centre frequency a amplified to a fixed level, the frequencies above the > centre frequencies are attentuated at a fixed rate (roll off) eg single > pole low pass rolloff = 6 db / octave. At the centre frequency band the > signal is attentuated or amplified by the gain setting. > > The single tone control on an old style tv or radio is just a parametric > eq with the centre frequency being the only control, and the band gain > set to unity. > > I set out to provide some clarity on Grahic EQ vs Parametric EQ, I hope > this all helps but I am afraid I may have muddied the water more. > > cheers > > Linton > > -----Original Message----- > From: audacity-users-admin@... > [mailto:audacity-users-admin@...]On Behalf Of David R. > Sky > Sent: Wednesday, 1 September 2004 1:16 AM > To: audacity-users@... > Subject: Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: > what do you want to see in it? > > > On Tue, 31 Aug 2004 mail@... wrote: > > > > > >>>>Gale wants a graphic equalizer, but David is building a parametric > equalizer, a different thing. > > > > > > I am afraid I am a bit ignorant here as I am only familiar with > graphic equalisers in player apps...... and thought each knob thereon > applied positive or negative gain only at the stated frequency not that > it was centred at that frequency with a range above and below. Could we > have an explanation of the difference between a graphic and parametric > equaliser? How would you define Audacity's native FFT filter which as > you know I use a lot but wish it could be made to stay enlarged. > > > I don't know the difference either. But hardware graphic equalizer knobs > do > boost or attenuate (increase or decrease in volume) a certain frequency > range. Just like the bass and treble knobs on simple playback devices > (those > ones would affect a larger band of frequencies than a graphic equalizer > with > many more knobs) (the number of freqneicies that can be changes for each > band tends to be larger the fewer the knobs there are). > > Also, hardware equalizers would alter the volume of a range of > frequencies, > and _only_ for those frequency ranges. The neat thing about software > (like > these plug-ins we've been talking about) is that many more things can be > changed. Like with the EQ plug-in I sent a couple days ago. Not only > does it > amplify or attentuate a frequency range, the bandwidth (how wide the > band > is, or how many frequencies each band will change) is changeable, as > well as > the actual center frequency of that range of frequencies. I doubt that > you > could do these last two alterations using hardware EQ units. > > David > > > > ------------------------------------------------------- > This SF.Net email is sponsored by BEA Weblogic Workshop > FREE Java Enterprise J2EE developer tools! > Get your free copy of BEA WebLogic Workshop 8.1 today. > http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click > -- ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: SMITH Linton - 2004-09-01 00:08:40 Attachments: Message as HTML ```See my other email for a "discussion" of graphic eq and parametric eq. =20 a FFT is a Fast Fourier Transform. FFT is a mathematical method of determi= ning the frequency content of a block of signal. ie it converts a signal r= epresented in the time domain (ie signal vs time) into a representation in = the frequency domain ((ie sginal vs frequency) From what I can gather from = my use of the FFT Filter, you set the frequency response that you want on a= gain vs frequency graph and by using the inverse FFT calculation, a time = domain filter is produced which is used to filter the signal. =20 HTH =20 Linton -----Original Message----- From: audacity-users-admin@... [mailto:audacity-users-adm= in@...]On Behalf Of mail@... Sent: Tuesday, 31 August 2004 8:33 PM To: audacity-users@... Subject: Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what = do you want to see in it? =20 >>>>Gale wants a graphic equalizer, but David is building a parametric equa= lizer, a different thing.=20 =20 I am afraid I am a bit ignorant here as I am only familiar with graphic equ= alisers in player apps...... and thought each knob thereon applied positiv= e or negative gain only at the stated frequency not that it was centred at = that frequency with a range above and below. Could we have an explanation o= f the difference between a graphic and parametric equaliser? How would you = define Audacity's native FFT filter which as you know I use a lot but wish = it could be made to stay enlarged.=20 ```
 Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: - 2004-09-01 00:21:09 ``` On Wed, Sep 01, 2004 at 10:02:21AM +1000, SMITH Linton wrote: > See my other email for a "discussion" of graphic eq and parametric eq. > > a FFT is a Fast Fourier Transform. FFT is a mathematical method of > determining the frequency content of a block of signal. ie it > converts a signal represented in the time domain (ie signal vs time) > into a representation in the frequency domain ((ie sginal vs > frequency) From what I can gather from my use of the FFT Filter, you > set the frequency response that you want on a gain vs frequency > graph and by using the inverse FFT calculation, a time domain filter > is produced which is used to filter the signal. Yes; the trick in using an FFT on a time domain signal is that a windowed FFT is a circular function whereas the time block is not. Windowing and padding tricks are needed to end up with a time-domain filter that isn't inadvertantly mixing data from the beginnings and ends of blocks. You also want the impulse response to be real and smooth in addition to noncircular. It's all in the details :-) Monty ```
 Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: - 2004-09-01 11:23:35 Attachments: Message as HTML ```>>>Linton.SMITH@...: a FFT is a Fast Fourier Transform. FFT is a mathematical method of determining the frequency content of a block of signal. ie it converts a signal represented in the time domain (ie signal vs time) into a representation in the frequency domain ((ie sginal vs frequency) From what I can gather from my use of the FFT Filter, you set the frequency response that you want on a gain vs frequency graph and by using the inverse FFT calculation, a time domain filter is produced which is used to filter the signal. If you take the flat line starting point with this tool and then make a control point , then I assume along this discontinuity you are in effect having the equivalent of many sets of knobs on a hardware equaliser with the knob set at greatest deviation being at the control point? I may be obtuse but if you look say at the control point, is this effect applied at the whole range of hearing or only at the width of some predefined range of octaves. In other words is this control point centred around a frequency as per previous discussions? ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: Jeff and Linda Klein - 2004-09-01 10:31:29 ``` 1 = amplification. Gain < 1 =
 Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: Bob Long - 2004-09-01 12:01:47 ```On Wednesday, September 01, 2004 8:27 PM [GMT+1=CET], Jeff and Linda Klein wrote: > 1 = amplification. Gain < 1 = > > I think you mean to say Gain = 0 = no signal level change at all. I would have thought that a gain of 1 (unity) is no change to the signal. Not sure if a gain of zero really means anything?? Bob Long ```
 Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: David R. Sky - 2004-09-01 16:39:15 ```On Wed, 1 Sep 2004, Bob Long wrote: > On Wednesday, September 01, 2004 8:27 PM [GMT+1=CET], > Jeff and Linda Klein wrote: > > > 1 = amplification. Gain < 1 = > > > > > I think you mean to say Gain = 0 = no signal level change at all. > > I would have thought that a gain of 1 (unity) is no change to the signal. > Not sure if a gain of zero really means anything?? > Some of this is in the language. A gain or attenuation of 0 (whether db or something else) means no change in either direction. david ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: Jeff and Linda Klein - 2004-09-01 20:45:34 ```? Jeff BTW, the B in dB is capitalized because it really means 1/10 of a Bel (named after Alexander Graham Bell) Jeff Klein (AIM: eggbert090) Head Volunteer Sound Geek Faith Baptist Church Glen Burnie, Maryland http://faithbc.net http://w3.mivlmd.cablespeed.com/~klein6/web-public/ --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.746 / Virus Database: 498 - Release Date: 8/31/2004 ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: David R. Sky - 2004-09-01 20:57:08 ```By referring to language, here's a simpler example. If you have \$100.00, and you increased it by 0 dollars, or 0 cents, or 0 pounds sterling, you still have \$100.00. If you again start with \$100.00 and increase it by 10 percent, you've added \$10 to it. Increase the \$100 by 10 pounds sterling and you have to do some converting. Adding gain or attenuation is an addition thing, not a multiplication thing. ... well, there is some multiplication. You can think of it as what was added to the original was multiplied