As a producer of music, on hold and voice-over packages for telephone, IVR and on hold marketing, I am very aware of the telephone greeting services and on hold music that I hear on other companies phone systems. I know this sounds rather geeky, but I actually like being placed on hold for a short period of time when calling companies. This period of time gives me a chance to analyze the quality of their music, on-hold audio and the quality of their voice-overs.
What I tend to find is that many on hold sections are distorted, too quiet or too loud. Now, you could just blame the supplier of the music on hold package for their lack of quality. However, as a provider of these packages, Media Music Now are experts in audio production and audio quality, but we do not know much about telephony hardware. Likewise, it seems that providers of telephony hardware know little about audio production or audio file formats. The result is that very often we produce high quality on hold packages but are requested to convert them into an audio format that the hardware provider deems suitable. In many cases, this is fine as we can convert to many formats with no problem at all, but occasionally, we are asked to convert the on hold package to a file type that is actually so compressed that it distorts the quality of the music and on hold voice-over, before it is even added to the telephone system. Once added to the system, the audio is further degraded by the bandwidth limitations of the telephone network.
In our experience, it is better to put the highest quality file possible into the system to start with. By doing this, the only degradation that is introduced is the bandwidth limitations of the telephone network. We find that MP3s or 16 bit, 22 kHz wave files provide the best audio quality on the end of the telephone receiver.
Many telephony hardware providers will argue this point and say that the telephone network degrades the quality to four bits, 8kHz anyway and for this reason they don’t need a higher quality audio file. I would have to say I really disagree with this. In my opinion, it is like taping a really old tape on to another really old tape, it doesn’t sound very good at all.
My view is that some of the highly compressed file formats they request make the audio sound very distorted before it even goes on the telephone system. With the added distortion/limitations of the telephone line, the final audio that the listener hears can end up excessively crackly or distorted. In my opinion, if you are spending money to have a bespoke on hold music & voice over package produced professionally, the last thing you want is it to sound distorted and poor quality.
I firmly believe that putting a higher quality file in the telephone system to start with, makes the delivered quality to your caller much higher and much more acceptable to the ear.
In conclusion, start with good quality in your telephone system to avoid distortion, crackle, bad volumes levels and any other audio anomalies that cause a bad listener experience. Granted, the telephone system will limit the quality anyway, but having the sound the best it can be will always present a much more positive image of your company.
For information about Media Music Now’s music & voice-over on hold service go to
http://www.mediamusicnow.co.uk/voice-overs/on-hold.aspx
—
Read other music on hold production related articles
Further articles related to music use & audio production
By Lee Pritchard
You are correct. CPUs are getting better and better. I manage several PBXs that have 16 processors per server!
I’m not sure of all the PBXs and what formats they support, but I know Asterisk can support MP3s.
I believe the physical limitation is due to the design of the common telephone carrier infrastructure. A typical telephone call is carried across a Digital Signal 0 (DS0) channel which is 64 Kbits/s. (In North America, 24 of these DS0s are multiplexed together to make a T1.) The audio is digitized at an 8 kHz sample rate using 8-bit pulse-code modulation (PCM). So, that seems to be the logical limit to our telephone quality (8Kbits x 8 = 64Kbits).
I hope this helps…
Jeff
Your article saved my rear end this morning. Thanks!
Jerry
Hi Jeff, great to hear a hardware expert explain the reason for using such compressed files. I am familiar with the file type you mention as we have to supply it from time to time. My preference is to keep the audio at a higher quality, however, I can understand that the hardware can dictate what is needed.
May be improvements in CPU and server speed will change this in the future, what do you think!? Lately, I notice a lot of client’s request mp3’s and our standard format seems quite widely accepted (16bit mono 22k wave).
Regards
Lee
I would tend to agree with this except some PBXs have to downgrade the audio to play across their hardware platform. In doing this, the PBX’s CPU cycles spike. For a large call center with hundreds or perhaps thousands of calls per hour, this could cause the server to fail. I have direct experience with this type of failure. The format I work with is 8k Ulaw.