0

I have a question regarding sound quality - its a question about whether I should choose betwteen two strategies where the purpose is to preserver good sound quality at the one hand - and to reduce memory occupation of the soundfile on the other.

Until now I have converted the actual sound-file from wave-format to a mp3-format. And the sampling-rate is 44000 HZ.

This is one way - and I certainly loose some quality when making this conversion.

The other way I have been thinking of lately is to NOT compress - instead of compressing the wave-file to mp3 I lower the samplingrate to 22000 HZ and in that keep "good quality" and at the same time reduce the size a lot.

Is it a drawback with this strategy? Which strategy is the best?

The soundfiles is used to an app I am developing in android. The soundfiles is phrases - a man and a womans voice - no longer than 2-5 seconds.

thanks in advance!!!!

2 Answers 2

3

If you down sample sounds as Stavrosound has said you need to be wary of the Nyquist limit. Roughly the sound must be sampled at twice it's frequency to be accurately represented.

In game sound people will down sample making sure that the majority of the useful data is below the limit, eg. If the audio has no useful info above 4k then the sound can be down sampled to 8k, although compromises are often made to quality at a loss to some high end vs size dependant on how key the sound is in the overall mix. So whether or not to down sample is down to the Audio's application, it's frequency content and it's importance on the overall mix. Often low end sounds are down sampled significantly with little compromise vs something like a cricket sound which has plenty in the high end to be obviously lost.

2
  • 1
    To expand on that, downsampling from 44kHz to 22kHz will remove all frequency content above 10kHz (rounded, it depends on the antialiasing filter used).
    – EMV
    Jan 16, 2014 at 17:35
  • @Rory you mis-quoted me. The Nyquis limit is why I said MP3 is the better option. I did not suggest the OP choose to downsample Jan 21, 2014 at 22:57
1

Drawback = Nyquist limit. MP3 does psychoacustically-pleasing compression, sampling rate will truncate the data. The former is the better of the choice.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.