Forums

VOIP, Call Center, and DTMF errors...

Started by David Morgan (MAMS) April 12, 2007
New to forum, new to industry, but 30 years in audio.
__________________________________________


The header is not very clear, I know...  but I have a question that I'd
love to have addressed by anyone with experience in SRC for
telephone call center software using low-Khz voice files for
automated prompting.



Issue: (assumed on my part)

Distortion within a compressed audio file triggers random DTMF tone,
interrupting or altering end user's software application which interfaces
with incoming caller.



Background:

Recording system in vocal booth creates audio files of voice talent at
16 / 22,050Khz

Audio files are edited into single 'prompts' and batch processed with filtering
and the end result is an 8Khz file used by the software handling the calls.

Updates to the voice 'prompts' for the software package are sent to the
client via VOIP.

Audio files are used by software in automated call centers to route calls
and handle customer service, etc...  (Press one for english)


Front end recording setup:

I have determined that there is a gain staging problem between at least
two devices in the path, but there are also equipment quality issues which
may or may not be related to the end result.

Bear in mind that this equipment has put this company (and it's unique
take on the software that uses these audio files) on the map to ten years
of multi-million dollar success.

Nice vocal booth... talent operates Cool-Edit on quiet PC (Hush) to self-
record voice.  (These files are networked to another office for editing)

ATM-33 on desk boom
DBX 286A - Mic Pre / channel strip
E-MU 2 channel mic pre (used only as A/D convertor to PC via
          proprietary wired CAT-5 cable)



One problem cured....  but is there anything else ???

The output of the DBX was overdriving the line input to the E-MU, and the
problem may be cured... but a multi-million dollar business is not going
to be happy if their incoming calls are dropped again when they bring the
software package back on line with the new voice prompts.

I want to see the E-MU replaced... it has faaar too much self noise.
I'd like to see them working with an RE-20 or something more than
the little AT.... but it's been making them happy for many years so
I don't know exactly how to recommend the change.  The DBX may
be suffering from slow capacitor death, as many of the process
controls seem sluggish and possible randomly inaccurate.

I think that altering the gain stages to eliminate any more potential
distortion in the initial recordings may cure the issue, but what if
(and *could*) the problem actually be occurring later in the conversion
process to 8Khz?   Are there any other possible causes of random
DTMF tone generation besides the light distortion in the originating
files which was creating some rather blatant crackling after the final
conversion to 8K?

I just don't want to overlook any potential problems and this isn't
something I address every day.

One of my first recommendations was to bump the original recording
level up to 48Khz.... making the conversion to 8K, even numbered math.

This is not my forte', but I'd love to gain enough knowledge on the
workings of call-center software and implementation in order to
converse logically with my contractor.

AM I wrong to think that some light distortion on the front end may
have been 100% the culprit?  After all, the company has been doing
just fine on this gear (unattended) for many years.

TIA for any input...



-- 
David Morgan (MAMS)
http://www.m-a-m-s DOT com
Morgan Audio Media Service
Dallas, Texas   (214) 662-9901
_______________________________________
http://www.artisan-recordingstudio.com










Anyone here have call-center sofware experience or
experience in implementing these packages with the
phone lines?

DM




CAN I GET SOME HELP HERE.....?       ;-)        Some discussion, maybe?