by some number. Maybe that's not the way EQ really works, but the metaphor is appropriate. Thanks for your explanation Jeff! And I'll send the plug-in in a separate post. You can check it out and give me feedback. Remember that I don't know yet if my math is correct, but I think it is. David On Wed, 1 Sep 2004, Jeff and Linda Klein wrote: > > Yup, dB is just a unit of change and depending on what you're comparing > there are two equations used: > dB = 20 LOG (X/Y) for changes in voltage or sound level changes over > distance calculations > dB = 10 LOG (X/Y) for changes in power level or stationary sound output > calculations > > In both cases X = output value and Y = input value. If you were calculating > gain of an amplifier/mixer or sound output from a speaker you would use the > "10 LOG" version, but for voltage changes or sound pressure level changes as > you move around in a room you would use the "20 LOG" version. > > In either case, a dB is a dB no matter what you compare because the units > cancel out; it's just a unit of change of anything. Increasing from 10W to > 20W is the same as increasing from 500W to 1000W; both are +3dB or twice the > power of the original. If dB is followed by another character, like dBu or > dBm, then that is a constant logarithmic unit because it's been referenced > to a known linear value, i.e. 0 dBm is always 1 milliwatt, 3 dBm is always 2 > milliwatts, and 6 dBm is always 4 milliwatts. > > Confusing, huh ? > > Jeff > > BTW, the B in dB is capitalized because it really means 1/10 of a Bel (named > after Alexander Graham Bell) > > Jeff Klein (AIM: eggbert090) > Head Volunteer Sound Geek > Faith Baptist Church > Glen Burnie, Maryland > http://faithbc.net > http://w3.mivlmd.cablespeed.com/~klein6/web-public/ > > > --- > Outgoing mail is certified Virus Free. > Checked by AVG anti-virus system (http://www.grisoft.com). > Version: 6.0.746 / Virus Database: 498 - Release Date: 8/31/2004 > > > > ------------------------------------------------------- > This SF.Net email is sponsored by BEA Weblogic Workshop > FREE Java Enterprise J2EE developer tools! > Get your free copy of BEA WebLogic Workshop 8.1 today. > http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click > -- ```
 Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: Matt Brubeck - 2004-09-01 23:54:03 ```Jeff and Linda Klein wrote: >> Gain: a multiplication factor. Gain > 1 = amplification. Gain < 1 = >> attenuation. Gain = 0 = no signal gets though at all. > > I think you mean to say Gain = 0 = no signal level change at all. You're both right -- one of you is using a linear scale (where unity = 1) and the other is using a logarithmic scale (where unity = 0 dB). Switching to logarithmic units like dB transforms multiplication into addition, and 1 into 0. ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: SMITH Linton - 2004-09-01 23:41:10 ```No. Gain (in the electronics world) is a multiplication factor.=20=20 Signal x 0 =3D 0=20=20 Signal x 1 =3D Signal -----Original Message----- From: audacity-users-admin@... [mailto:audacity-users-admin@...]On Behalf Of Jeff and Linda Klein Sent: Wednesday, 1 September 2004 7:57 PM To: audacity-users@... Subject: RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? 1 =3D amplification. Gain < 1 =3D
 [Audacity-users] re: Gain and multiplying and adding again From: David R. Sky - 2004-09-01 23:55:00 ```Hmmm. I meant, if you apply 0 db gain to a signal, there is no change. ie. signal plus 0 db gain equals original signal. Maybe it's that log thing, you know, adders multiplying on a log table. Also known as breeding snakes. So applying 0 db is the same as multiplying by 1? David On Thu, 2 Sep 2004, SMITH Linton wrote: > No. Gain (in the electronics world) is a multiplication factor. > Signal x 0 = 0 > Signal x 1 = Signal > > -----Original Message----- > From: audacity-users-admin@... > [mailto:audacity-users-admin@...]On Behalf Of Jeff and > Linda Klein > Sent: Wednesday, 1 September 2004 7:57 PM > To: audacity-users@... > Subject: RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: > what do you want to see in it? > > > 1 = amplification. Gain < 1 = > > I think you mean to say Gain = 0 = no signal level change at all. > > BTW, if it hasn't been mentioned before equalizer gain/attenuation settings > are in dB, which is a logarithmic ratio of output to input level or a change > in level. Doesn't matter what the input power is, only the actual change. 