Advice on Voice Recording

[Flint]

I need to record some voice messages for a “hold music” device on a phone system. I need to do this on a shoestring budget, but I am concerned about the audio quality.

The device can import mp3, wav, ogg, or wma files. It converts them to 64kbps ogg files and stores them on flash memory. There is a “music” queue and a “messages” queue which are crossfaded, so the audio level needs to be consistent.

My basic idea is using a mic headset (which I do have, and was intended for use with speech recognition software) plugged into a PC (does it need a good sound card?) and record into...some free software? Do I need standalone software? I think Windows records wav files...

For recording voice messages, can I expect a decent result? Or else, what would you recommend?

[Clodfobble]

Download Audacity. It can record the signal off your mic jack in the back of your PC. For something that will be downgraded to phone system quality, your standard audio card should be perfectly fine.

But seriously, assuming it's not some 1,000 word script, just send it to me. I'm going to be recording for a client in the next few days, I'll just do it while my system's all set up, and email you the files. Easy peasy.

[jimhelm]

and her voice is just like they sound on TV. you'll get promoted!

[footfootfoot]

Originally Posted by jimhelm

and her voice is just like they sound on TV. you'll get weak in the knees!

FTFY

[Flint]

Originally Posted by Clodfobble

Download Audacity. It can record the signal off your mic jack in the back of your PC. For something that will be downgraded to phone system quality, your standard audio card should be perfectly fine.

But seriously, assuming it's not some 1,000 word script, just send it to me. I'm going to be recording for a client in the next few days, I'll just do it while my system's all set up, and email you the files. Easy peasy.

Wow, that is a generous offer. Thank you. In these cases I always think to myself I'd have to figure up the salary and pay it to the Cellar tip jar.

Unfortunately I don't have any kind of script ready nor do I expect this could possibly happen soon. I'm just the IT guy, if you know what I mean.

Thanks for the tip on Audacity--the CNET review sounds great. And it's free!

[Clodfobble]

Originally Posted by Flint

Wow, that is a generous offer. Thank you. In these cases I always think to myself I'd have to figure up the salary and pay it to the Cellar tip jar.

Pah, it's a phone script, it'll take like 5 minutes. Setting up the system takes longer than that. As long as you piggy-back on some other existing session, it's nothing. Just let me know if you ever get a script.

[xoxoxoBruce]

Originally Posted by Flint

I need to record some voice messages for a “hold music” device on a phone system. I need to do this on a shoestring budget, but I am concerned about the audio quality.

What? Are you trying to be the first person in history to actually give a shit about the caller being jarred by multi-volume recordings? You're such a rebel.

[Undertoad]

Whatever you do, don't use the words "please listen closely as our options have changed."

[Flint]

Originally Posted by Undertoad

Whatever you do, don't use the words "please listen closely as our options have changed."

Wondermark #651; Those Volatile Menu Options

[beauregaardhooligan]

I'm a big fan of CDWaveEditor.
http://www.milosoftware.com/en/index...ody=cdwave.php

[xoxoxoBruce]

That sounds like a powerful tool, beauregaardhooligan.

[beauregaardhooligan]

It is amazingly simple to use. I use it to produce digital files from analog sources. Just plug in a microphone, your record player or cassette deck and save those goldenoldies..
It only reads .wav files, but can save to several different formats including mp3, org vorbis, and flac, each at several different resolutions. I have even changed mp3s into a .wav file just by changing the file extension to make some edits and it works just fine.
The graphical representation of the .wav file is very easy to understand and lets you go right to the spot you want to put a track marker. No worries about the zero point and so forth. Place the marker where the wave starts to get bigger to indicate the start of the music or whatever and hit play to start from that spot. You can chop the file up into as many tracks as you want, then uncheck the box on the left and save just the checked boxes in whatever format you desire. It's non-destructive, so it won't change the original file.
It's so simple that it doesn't even have a record level volume control. You have to modulate it before it goes into the recording process. There is an overlimit indicator, but it seems to be very forgiving.
And, it's free!...although donations are encouraged.

[beauregaardhooligan]

Flint, the quality will all be in the original source so find the best mic you can, or at least one you're happy with. The higher resolution the better, though, because you never know what will step on it once it gets in the system.
Your headset mic plugged into a "mic" input may be satisfactory, but it would probably be better in a "line in" input. The sound card in your 'puter should be sufficient for voice recording.
Right click on the volume icon to open up your mic, and other record levels.

[smoothmoniker]

+1 for Audacity.

If you're trying to ensure consistent volume levels, there are two tools that you should familiarize yourself with.

The first is a plugin called a "compressor" that should come included with Audicity. It reduces dynamic range by a fixed ratio. For example, if you set the ratio at 6:1, then if the volume of your voice goes up by 6 dB, the output volume from the compressor would only go up 1dB. The end result is a more consistent dynamic range, without big spikes from loud syllables. You'll have to adjust the settings based on your particular recording, but a good starting place for speech is a threshold of -20 dB, a ratio of 9:1, and the fastest attack and release time possible, if the plugin has those settings.

The second tool is something called "normalize". This takes a digital audio file, scans it for the loudest point in the file, then raises the overall level to that the loudest point is at the maximum volume for digital audio. In other words, after you've recorded and compressed the audio, if the loudest point in your file is at -9 dB, normalizing will raise the volume of the entire audio file by 9 dB.

Most of all, trust your ears. What you hear is what it actually sounds like, and we can all tell the difference between something sounding good and bad.

[smoothmoniker]

Just to clarify, in case it was confusing, in digital audio all volume levels are measured as negative decibels away from the maximum possible volume of zero.