W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > July to September 2018

Re: Measuring Decibel Separation

From: Greg Gay <greggay@rogers.com>
Date: Tue, 18 Sep 2018 14:42:24 -0400
To: w3c-wai-ig@w3.org
Message-ID: <61058d48-defa-30c9-8714-7acf94dc9d50@rogers.com>
Thx David,

After posting the original question, I did come across an old article by 
David MacDonald that essentially suggests the same strategy you've 
outlined. Take a sample of the background, take a sample of the 
background with speech in it, then compare the decibels of each sample.

http://www.eramp.com/david/audio_contrast_general_techs.htm

I happen to have ProTools available, which apparently will do the job.

https://www.avid.com/pro-toolss

Does anyone know of an *open source or free tool* that can be 
recommended for client who want to do their own audio sampling, who 
might not otherwise have a use for a tool like ProTools?

greg



On 2018-09-18 1:50 PM, David Woolley wrote:
> On 18/09/18 18:23, Greg Gay wrote:
>> How does one measure whether speech in multimedia is 20db louder than 
>> background sound, from an auditing perspective.
>>
>> During production its easy enough to control speech and background 
>> levels, but what I'm looking for is a way to show clients post 
>> production, that the speech volumn level in a video is less than 20db 
>> greater than the volume of the background audio in that video. (re 
>> WCAG 2.1 SC 1.4.7)
>
> This really needs to be done white box to properly demonstrate 
> compliance, but 20dB is 10 times more power, so, I'd suggest 
> determining sections of the audio with no speech and measuring peak 
> amplitude (probably fast A weighted**), then measuring the peak 
> amplitude in a section with speech.  If the ratio is more than 11 
> times power (about 20.8 dB, say 21 dB)you are getting close to 
> complying, although one needs some safety margin to account for poor 
> selection of the quiet and speech sections.
>
> You'll need to exclude the periods with the allowable occasional 
> sounds, first.
>
> ** The criteria should specify this, although they appear not to do so!
>
>
Received on Tuesday, 18 September 2018 18:42:52 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:37:21 UTC