"David Morgan (MAMS)" <findme@m-a-m-s.comC/Odm> wrote in message news:cAvTh.3029$vo2.2667@trnddc01...
> New to forum, new to industry, but 30 years in audio. > __________________________________________ > > > The header is not very clear, I know... but I have a question that I'd > love to have addressed by anyone with experience in SRC for > telephone call center software using low-Khz voice files for > automated prompting. > > > > Issue: (assumed on my part) > > Distortion within a compressed audio file triggers random DTMF tone, > interrupting or altering end user's software application which interfaces > with incoming caller. > > > > Background: > > Recording system in vocal booth creates audio files of voice talent at > 16 / 22,050Khz > > Audio files are edited into single 'prompts' and batch processed with filtering > and the end result is an 8Khz file used by the software handling the calls. > > Updates to the voice 'prompts' for the software package are sent to the > client via VOIP. > > Audio files are used by software in automated call centers to route calls > and handle customer service, etc... (Press one for english) > > > Front end recording setup: > > I have determined that there is a gain staging problem between at least > two devices in the path, but there are also equipment quality issues which > may or may not be related to the end result. > > Bear in mind that this equipment has put this company (and it's unique > take on the software that uses these audio files) on the map to ten years > of multi-million dollar success. > > Nice vocal booth... talent operates Cool-Edit on quiet PC (Hush) to self- > record voice. (These files are networked to another office for editing) > > ATM-33 on desk boom > DBX 286A - Mic Pre / channel strip > E-MU 2 channel mic pre (used only as A/D convertor to PC via > proprietary wired CAT-5 cable) > > > > One problem cured.... but is there anything else ??? > > The output of the DBX was overdriving the line input to the E-MU, and the > problem may be cured... but a multi-million dollar business is not going > to be happy if their incoming calls are dropped again when they bring the > software package back on line with the new voice prompts. > > I want to see the E-MU replaced... it has faaar too much self noise. > I'd like to see them working with an RE-20 or something more than > the little AT.... but it's been making them happy for many years so > I don't know exactly how to recommend the change. The DBX may > be suffering from slow capacitor death, as many of the process > controls seem sluggish and possible randomly inaccurate. > > I think that altering the gain stages to eliminate any more potential > distortion in the initial recordings may cure the issue, but what if > (and *could*) the problem actually be occurring later in the conversion > process to 8Khz? Are there any other possible causes of random > DTMF tone generation besides the light distortion in the originating > files which was creating some rather blatant crackling after the final > conversion to 8K? > > I just don't want to overlook any potential problems and this isn't > something I address every day. > > One of my first recommendations was to bump the original recording > level up to 48Khz.... making the conversion to 8K, even numbered math. > > This is not my forte', but I'd love to gain enough knowledge on the > workings of call-center software and implementation in order to > converse logically with my contractor. > > AM I wrong to think that some light distortion on the front end may > have been 100% the culprit? After all, the company has been doing > just fine on this gear (unattended) for many years. > > TIA for any input... > > > > -- > David Morgan (MAMS) > http://www.m-a-m-s DOT com > Morgan Audio Media Service > Dallas, Texas (214) 662-9901 > _______________________________________ > http://www.artisan-recordingstudio.com > > > > > > > > > >
David,

Do you have a sample .wav file that triggers the DTMF tone?

--Randy

"David Morgan \(MAMS\)" <findme@m-a-m-s.comC/Odm> writes:

