Page 2 of 2
Re: Trouble getting the software to record, windows 7 comput
Posted: Mon Mar 08, 2010 1:10 am
by whomper
stevethefiddle wrote:whomper wrote:
32 is totally unnecessary
32 bit processing give significant quality improvements during audio processing, particularly with low level signals.
By way of demonstration, try amplifying a 16 bit track by -50dB, then amplifying by +50dB. You will notice a significant deterioration in sound quality. Do the same in 32 bit and the result is virtually identical to the original. Processing in 32 bit float is also faster than 16 bit processing due to the architecture of modern computers.
whomper wrote:but default export to 16
The default Export format in Audacity 1.3.11 is 16 bit WAV.
whomper wrote:
and of course all should be manually overrideable in the options
Theses defaults may be overridden in Preferences.
24 bits should be enough
imho
there is headroom and there is poor recording technique
18 db is plenty - euro standard
i allow 24
50db is way too much
Re: Trouble getting the software to record, windows 7 comput
Posted: Mon Mar 08, 2010 7:53 am
by steve
If you are just doing a simple "record/trim/export/burn to CD", or "record/trim/export as MP3/dump on iPlayer", then 16 bit is quite sufficient. However, if you need bit perfect copies, or if you are doing a lot of processing and mixing, then 32 bit is the way to go. While the improvement in sound quality of 32 bit over 16 bit may be negligible in most cases there is no overhead penalty in increasing the bit depth from 24 bit to 32 bit, in fact quite the reverse - 32bit float calculations are performed faster than 24 bit or 16 bit calculations.
The 16 bit 44100Hz standard for CD audio was based on technical limitations that existed at the time that the standard was created. It is generally accepted that 16/44.1 is the minimum necessary for high fidelity audio, and there is no shortage of people that consider it to be too low.
Audacity handles all calculations internally as 32bit float calculations (for reasons of quality and efficiency). If the results of such calculations are required at a lower bit depth (16 or 24 bit) then the results must be down sampled to the track bit depth each time, resulting in a cumulative loss of sound quality.