[audacity4blind] Re: high shelf filter
- From: Robert Hänggi <aarjay.robert@xxxxxxxxx>
- To: audacity4blind@xxxxxxxxxxxxx
- Date: Tue, 9 Sep 2014 02:47:43 +0200
Hi Gary
I've been a little busy lately, so the plug-in couldn't be finished
(at least not for all platforms).
I'm just wondering if I should take the way over the equalizer effect.
The filter could as well be directly implemented as a stand-alone
effect plug-in. This allows more freedom in the choice of the
parameters.
I use sometimes your described way using the plot spectrum to compare audio.
The equalizer is unfortunately only accessible in the graphic mode.
This gives some constraints on what one can do. Firstly, the sliders
go only from 20 to -20 dB. Secondly, the center frequency of the
sliders is fixed and they are exactly 1/3 octave apart.
(detailed audiograms are similarly spaced, I think)
The plot spectrum export is on the other hand linearly, i.e. a
512-size spectrum at 51.2 kHz (for simplicity sake) gives the dB
values for 100, 200, 300, ... 25600 Hz.
This means that you have to add up some frequencies and average them
to get the value for the specific 1/3 octave point.
For instance, I have a template sheet where I import the spectrum
values. The comparison is either another voice spectrum or a general
one where the dB values are calculated from 1/frequency (e.g. 1/500
hz-->-54 dB, 1/1000--> -60 dB, thus 6 dB roll-off per octave).
Another columns holds the interpolated values to use in the equalizer.
This procedure is only intended to roughly adapt the overall speech
quality (clarity, brightness, boominess etc.) and not to exactly copy
another voice.
As you've already mentioned, this involves treating each phonem
separately and analysing the formants.
By no means a trivial task but even possible with Nyquist and Audacity.
Regards
Robert
2014-09-08 8:56 GMT+02:00, Gary Campbell <campg2003@xxxxxxxxx>:
> That sounds interesting. I use hearing aids and my wife is always
> frustrated because I don't hear her, and no it's not selective hearing
> loss :-). I got the data points from my audiogram from my audiologist
> and sometime plan to try to duplicate it with the equalizer, that is,
> after I figure out how to use the equalizer. Your plugin sounds like it
> does the same thing except that it gives you the hearing test, too.
>
> I was also wondering if significant components of her voice happened to
> fall within dips in my audiogram. I was wondering if the plot spectrum
> feature of Audacity is accurate enough to analyze a voice, since
> Audacity's plots can be output to a spreadsheet, which might actually
> give me a way to "look" at the plot.
>
> BTW: Several years ago I tried to search for information on voice
> formants and didn't find much. Does anyone know of any stuff on the web
> that explains how frequencies are distributed in speech? or differences
> between voices (speaker identification)? I have heard of voice prints--
> I would be interested in the details of how they work.
>
> Thanks.
> Gary
>
>
> On 5/29/2014 4:28 PM, Robert Hänggi wrote:
>> Hi Brad
>>
>> I'm currently writing a plug-in that measures the threshold of hearing
>> and might be of use to you (at least if you're on Windows).
>> It's more accurate than a simple high-shelf filter because it uses the
>> Audacity EQ bands.
>> It works like this:
>> Each plug-in call plays a specific frequency with ascending volume (3
>> dB per second).
>> The user presses space as soon as he hears a tone. This value will be
>> stored and Ctrl-R plays the next sequence.
>> After all bands, the result is stored and can be imported as a XML
>> file into the Equalizer effect.
>>
>> The resulting curve makes that you hear all frequencies equally loud.
>> Inverting the curve will attenuate frequencies where your hearing is
>> weak and does therefore simulate your hearing experience for others.
>> However, the test persons should actually take their profile too, or
>> the filtered audio will not sound quite the same, i.e. it will
>> probably be too strong.
>>
>> Although the concept is very easy, the programming is really hard due
>> to the Nyquist limitations. Nevertheless, the main code runs and I
>> hope to finish it by the end of the week. You and all other
>> subscribers are welcomed to test it.
>>
>> Best regards
>> Robert
>>
>> The audacity4blind web site is at
>> //www.freelists.org/webpage/audacity4blind
>>
>> Subscribe and unsubscribe information, message archives,
>> Audacity keyboard commands, and more...
>>
>> To unsubscribe from audacity4blind, send an email to
>> audacity4blind-request@xxxxxxxxxxxxx
>> with subject line
>> unsubscribe
>>
>
>
> The audacity4blind web site is at
> //www.freelists.org/webpage/audacity4blind
>
> Subscribe and unsubscribe information, message archives,
> Audacity keyboard commands, and more...
>
> To unsubscribe from audacity4blind, send an email to
> audacity4blind-request@xxxxxxxxxxxxx
> with subject line
> unsubscribe
>
>
--
The audacity4blind web site is at
//www.freelists.org/webpage/audacity4blind
Subscribe and unsubscribe information, message archives,
Audacity keyboard commands, and more...
To unsubscribe from audacity4blind, send an email to
audacity4blind-request@xxxxxxxxxxxxx
with subject line
unsubscribe
Other related posts:
