Since the first days of telephony, noise became an integral part of communication.  Modern life caused the problem of ambient noise during phone calls to be even more acute than before.  There are several reasons for this, like:

1. Mobility. Today people make calls from almost every location: street, road.. and even from the toilette.

2. Home office. People tend to make many phone calls from their home. Usually the home office suffers from high ambient noise originated from: other people, kids, television, pets, etc.

3. Open space & Call centers. In both cases, phone calls tend to suffer from a noisy environment.

As a result companies invested and keep investing effort in improving the quality by identifying and removing ambient noise from the calls.  There are two main types of solutions for ambient noise reduction: hardware-based and software-based. We have covered these techniques in a previous blog post on ambient noise. In the current post we will concentrate on a unique solution that SoliCall patented – Personalized Noise Reduction.

Personalized Noise Reduction

Personalized noise reduction is a unique technique to remove ambient noise by using registered voice pattern of the user. The technique of registering voice pattern for later use is not new in computer science. For example, there are companies that use the registered voice pattern to  perform biometric identification or to improve speech-to-text translation. SoliCall is the only vendor that uses the technique of resisted voice patterns  to do real-time cleaning of voice calls. This novel concept of using registered patterns for real-time cleaning of voice calls provide very rewarding results. Following is a schematic drawing of the personalized noise reduction approach.

Use Cases

Since the personalized noise reduction  does not require any change to the existing hardware that is currently being used for the call, for example no additional sensors  or microphones need to be added,  the number of uses cases are enormous. In addition, the Personalized Noise Reduction can be performed in any location of the audio path: end devices, routers, PBX etc. Following are few samples of use cases:

1. Call center operators that want to provide higher quality experience to their customers by cancelling the noise that is generated in the call center.

2. Enterprises that want to enable their employees, to make high-quality business calls from noisy environment including form their home office.

3. VoIP service providers that offer hardware-independent telephony service and want to provide their users high quality calls regardless the hardware that they are using in each call.

4. Phone manufactures that want to offer superior noise reduction on top of their existing noise reduction hardware.

How to try  the personalized noise reduction ?

As of today,  SoliCall did not release the professional personalized noise reduction filter as part of its freely available products. Companies that are interested in evaluating this unique technology are invited to contact us.

In the previous posts of this series we discussed the challenges of echo cancellation algorithms to efficiently handle the echo path,  the complexity introduced by  echo invalidation and the howling effect. In this post we will discuss another important aspect with regards to echo cancellation called Saturation.

What is Saturation ?

The formal definition of saturation is the state that a digitized value reaches its limit and can not be increased even if the audio signal is amplified.  In most modern computers, the digitized wave samples are stored in a memory size of two bytes. When the energy in the audio signal exceed the value that can be stored in two bytes – the result is saturation. The visual impact on the audio signal is very clear, instead of having a nice sin-wave form, it looks like it has been cut in the edges.

sin wave audio signal

saturated audio signal

So now that we know what is saturation, how does it affect echo cancellation?

Saturation and Echo Cancellation

One of the main components in echo cancellation algorithms is the “correlation engine”. The correlation engine tries to match the reference signal (i.e. the signal that is being rendered to the speakers) with the input signal (i.e. the signal that is being captured by the microphone).  The correlation engine is looking for resemblance between the reference signal and the input signal. Whenever there is saturation in the input signal, it make it harder to find resemblance between the saturated input signal (…with the cuts..) and the reference signal that does not have these cuts. As a result, the correlation might fail or find low quality correlation. Therefore, the echo cancellation overall performance will be less optimal in such conditions. Modern echo canceller have heuristics to enable them to operate in these conditions but they will never reach the optimal results.

How can I avoid saturation ?

If you are making calls and your party complains that he/she can hear echo, one of the possible reasons might be that your input signal is saturated. In such case try doing the following:

1. Lower the volume of your microphone and disable/reduce microphone boost.

2. Lower the volume of your speakers and try to keep them as far away as possible from your microphone.

ISV/Manufactures/Operator that want to improve the audio quality for their users regardless of uncontrolled saturation conditions, should use an echo canceller that is robust to saturation conditions. SoliCall offers multiple products both as OEM, like the SDKs,  and as stand-alone plug-and-play products like PBXMate.

SoliCall offers different software solutions for audio quality improvements. One of the areas with an increasing demand is the need to solve echo issues with Android devices.  As discussed in few of our previous posts on echo cancellation, in real-life conditions performing a good echo cancellation is a rather complex task. In addition, it turns out that for Android devices this task is even harder. In this post we will review the main reasons and alternative solutions.

Audio Channels and Synchronization

On Android’s audio system it is harder to synchronize both sides of the audio. In other words, it is hard to synchronize the audio packets that are being sent to the speakers with the audio packets that are being captured by the microphone. Since the former audio stream is used as a reference to cancel echo from the latter audio stream, unsynchronized audio streams forces the AEC to operate in a volatile environment.

CPU Power

Since Android devices are run on many brands of CPU and hardware platforms, in different cases the CPU is simply not powerful enough to simultaneously run all the applications and in addition to do a proper acoustic echo cancellation. Moreover, the shortage in CPU power might also lead to an increase in echo tail/delay that can quickly reach few hundred ms.

Solutions

There are two options to solve this problem:

1.      Use an SDK with a professional AEC filter that has at least all the following capabilities: (a) Quickly recover from synchronization lost between the audio captured from the microphone and the audio rendered to the speakers; (b) Can handle an echo tail that might exceed the basic 128ms delay; (c) Can be configured to work with different CPU requirements in order to provide enough flexibility during integration.

2.      Use a centralized solution for echo cancellation. In this option, you provide a centralized professional acoustic echo cancellation on behalf of the Android mobile devices. One of products that offer this kind of centralized acoustic echo cancellation is the PBXMate.

Many VoIP network administrator face a common problem that their users complain about the quality of their phone calls. The most common complains are about noise, echo and other artifacts.

If the network is distributed, it becomes very difficult to analyze and control the multiple components in the network starting from the end devices (phones/computers) moving to the network components (Cables/Wiring/Gateways), and ending with the centralized units (PBX/ Media Device/Conference Bridge).

The PBXMate is popular software that is used to help monitoring the network and actively resolve quality issues regardless the cause of the problem.

In this post we will review few practical tips for PBXMate installation in a distributed network.

Firewall

By default the PBXMate is using the below ports. You need to make sure your firewall does not block them.

• 5061 UDP. This is the default listening port for SIP messaging. Note: the default is port 5061 and not port 5060 in order to avoid conflicts with any other SIP software that might already exists on the computer.
• 4000-4100 UDP. These ports are used for the RTP messaging. If these ports are blocked then you might not hear audio that is send from phones that are located in a remote location.
• 8083 TCP. This port is used for the web-based administration. If you are unable to view the web interface then you should open this port. Note: for Elastix users, that installed the PBXMate addon, this port is not used since all the web interface is handled directly by the Elastix server.

Internal and External IP Address

If the PBXMate is installed on a server that has both internal IP and external one, it is recommended to force the PBXMate to bind to the external IP. This is done by modifying the binding parameter in pbxmate.conf file:

SipUAIPForBinding=<external_ip>

Don’t forget to restart PBXMate after you have changed the configuration file.

Symmetric Response Routing

In some cases, not all the equipment you are using fully complies with this extension to the SIP standard. If this is the case, in your network, the audio might not be heard by remote phone. In such a case modify the following parameter in the configuration file:

UseConnectionSessionDescriptorForFarEndMediaAddress=0

Don’t forget to restart PBXMate after you have changed the configuration file.

Sip Debug

If after implementing the above, there are unresolved issues,  activate SipDebug in the configuration file. This is done by setting

SipDebug=1

After the PBXMate is restarted, all SIP messages (incoming & outgoing) will be recorded to the errLog file that is located in the log subdirectory. Please send this file together with the configuration file to support.

