Hi,
Would greatly appreciate any recommendation on a software that can introduce these effects to speech files (.wav). The idea is to be able to simulate voice quality degradation for:
Absolute silence
Clicking
Crackling
Crosstalk
Hissing
Hum
Motor Sound
Popping
Static
Talker Echo
Tunnel Voice
Thanks a lot for any hint in advance!
Regards,
Vallu
Sevana Oy
http://www.sevana.fi
The problem with software that can simulate these conditions is the same problem people have making a highly controlled, orderly, precise digital system generate random numbers. Anything you can do with a digital system can easily be undone with a computer cranking long enough. Not so with chaos in an analog system which is what many if not all of those defects represent. It doesn’t matter how long you crank, you will never get rid of all the trash in a defective analog recording. You can get Close Enough for Jazz, and maybe produce a pleasant recording, but you will never fix all the original problems.
So anybody generating, for example, “white noise” with a digital system will soon fall to somebody else recognizing that the pattern in the noise happens to correspond to Krajniak’s Algorithm with a progressing exponent. And you’re dead. Undo the algorithm and you’re back to silence. Not so a recording of interstation hiss in an FM receiver. No pattern there at all.
I know what you’re doing. Scientists love to reduce things to measurable conditions and then draw conclusions from the results. The vitamin industry lives on such measurements. (Vitamin D is good for you. Vitamin D is bad for you.)
Even silence can be misleading. To what accuracy? 44100, 16-bit, Stereo? That’s the specification of Music CDs which was a compromise a million years ago when CDs were developed. Developers at the time knew it wasn’t even close to perfect, but they needed to get something out the door. That and complete silence doesn’t exist in the real world. The SPL measurement system is open ended on the bottom and more than one electronics system has fallen face-first in the mud from not taking that into account.
So what’s the job? Why do you want these simulations? Some of the existing tools in Audacity may get close enough for your Jazz.
Koz
Thank you for your comments! Well, I do realize it may be not a common feature… Audacity helps us very well with audio editing and analysis, and echo is available as well. However, our task is to be able to generate files with different type of degradation and maybe even different set of parameters in order to train our voice quality assessment software to detect these degradations in telecommunication systems. We are not afraid that approaches to achieve these effects may vary, the training algorithm we have is quite powerful and captures specific signal characteristics, so we think that no matter f.e. whether echo samples were made by Audacity or another software, we will detect echo presence and thus identify the main reason for audio quality degradation.
To take echo as one example. Your bathroom degenerates into infinite echoes and reflections very rapidly and no software package can hope to keep up either putting echo in or taking it out. That’s why echo generation is a religion. “My echoes sound much better than yours.”
You do have to be careful of tricks. White noise isn’t really white because we use tricks to make it “sound” white.
All that said, I would probably take a recorder out and record a slamming car door if I wanted that sound. No tricks and no software and algorithm juggling. You can probably use the basic Audacity sounds built-in. Silence, hiss, etc. Those are or should be close enough.
Power Hum. Red flag. Power Hum can’t be heard. Pure 60 and 50 are very close to the threshold of human hearing. Those are two large pipe organ tones. However what people perceive as power hum is really power damage creating tons of harmonics and intermodulation distortion. Really buzz. This is what people complain about when the shielding on their microphone goes open allowing motor noise and fluorescent lamp trash to enter.
Are you in Britain? Plan on 50 Hz and harmonics. In the US, plan on 60.
And I can tell you now what your challenge would be. All at once. A noisy distorted microphone with echo and white noise with a Metrobus starting up outside the window and comb effects from two parallel opposing walls.
Have a happy day.
Koz
The obvious answer would be to make “field recordings” of the various sounds, import them along with the voice recording into Audacity, then adjust the levels to suit.
I also think that that you already got what you need. And I don’t think you will find much better quality. (Unless you switch to hardware maybe…)