> New to forum, new to industry, but 30 years in audio. > __________________________________________ > > > The header is not very clear, I know... but I have a question that I'd > love to have addressed by anyone with experience in SRC for > telephone call center software using low-Khz voice files for > automated prompting. > > > > Issue: (assumed on my part) > > Distortion within a compressed audio file triggers random DTMF tone, > interrupting or altering end user's software application which interfaces > with incoming caller. > > > > Background: > > Recording system in vocal booth creates audio files of voice talent at > 16 / 22,050Khz > > Audio files are edited into single 'prompts' and batch processed with filtering > and the end result is an 8Khz file used by the software handling the calls. > > Updates to the voice 'prompts' for the software package are sent to the > client via VOIP. > > Audio files are used by software in automated call centers to route calls > and handle customer service, etc... (Press one for english) > > > Front end recording setup: > > I have determined that there is a gain staging problem between at least > two devices in the path, but there are also equipment quality issues which > may or may not be related to the end result. > > Bear in mind that this equipment has put this company (and it's unique > take on the software that uses these audio files) on the map to ten years > of multi-million dollar success. > > Nice vocal booth... talent operates Cool-Edit on quiet PC (Hush) to self- > record voice. (These files are networked to another office for editing) > > ATM-33 on desk boom > DBX 286A - Mic Pre / channel strip > E-MU 2 channel mic pre (used only as A/D convertor to PC via > proprietary wired CAT-5 cable) > > > > One problem cured.... but is there anything else ??? > > The output of the DBX was overdriving the line input to the E-MU, and the > problem may be cured... but a multi-million dollar business is not going > to be happy if their incoming calls are dropped again when they bring the > software package back on line with the new voice prompts. > > I want to see the E-MU replaced... it has faaar too much self noise. > I'd like to see them working with an RE-20 or something more than > the little AT.... but it's been making them happy for many years so > I don't know exactly how to recommend the change. The DBX may > be suffering from slow capacitor death, as many of the process > controls seem sluggish and possible randomly inaccurate. > > I think that altering the gain stages to eliminate any more potential > distortion in the initial recordings may cure the issue, but what if > (and *could*) the problem actually be occurring later in the conversion > process to 8Khz? Are there any other possible causes of random > DTMF tone generation besides the light distortion in the originating > files which was creating some rather blatant crackling after the final > conversion to 8K? > > I just don't want to overlook any potential problems and this isn't > something I address every day. > > One of my first recommendations was to bump the original recording > level up to 48Khz.... making the conversion to 8K, even numbered math. > > This is not my forte', but I'd love to gain enough knowledge on the > workings of call-center software and implementation in order to > converse logically with my contractor. > > AM I wrong to think that some light distortion on the front end may > have been 100% the culprit? After all, the company has been doing > just fine on this gear (unattended) for many years. > > TIA for any input... > > > > -- > David Morgan (MAMS) > http://www.m-a-m-s DOT com > Morgan Audio Media Service > Dallas, Texas (214) 662-9901 > _______________________________________ > http://www.artisan-recordingstudio.com > > > > > > > > > >
-- % Randy Yates % "Midnight, on the water... %% Fuquay-Varina, NC % I saw... the ocean's daughter." %%% 919-577-9882 % 'Can't Get It Out Of My Head' %%%% <yates@ieee.org> % *El Dorado*, Electric Light Orchestra http://home.earthlink.net/~yatescr
"Randy Yates" <yates@ieee.org> wrote in message news:m3bqhora6x.fsf@ieee.org...