In the last couple of years the percentages of employees that work from home rapidly increases. There are three main reasons for this phenomenon: (1) Efficiency. No need to spend hours driving to/from the office and in many cases the amount of interrupts while at home is lower than at the office. (2) Cost. When the employees work from home the infrastructure/office cost for the company are dramatically reduced. (3) Technology. This is the key enabler for this phenomenon. In the last years robust communication lines and tools have evolved that enable to seamlessly bridge the geographic gap between the physical office and the virtual work place.

In this post we will cover the technology area.

Technology Enablers

Following are the top important technology improvements that enable affordable high quality video conferences:

• Worldwide adoption of the VoIP technology. This technology enables passing audio & video using standard network protocol. Many tools use VoIP to pass the call on the freely available Internet cloud.

• Sufficient Internet bandwidth in the “last mile”. In the last couple of years the Internet bandwidth that reaches most households in the world has increased dramatically. This is a critical factor especially for video calls which require a lot of bandwidth compared to the legacy audio calls.

• The cost of powerful computers (including mobile ones) and high quality web-cams was significantly reduced.

Although the technology is great and the usage of video conferences for business purpose has risen dramatically, there are still few issues remain that yet needs to be solved.

Echo and Feedback

Whenever one of the participants in the call is in hands-free environment, there is a risk of echo and feedback that might also ruin the quality of the call for all other participants. This can be caused by many factors, like:  software or hardware malfunction, low quality of devices/software, inappropriate/conflict of drivers etc. In some cases drastic measures like forcing “half-duplex” calls are deployed to minimize the risk for echo but these drastic measures reduce the overall quality of the call.

Background Noise

If one of the participants is located in a noisy environment, this will influence the quality of the call for all other participants. Noisy environment does not necessarily have to be outdoors it can as well be indoors – for example whenever children are  playing/talking/shouting near one of the participants in the call, the noisy will be heard by all other participants.

Is there a solution ?

SoliCall provides a variety of software solutions for echo cancellation and noise reduction. For example SoliCall Pro was designed to be the ‘Go To’ solution for PCs that are used in video conferences. In addition, SoliCall has recently teamed up with Jell Networks to provide tailored high quality solutions for business in the US.

Following our previous post on acoustic echo cancellation on flash calls, we got requests to improve audio quality also on conference calls that are based on flash technology. The case study that was described in the previous post dealt with audio quality for regular flash calls, and showed how to convert them to SIP in order to improve their quality. In this post we will show you how to utilize the great tools that exists in the SIP world in order to implement conference systems for flash-based calls.

Conferencing in SIP

SIP is the leading standard for VoIP communication world-wide. It has many high quality features for almost any requirement in telephony including the options to build high quality conferencing systems. In this post we will describe a very simple conference system that you can build in a click of a button. For simplicity we will continue the same case study that was described in the previous post.  As you remember, in this case study we used Elastix as our SIP PBX provider.

Conferencing with Elastix

Elastix, which uses Asterisk as its telephony infrastructure, give you the option to build conference rooms in a click of a button. All you need to do is to open the graphical web-based menu in Elastix and select PBX configuration

In the PBX configuration go to “conferences” sub-menu.

In this section you can define conference rooms that enables everyone dialing to them to automatically share their audio. As you remember from the previous post, the audio is routed to Elastix via PBXMate therefore, it will be free of acoustic echo and free of disturbing background noise. In addition, the PBXMate will take care to adjust the audio volume to a comfortable level.

Conclusion

Once you have access to the standard SIP in your VoIP system you have endless options to implement your telephony requirements. For example, if the above conference system is not powerful enough to your needs or you are looking for additional features, there are plenty of other solutions out there. If you need specific guidance, do not hesitate to contact us.

Background

In a previous post we reviewed the basics of installing of PBXMate with Elastix. In this post we will review how to configure your phones to work with PBXMate. This post will discuss the configuration for VoIP phones.  In a future post we will review the configuration required for analogue phones.

