As of right now I’m working on editing audio files in Audacity, which seems to be going fine for the most part. But i’ve been running into some problems that I don’t know how to solve. Most of my files are meant for e.g: answering machines or voicemail messages.
I’m currently working with the latest Audacity version, 2.1.1.
I prepare the files in Audacity for the customers (mostly only cutting & cleaning), but before these files are being sent to the customers, my colleague in Switzerland receives them so she can check it before delivering the files. My colleague uses Wavelab 5 (yes, an old version but still with lots of options). And she always opens my files in Wavelab to check. The following problem keeps occuring:
She tells me the file hasn’t been cleaned up ‘good enough’. In Audacity: I use the Noise Reduction tool or I generate a silence in places where there’s no talking. Audacity doesn’t show that there’s still any noise going on in the background of the files, yet in Wavelab it shows that the files are not completely cleaned up.
Are there better options in Audacity to remove background noices than Noise Reduction?
Please understand that I’m still fairly new to Audacity and I’m still trying figure everything out
Audacity doesn’t show that there’s still any noise going on in the background of the files, yet in Wavelab it shows that the files are not completely cleaned up.
The files shouldn’t sound any different when played in Wavelab, Audacity, or Windows Media Player, etc.
So, I think you should ask her exactly what (and where on the recording) the problems are. “Cleaned-up” would seem to relate to noise, but it’s a bit vague and she may mean something different from what you think she means…
Once you know what you are listening for, if you are not doing this already, try listening on headphones… You can often hear little noises and little defects/details better on headphones than on speakers. (Pros normally advise against mixing and editing with headphones as your main monitors, but they are good for certain specific things.)
She tells me the file hasn’t been cleaned up ‘good enough’.
She’s going to have to tell you a lot more than that. The words “clean up” in a forum posting is usually the kiss of death. That means the sound in the files is not recoverable.
Home recording for commercial application and sale isn’t easy. Posters who have troubles, for example, reading audiobooks usually go weeks and multiple forum chapters trying to fix their studios so that the Audacity tools have a fighting chance. Larger publishing companies have a robot which listens to the works and automatically flunks them if they fail general volume and noise. They never actually get to a human, so you got lucky.
Post about ten seconds of one of the “dirty” voice segments on the forum. Scroll down from a text entry window > Upload Attachment > Browse…
It will tell you if the clip is too big. If it’s in stereo (pretty unusual for voice) you will only get five seconds.
I have added a raw clip of about 10 seconds in this reply. This file is recorded and
after that there is nothing we did with it yet. Could you please be so kind to explain how
you would clean up the file in Optima Forma, in order to use the file on answering machines?
By the way, I am using Windows 7 64bit on my computer.
If your audience will hear via telephone, that only has a bandwidth of about 200Hz-4000Hz.
So IMO it’s best to apply a band-pass filter to simulate the bandwidth of the phone, then work from there to optimise the intelligibility.
If your audience will hear via telephone, that only has a bandwidth of about 200Hz-4000Hz.
But the client is a human and I bet they’re expecting a clean, clear recording. >
my colleague in Switzerland receives them so she can check it before delivering the files.
There is a tool called a noise-gate which will do that automatically.
But I’m betting not in this case. The performer has hissy words over a relatively clean background. I suspect that’s what your client was complaining about. You cleared all the sound between words by hand assuming that’s what the complaint was. Noise-Gate can be adjusted to do that automatically. But it didn’t work. The sound came back as ‘not clean enough’.
If you play the clip on a high quality sound system multiple times, you notice the words are not clear and crisp. On one or two words at the beginning, there’s actual hiss hang-over.
I suspect the performer already “noise removed” the clip, or they’re using a system that tries to noise remove the sound with no notice to the user. Windows is famous for this.
And that is why I wanted a very specific complaint. Not just “clean it up.” If actual word quality is the complaint, there is no fix that I know of. That has to be done at the performance stage.
There’s another level of this. The performer sounds really good, like she does this all the time. “Everybody knows” an announcer/presenter needs to noise remove work before submitting it to a client (you).
or compression damage : the “raw” file is in “wma” format , which is lossy-compression like mp3.
Perfectly correct and I was going to complain about that, too.
Whereas WAV (Microsoft, Audacity Default) is close to a universal, perfect, uncompressed sound format, Windows Media Audio is not. Macs will not play WMA sound files without additional software. I had to use VLC Player to play your clip. It’s a download program, not part of the Mac normal tool set.
WMA has sound compression damage and supports copy protection. Strike two and three.