> David, > > Do you have a sample .wav file that triggers the DTMF tone?
Yes.... but I have no clue what *actually* could have happened where the triggering occurred... it appears, from some very light conversation with user techs, that any number of things might be the culprit triggering the tone. I have it in the original 22,050 recorded format; and, in the 8K, peak normalized file that gets played by the telephone interfacing SW. (I can see your interest in lowering the peak value). This is how I got the gig... technicians on the client side (the users of the software package) say their end checks out ok, so the seller and designer of the package called in an audio tech (me) to examine the front-end recording set-up. Apparently editors and QC people, non-audio types, were missing the distortion in it's entirety. I'd like to be informed enough to challenge the end user's implementation, hardware, and phone lines, should it become necessary after the front end is repaired. What happened on the front end, was an overdriven component (cheap A/D convertor) created distortion, which was then enhanced not only by normalization, but multiplied by aliasing during the odd-math conversion from 22,050 to 8K; then shipped via VOIP to the customer. The convertor is bad, inputs distorted and digital outputs full of hash... it'll be replaced this week, but I need some fast knowledge on the user end of things to help me aid in protecting my contract employer and increase my own personal net worth. DM I can put them on my web site later today...
> --Randy
> "David Morgan \(MAMS\)" <findme@m-a-m-s.comC/Odm> writes: > > > New to forum, new to industry, but 30 years in audio. > > __________________________________________ > > > > > > The header is not very clear, I know... but I have a question that I'd > > love to have addressed by anyone with experience in SRC for > > telephone call center software using low-Khz voice files for > > automated prompting. > > > > > > > > Issue: (assumed on my part) > > > > Distortion within a compressed audio file triggers random DTMF tone, > > interrupting or altering end user's software application which interfaces > > with incoming caller. > > > > > > > > Background: > > > > Recording system in vocal booth creates audio files of voice talent at > > 16 / 22,050Khz > > > > Audio files are edited into single 'prompts' and batch processed with filtering > > and the end result is an 8Khz file used by the software handling the calls. > > > > Updates to the voice 'prompts' for the software package are sent to the > > client via VOIP. > > > > Audio files are used by software in automated call centers to route calls > > and handle customer service, etc... (Press one for english) > > > > > > Front end recording setup: > > > > I have determined that there is a gain staging problem between at least > > two devices in the path, but there are also equipment quality issues which > > may or may not be related to the end result. > > > > Bear in mind that this equipment has put this company (and it's unique > > take on the software that uses these audio files) on the map to ten years > > of multi-million dollar success. > > > > Nice vocal booth... talent operates Cool-Edit on quiet PC (Hush) to self- > > record voice. (These files are networked to another office for editing) > > > > ATM-33 on desk boom > > DBX 286A - Mic Pre / channel strip > > E-MU 2 channel mic pre (used only as A/D convertor to PC via > > proprietary wired CAT-5 cable) > > > > > > > > One problem cured.... but is there anything else ??? > > > > The output of the DBX was overdriving the line input to the E-MU, and the > > problem may be cured... but a multi-million dollar business is not going > > to be happy if their incoming calls are dropped again when they bring the > > software package back on line with the new voice prompts. > > > > I want to see the E-MU replaced... it has faaar too much self noise. > > I'd like to see them working with an RE-20 or something more than > > the little AT.... but it's been making them happy for many years so > > I don't know exactly how to recommend the change. The DBX may > > be suffering from slow capacitor death, as many of the process > > controls seem sluggish and possible randomly inaccurate. > > > > I think that altering the gain stages to eliminate any more potential > > distortion in the initial recordings may cure the issue, but what if > > (and *could*) the problem actually be occurring later in the conversion > > process to 8Khz? Are there any other possible causes of random > > DTMF tone generation besides the light distortion in the originating > > files which was creating some rather blatant crackling after the final > > conversion to 8K? > > > > I just don't want to overlook any potential problems and this isn't > > something I address every day. > > > > One of my first recommendations was to bump the original recording > > level up to 48Khz.... making the conversion to 8K, even numbered math. > > > > This is not my forte', but I'd love to gain enough knowledge on the > > workings of call-center software and implementation in order to > > converse logically with my contractor. > > > > AM I wrong to think that some light distortion on the front end may > > have been 100% the culprit? After all, the company has been doing > > just fine on this gear (unattended) for many years. > > > > TIA for any input... > > > > > > > > -- > > David Morgan (MAMS) > > http://www.m-a-m-s DOT com > > Morgan Audio Media Service > > Dallas, Texas (214) 662-9901 > > _______________________________________ > > http://www.artisan-recordingstudio.com > > > > > > > > > > > > > > > > > > > > > > -- > % Randy Yates % "Midnight, on the water... > %% Fuquay-Varina, NC % I saw... the ocean's daughter." > %%% 919-577-9882 % 'Can't Get It Out Of My Head' > %%%% <yates@ieee.org> % *El Dorado*, Electric Light Orchestra > http://home.earthlink.net/~yatescr
David Morgan (MAMS) wrote:
> CAN I GET SOME HELP HERE.....? ;-) Some discussion, maybe?
Maybe if you tried to give a coherent description of your problem, you might get more response. I have no real clue what you are actually getting. You keep asking about call centres, but seem to be working with IVRs. You make references to VoIP, but don't describe the path. You keep going on about distortion but don't identify the relevance. So.... Is this a pure IVR system, which is now encountering trouble for the first time, as people begin to use it over VoIP paths? Is the problem with a particular VoIP path, or over various services? Is the problem that when your voice prompt goes out, something in the network is falsely interpreting this as outgoing DTMF and sending a DTMF beep back? Is it your own IVR which is falsely detecting the DTMF from the voice prompt? My guess would be you are getting this with a particular path, and getting beeps back, which your IVR is seeing as user input. This would mean that something in the VoIP path has a really crappy DTMF detector. Good detectors generate no more false DTMF on distorted signals than they do on clean ones. Regards, Steve
"Steve Underwood" <steveu@dis.org> wrote in message...