Architecture

Currently your phones are configured to send SIP messages to the Elastix server on port 5060.

After installing the PBXMate, it  acts as a SIP Proxy on port 5061.  It means that in order for a phone to work with PBXMate, the Phone’s SIP Proxy port needs to be change to 5061.

Do not confuse this parameter with the phone’s internal sip port. The phone’s internal sip port has nothing to do with this architecture and there is no need to change it. What you should be looking for is a parameter called “Sip Proxy Port”.

And If there is no parameter “Sip Proxy Port” ?

On some phones, the parameter Sip Proxy Port does not appear explicitly and the only available parameter is “Sip Proxy”.  This parameter contains the IP address of the Elastix server. For example: 192.168.0.133

In such case all you need to do is to add “:5061″ at the end of the IP address. For example: 192.168.0.133:5061

What about Registrar information ?

If the Registrar information on the phone exists and is currently filled with the IP of the Elastix server, then you need to change it the same way as with the Sip Proxy. If there is a port field, change it to 5061, otherwise add “:5061″ at the end of the IP.

How do I know that I successfully connected the phone to PBXMate ?

Make a call to/from the phone and you should be able to see the call details in the statistics menu.

Continue Reading

If your network is distributed, you might also wand to read the following post on tips for installing PBXMate in a distributed environment.

One of the disturbing problems that appears when using intercoms, or other low-end full-duplex communication devices, is the generated echo.  Usually the echo is significant in these situations due to the following two reasons:

1. The intercom is equipped with an amplifier on its speakers that makes sure the voice is  heard even in a noisy environment. The down-side of this amplified is that it significantly increases the echo that is picked up by the microphone.

2. The intercom is usually a low-end device that has limited CPU and it is unable to do prover echo cancellation. Some intercoms do echo suppression but this method reduces the quality of the call by turning it to be half-duplex.

Centralized Echo Cancellation

The solution to this problem is to use a professional echo cancellation in the network that is capable of doing far-end echo cancellation on behalf of the intercom devices. A high quality centralized echo canceller should be easily capable of handling the long delay (echo tail) that exists in such environments.

PBXMate

PBXMate is a software product that is designed to do centralized acoustic echo cancellation. In addition, the PBXMate has many more features like:

1. Removing background noise.

2. Adjust call volume – for example amplify a low speech volume.

3. Real-time monitoring of call’s quality.

Seamless Integration

The PBXMate is designed to be easily integrated in any SIP network. Most modern communication networks have support for SIP. So, even if the intercom device and all other end-devices are analogue, the call can be easily converted to SIP via most modern PBX systems.

As recently announced, the PBXMate is now an Addon to Elastix. As a result, installing PBXMate with Elastix became a very simple and straight forward task. In this post we will perform this task step-by-step.

Install Elastix

The first thing you need is to have is an  active Elastix machine. If you do not have such a machine, go to the Elastix web site and download the latest stable ISO image.

You can install the ISO image either on a physical computer or on a virtual machine.

After installation, from your browser navigate to the Elastix machine and you will see the following welcome screen.

Navigate to the Addon menu

After you log-in, press on the arrow in the upper left side of the menu and select Addons.

In some cases, especially with older versions of Elastix, you might encounter difficulty opening the Addon screen. Also, in some cases the commercial Addons will not be listed. If you encounter any difficulty browsing the addons, we suggest to try the following two optional fixes:

Fix-1
On the Elastix console write the following two commands:

kill $(pidof python)
service httpd restart

Fix-2

On the Elastix server, create a file in /etc/yum.repos.d/ called commercial-addons.repo and paste the following:

[commercial-addons]
name=Commercial-Addons RPM Repository for Elastix
mirrorlist=http://mirror.elastix.org/?release=2&arch=$basearch&repo=commercial_addons
#baseurl=http://repo.elastix.org/elastix/2/commercial_addons/$basearch/
gpgcheck=1
enabled=1
gpgkey=http://repo.elastix.org/elastix/RPM-GPG-KEY-Elastix

[LowayResearch]
name=Loway Research Yum Repository
baseurl=http://yum.loway.ch/RPMS
gpgcheck=0
enabled=1

Afterwards, execute the following command:

yum clean all

At this point you can also install the PBXMate directly from the console by executing the following command:

yum install elastix-pbxmate

Install PBXMate

From the list of the Addons, select the PBXMate and push the “Trial” button to begin the installation.

