Twitter LinkedIn
Return to Wellspring


Revised Envelope Difference Index (rEDI)


Project TitleRevised Envelope Difference Index (rEDI)
Track Code08-033
Short Description

Software is available for download to aid in specifically analyzing the alterations to real speech signals associated with hearing aid processing


The software described here is needed specifically to analyze the alterations to real speech signals associated with hearing aid processing. Previous efforts to analyze the acoustic alterations of hearing aid processed speech have typically used a measure known as the Speech Transmission Index (STI). The Envelope Difference Index (EDI) approach differs from the STI in that it can be used on real speech signals of any length and is more appropriate for the types of acoustic changes caused by hearing aid processing.

Tagssoftware, algorithm, envelope, speech, hearing, audiology
Posted DateSep 17, 2010 6:31 PM


The term Envelope Difference Index (EDI) was first coined by Fortune, Woodruff, & Preves (Ear & Hearing, 1994) and uses an envelope subtraction technique to provide an index quantifying the difference between two temporal envelopes.  The EDI allows for a single value to be assigned to the differences between digital signals that have in excess of 40000 points.  The EDI was used to examine the differences between two envelopes across a single band but such systems have been constrained to one envelope extraction schema. 


The new rEDI software developed is unique from the originally-described EDI procedure in its ability to process envelope differences across a user defined number of frequency bands with weighting values attached to each band according to user specifications. The envelope extraction methodology is inherently flexible and can be varied by the user to capture finer or coarser envelopes.

Since human auditory filters are extremely sharp and narrow, the bandwidth of each filter band must be configurable to the same values. There are inherent limitations in engineering extremely steep and narrow filters. If a filter is too narrow there is insufficient support for the filter to be stable. Similarly, if the filter's cutoff frequencies are placed to close to 0 or the pi the filter remains unstable. Conventional finite impulse response (FIR) filters, if used to meet the requirements, are too high to be implemented in a near real-time environment. To overcome some of the limitations a novel filterbank is implemented in the rEDI software.


Any publications or presentations of material that arise from the use of the Software must acknowledge Dr. Lorienne Jenstad and Mr. Adrian Lister as the source of the Software.

Minimum Software Specifications

CPU 486, 33MHz


Free disk space 220MB

About the Software

Language Python

Platform X86

Trouble Shooting

The rEDI software  is not configured to run on Windows 7. If you get the following error:

"The program can’t start because MSVCP71.dll is missing from your computer. Try reinstalling the program to fix this problem."

you can try downloading the following application:


Name Price
rEDI software for download Free Download