> David Morgan (MAMS) wrote: > > > > CAN I GET SOME HELP HERE.....? ;-) Some discussion, maybe?
> Maybe if you tried to give a coherent description of your problem, you > might get more response. I have no real clue what you are actually > getting. You keep asking about call centres, but seem to be working with > IVRs. You make references to VoIP, but don't describe the path. You keep > going on about distortion but don't identify the relevance. So....
In spite of your inability to acknowledge my admissions of being a 'newbie' at the technology I am inquiring about, and your resultant rude-assed albeit accurate observation, I sincerely appreciate your comments. I am an audio guy... a recordist, a studio engineer. I record, mix, and produce music and voice, and I have done so for over 33 years. I don't know jack-shit about telephone lines, IVR, VoIP, or even DTMF usage other than how it is generated and applies to basic telephone. I need some links to how this crap works... I don't have hours to mull over hundreds of thousands of search results inclusive of those that contain no useful data at all.
> Is this a pure IVR system, which is now encountering trouble for the > first time, as people begin to use it over VoIP paths?
I don't know what an IVR is. Can you send me to a non-mathematical description of it's concept and application? (Pleeeease........)
> Is the problem with a particular VoIP path, or over various services?
What are various services? To the best of my knowledge, the problem occurred when a company put a new package of voice prompts into place for an additional product line, and those prompts contained distortion.
> Is the problem that when your voice prompt goes out, something in the > network is falsely interpreting this as outgoing DTMF and sending a DTMF > beep back? Is it your own IVR which is falsely detecting the DTMF from > the voice prompt?
The company who has hired me to address the front-end recording issues, sells a software package that failed to operate properly when a new set of voice promts were added to accomodate a new product. I can repair (or update) the front end voice prompt recording system until I am blue in the face, but we're talking about some seriously big losses and troubles if the problem persists downstream. If it does, I would certainly like to be able to communicate with more accuracy and knowledge, thus more input to my employer, than I obviously have the ability to at this time. So... I need all the guidance I can get to help me be assured that I ask my employer all of the right questions regarding how the system is implemented further downstream (and now, obviously, what *kind* of system has some bearing on the search for answers).
> My guess would be you are getting this with a particular path, and > getting beeps back, which your IVR is seeing as user input.
This is basically what was explained to me when I was asked to repair the front-end recording setup. But this makes little sense to me without garnering a little experience or some references for me to learn how this gear (system) operates and is implemented on the phone lines.
> This would mean that something in the VoIP path has a really crappy > DTMF detector. Good detectors generate no more false DTMF on > distorted signals than they do on clean ones.
I'll add this to my notes and continue my research. Any further guidance would **seriously** be appreciated. -- David Morgan (MAMS) http://www.m-a-m-s DOT com Morgan Audio Media Service Dallas, Texas (214) 662-9901 _______________________________________ http://www.artisan-recordingstudio.com
"David Morgan (MAMS)" <findme@m-a-m-s.comC/Odm> wrote in message...

> "Steve Underwood" <steveu@dis.org> wrote in message...
> > Is this a pure IVR system, which is now encountering trouble for the > > first time, as people begin to use it over VoIP paths? > > I don't know what an IVR is. Can you send me to a non-mathematical > description of it's concept and application? (Pleeeease........)
OK... I'm getting antsy, in a hurry, and a little rude myself, sorry. It is both incoming VR and DTMF response oriented.
On Apr 16, 4:18 pm, "David Morgan \(MAMS\)" <fin...@m-a-m-s.comC/Odm>
wrote:
> "David Morgan (MAMS)" <fin...@m-a-m-s.comC/Odm> wrote in message... > > > "Steve Underwood" <ste...@dis.org> wrote in message... > > > Is this a pure IVR system, which is now encountering trouble for the > > > first time, as people begin to use it over VoIP paths? > > > I don't know what an IVR is. Can you send me to a non-mathematical > > description of it's concept and application? (Pleeeease........) > > OK... I'm getting antsy, in a hurry, and a little rude myself, sorry. > > It is both incoming VR and DTMF response oriented.
I was hoping the group here could provide some details about how various system variables impact "TALK OFF"....such as source audio distortion, phone line termination, hybrid balance, echo supression, outgoing audio levels, and last but not least...DTMF decoder design, etc... To clarify the problem..., I believe Mr Morgans situation is that the DTMF decoder in the call center is falsing (talk off) on the OUTGOING audio voice prompts from the call center that he has pre-recorded. Mark