Viewing PBXMate menu

Now, that the PBXMate was installed, you can see its menu under the PBX menu.

Using PBXMate

The PBXMate Addon for Elastix is pre-configured for seamless integration.  There is *NO* change required in the Elastix server.  You simply change the SIP Proxy on the phones you want to connect to PBXMate. We recommend reading the README section in the PBXMate menu.

From time to time we are getting requests to improve audio quality in flash calls and mainly to help remove echo. In a previous post, from 2011, we discussed few techniques that could be helpful. In this post we show a case study for integrating a centralized audio filtering to improve audio quality and perform server side echo cancellation. The ingredients for this solution combine several off-the-shelf products that integrate together using standard interfaces.

The Architecture

The system that we built has the following architecture:

Architecture

The Flashphoner Client is a flash based client. It could be replaced by other flash based clients.

The Wowza server is a standard streaming server.

The Flashphoner is responsible for translating the protocol of the streaming data to the standard SIP protocol.

The Elastix server is a well known unified communication server.

The PBXMate is an Elastix AddOn for audio filtering.

Installing Elastix Server

Installing the Elastix server is straight forward. All you need to do is to load a pre-configured ISO image. From the AddOn menu of Elastix, we installed the PBXMate. The following post provides more details on PBXMate addon installation.

After Elastix server is running, we used its simple GUI to define two SIP devices – we called them ”2” & “3”.

From the GUI of the PBXMate we enabled echo cancellation (AEC) filtering and restarted the PBXMate so the changes will take effect.

Elastix-PBXMate

Installing Wowza Server

We used a standard CentOS 5 32bit machine. On this machine, we downloaded and installed the Wowza RPM. We used an evaluation license for the Wowza server.

After we got the Wowza server running, we continued and installed the additional prerequisites for Flashponer: JVM, GCC & Make.

Afterwards, we installed on the server both the Flashphoner server & and Flashphoner client.  We used version 1.0.5.1523 of the server and version 1.0.258 of the client.
We installed an evaluation license for the Flashphoner server.

Preparing the Client Machines

We copied the Flashphoner client directory, that was created on the server, into two different client machines (our Client1 & Client2 machines). As explained in the Flashphoner documentation, we had to add the local Flashphoner client folders (including its subfolders) to the trusted locations of Flash otherwise the client did not load and got stuck in “Loading” state.  The change to the flash global security setting is done on macromedia site. Of course, this change needs to be done on both clients.

Making a Call

On each of the two client machines, we browsed to the flash client (PhoneJS.html) and pressed the Login button. The Login information that we provided was:

·         The Login & Auth name was “2” on the first machine and “3” on the second machine.
·         Password – is whatever we entered when defining these extensions in Elastix.
·         The Domain & Outbound Proxy is the IP of the Elastix machine.
·         The Port is 5061. If you use 5060, the call will be directed to Elastix without being filtered by PBXMate.

Client Setup

Monitoring Statistics

Once the call is established, we could see its statistics in the PBXMate tab in the Elastix GUI.

Statistics

Tips
·         Do not repeat this case study using the same machine for both Elastix & Wowza. We encountered port conflicts when trying to do so.
·         If the delay in your network is big and it is preventing from the AEC to properly remove the echo, you should increase the AEC delay parameter in the PBXMate Control GUI. Afterwards, don’t forget to restart PBXMate so the change will take effect.

EDITED – a later post was added to discuss high quality conferencing with flash-technology.