0 > dB means no change in level level, a 3 dB increase doubles level, and a 3 dB > decrease cuts power in half. > > Jeff > > Jeff Klein (AIM: eggbert090) > Head Volunteer Sound Geek > Faith Baptist Church > Glen Burnie, Maryland > http://faithbc.net > http://w3.mivlmd.cablespeed.com/~klein6/web-public/ > > > --- > Outgoing mail is certified Virus Free. > Checked by AVG anti-virus system (http://www.grisoft.com). > Version: 6.0.746 / Virus Database: 498 - Release Date: 8/31/2004 > > > > ------------------------------------------------------- > This SF.Net email is sponsored by BEA Weblogic Workshop > FREE Java Enterprise J2EE developer tools! > Get your free copy of BEA WebLogic Workshop 8.1 today. > http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click > -- ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: SMITH Linton - 2004-09-01 23:59:27 Attachments: Message as HTML ```(Best viewed in a fixed point font) =20 +db ----------------------X \ \=20 -db \x--------------------------- f1 f2 =20 Although the tool displays what appears to be a continuous line, each pixel= is really representing the frequency at that point on the display. Theref= ore the tool is just displaying a lot of discrete points really close so it= looks like a continuous line. The FFT algorithm, reads each pixel as bein= g separated by a frequency step of F/N where F is the frequency band displa= yed (eg 20 hz to 20khz) and N is the number of pixels, (eg 1024). =20 If you arranged the sliders on a graphic equalizer in the same pattern it w= ould have the same effect and can be shown mathematically to have the same = effect as the FFT filter.=20 ie X =3D slider =20 +db X X X X X | | | | | |=20=20 | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | -db | | | | | X X X X X X f1 f2 =20 cheers =20 Linton -----Original Message----- From: audacity-users-admin@... [mailto:audacity-users-adm= in@...]On Behalf Of mail@... Sent: Wednesday, 1 September 2004 8:53 PM To: audacity-users@... Subject: Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what = do you want to see in it? >>> Linton.SMITH@...: a FFT is a Fast Fourier Transform. FFT is a ma= thematical method of determining the frequency content of a block of signal= . ie it converts a signal represented in the time domain (ie signal vs tim= e) into a representation in the frequency domain ((ie sginal vs frequency) = From what I can gather from my use of the FFT Filter, you set the frequency= response that you want on a gain vs frequency graph and by using the inver= se FFT calculation, a time domain filter is produced which is used to filt= er the signal. =20 If you take the flat line starting point with this tool and then make a con= trol point , then I assume along this discontinuity you are in effect havin= g the equivalent of many sets of knobs on a hardware equaliser with the kno= b set at greatest deviation being at the control point? I may be obtuse b= ut if you look say at the control point, is this effect applied at the whol= e range of hearing or only at the width of some predefined range of octave= s. In other words is this control point centred around a frequency as per p= revious discussions?=20 =20 =20 ```
 Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: - 2004-09-02 00:21:06 ``` On Thu, Sep 02, 2004 at 09:52:00AM +1000, SMITH Linton wrote: > If you arranged the sliders on a graphic equalizer in the same pattern it would have the same effect and can be shown mathematically to have the same effect as the FFT filter. > ie X = slider > > +db X X X X X | | | | | | > | | | | | | | | | | | > | | | | | | | | | | | > | | | | | | | | | | | > -db | | | | | X X X X X X > f1 f2 I assume the FFT filter is actually a bit more sophisticated and this is just glossing a bit by way of keeping the explanation clear? This explanation is intuitive but subtly wrong in that you can see why it works, but what a proper implementation is actually doing is using the FFT to perform a fast time-domain convolution. Ie, you can't just multiply an FFT by an ideal desired frequency response and get acceptible quality output; it would mostly work, but not really work well. At a minimum, you need to window the time response of both the input data *and* the EQ 'curve' to smooth, symmetric windows and pad both the data and response to double-length with zeroes to avoid circular pollution in the transform, then add blocks with 75% overlap. Choose your window carefully! A poor window will be worse than no window at all (well, a square window). Monty ```
 Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: Monty - 2004-09-02 00:43:19 ```(Oh and don't forget; your window has to actually be wide enough to offer the frequency resolution in the low end that the UI claims it does!) Monty ```
 Re: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: David R. Sky - 2004-09-02 02:08:02 ```On Wed, 1 Sep 2004, Monty wrote: > (Oh and don't forget; your window has to actually be wide enough to > offer the frequency resolution in the low end that the UI claims it > does!) Well, this is a Nyquist plug-in. Since I'm using a screen reader, I have no idea what kind of visual stuff besides text is displayed. I have already posted the plug-in for testing earlier today... 'Single-band multichannel E Q'. Actually, could you tell me what you do see (besides text) in the Nyquist plug-ins? When you open the effects menu, the first group of alphabetized plug-ins are VST or Audacity's built-in effects. The second group of alphabetized effects are Nyquist. The first group are .dll files, the second group are made using plain text. David ```
 [Audacity-users] Widening stereo spread From: - 2004-09-05 12:15:44 Attachments: Message as HTML ```Someone (sorry, I forget who) posted a method other than panning to widen stereo effect by cancelling out some of the centre: Invert phase effect and splitting the stereo track. You can then use the inverted mono to cancel out some of the centre, giving you wider stereo. This is a variation on the MS technique commonly used by mastering engineers to widen stereo spread. 1. Open your stereo track 2. Copy the entire track to another stereo track 3. On your original track, split stereo. You now have 2 mono and 1 stereo tracks 4. On each of the mono tracks: make mono, then invert (from the effect menu), then drop level by about 20dB. 5. Export or quickmix. I found that did widen the stereo field whilst (like the full Karaoke/voice removal effect) bring to light unexpected detail which panning does not. However I am interested if there is a non technical explanation why cancelling some of the centre appears to push the left and right outwards rather than leaving them where they were and just making the centre quieter. Sorry if it is a stupid question but it is of interest to me. ```
 Re: [Audacity-users] Widening stereo spread From: David R. Sky - 2004-09-05 15:18:14 ```On Sun, 5 Sep 2004 mail@... wrote: > Someone (sorry, I forget who) posted a method other than panning to widen stereo effect by cancelling out some of the centre: > {snip} I found that did widen the stereo field whilst (like the full Karaoke/voice removal effect) bring to > light unexpected detail which panning does not. However I am interested if there is a non technical explanation why cancelling some of the centre appears to push the left and right outwards rather than leaving them where they were and just making the centre quieter. Sorry if it is a stupid question but it is of interest to me. > I don't know the answer specifically and would be interested too. As far as I'm concerned, the only time a question is stupid is when you don't ask it. This is kinduva hot button for me. David ```
 Re: [Audacity-users] Widening stereo spread From: - 2004-09-06 07:31:28 ``` On Sun, Sep 05, 2004 at 01:15:47PM +0100, mail@... wrote: > Someone (sorry, I forget who) posted a method other than panning to widen stereo effect by cancelling out some of the centre: This does work. Another easy way to artificially widen apparent stereo spred is to crossfeed an attenuated, slightly time delayed version of each channel to the other channel. To do that even better, use a bandpassed version of the delayed signal (roll off below 100Hz and above 4kHz). Monty ```
 Re: [Audacity-users] Widening stereo spread From: - 2004-09-06 11:29:01 ```>>>Monty: Another easy way to artificially widen apparent stereo spred is to crossfeed an attenuated, slightly time delayed version of each channel to the other channel. To do that even better, use a bandpassed version of the delayed signal (roll off below 100Hz and above 4kHz). Is that not also typically used to add a stereo effect to a mono recording (something I don't normally like, however done) ```
 Re: [Audacity-users] Widening stereo spread From: - 2004-09-06 17:09:12 ``` On Mon, Sep 06, 2004 at 12:29:05PM +0100, mail@... wrote: > > >>>Monty: Another easy way to artificially widen apparent > stereo spred is to crossfeed an attenuated, slightly time delayed > version of each channel to the other channel. To do that even better, > use a bandpassed version of the delayed signal (roll off below 100Hz > and above 4kHz). > > > Is that not also typically used to add a stereo effect to a mono recording (something I don't normally like, however done) Not by itself, as the channels start out the same and that would simply produce an echoey mono sample. A variant is often used to layer mono tracks into an apparent position in a stereo studio-produced track. The effect is much more convincing (and less fatiguing when done well) than trying to build a stereo image out of only L-R intensity differences. Monty ```
 RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? From: SMITH Linton - 2004-09-02 00:05:05 ```Sorry David,=20 But in my introduction I indicated the domain of my discussion ie electronic engineering Gain is a multiplication factor not an addition. Having said that and adding on Jeff or Grahams comments on dB (Sorry Mr Bell) Since it is a Logarithmic Scale, multiplication of "normal" numbers becomes addition when those numbers are converted to logarithms. So a gain of 1 is the same as a gain of 0dB. a gain > 1 will be a gain of XdB where X > 0. A gain < 1 will be a gain of Y dB where Y < 0. Now are we all totally confused? -----Original Message----- From: audacity-users-admin@... [mailto:audacity-users-admin@...]On Behalf Of David R. Sky Sent: Thursday, 2 September 2004 6:27 AM To: audacity-users@... Subject: RE: Spam:Re: [Audacity-users] Test multichannel EQ plug-in:: what do you want to see in it? By referring to language, here's a simpler example. If you have \$100.00, and you increased it by 0 dollars, or 0 cents, or 0 pounds sterling, you still have \$100.00. If you again start with \$100.00 and increase it by 10 percent, you've added \$10 to it. Increase the \$100 by 10 pounds sterling and you have to do some converting. Adding gain or attenuation is an addition thing, not a multiplication thing. ... well, there is some multiplication. You can think of it as what was added to the original was multiplied by some number. Maybe that's not the way EQ really works, but the metaphor is appropriate. Thanks for your explanation Jeff! And I'll send the plug-in in a separate post. You can check it out and give me feedback. Remember that I don't know yet if my math is correct, but I think it is. David On Wed, 1 Sep 2004, Jeff and Linda Klein wrote: > > Yup, dB is just a unit of change and depending on what you're comparing > there are two equations used: > dB =3D 20 LOG (X/Y) for changes in voltage or sound level changes over > distance calculations > dB =3D 10 LOG (X/Y) for changes in power level or stationary sound output > calculations > > In both cases X =3D output value and Y =3D input value. If you were calculating > gain of an amplifier/mixer or sound output from a speaker you would use the > "10 LOG" version, but for voltage changes or sound pressure level changes as > you move around in a room you would use the "20 LOG" version. > > In either case, a dB is a dB no matter what you compare because the units > cancel out; it's just a unit of change of anything. Increasing from 10W to > 20W is the same as increasing from 500W to 1000W; both are +3dB or twice the > power of the original. If dB is followed by another character, like dBu or > dBm, then that is a constant logarithmic unit because it's been referenced > to a known linear value, i.e. 0 dBm is always 1 milliwatt, 3 dBm is always 2 > milliwatts, and 6 dBm is always 4 milliwatts. > > Confusing, huh ? > > Jeff > > BTW, the B in dB is capitalized because it really means 1/10 of a Bel (named > after Alexander Graham Bell) > > Jeff Klein (AIM: eggbert090) > Head Volunteer Sound Geek > Faith Baptist Church > Glen Burnie, Maryland > http://faithbc.net > http://w3.mivlmd.cablespeed.com/~klein6/web-public/ > > > --- > Outgoing mail is certified Virus Free. > Checked by AVG anti-virus system (http://www.grisoft.com). > Version: 6.0.746 / Virus Database: 498 - Release Date: 8/31/2004 > > > > ------------------------------------------------------- > This SF.Net email is sponsored by BEA Weblogic Workshop > FREE Java Enterprise J2EE developer tools! > Get your free copy of BEA WebLogic Workshop 8.1 today. > http://ads.osdn.com/?ad_id=3D5047&alloc_id=3D10808&op=3Dclick > --=20 ------------------------------------------------------- This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools! Get your free copy of BEA WebLogic Workshop 8.1 today. http://ads.osdn.com/?ad_id=3D5047&alloc_id=3D10808&op=3Dclick --=20 Mailing list: Audacity-users@... To UNSUBSCRIBE, use the form at the bottom of this web page: https://lists.sourceforge.net/lists/listinfo/audacity-users ```
1 2 > >> (Page 1 of 2)