[pulseaudio-tickets] [PulseAudio] #895: Naive method for mapping between integer and float sample is numrically instable

[PulseAudio]

#895: Naive method for mapping between integer and float sample is numrically
instable
--------------------------+-------------------------------------------------
  Reporter:  datenwolf    |       Owner:  lennart                                   
      Type:  enhancement  |      Status:  new                                       
 Milestone:               |   Component:  core                                      
Resolution:               |    Keywords:  sample format conversion numeric stability
--------------------------+-------------------------------------------------
Changes (by cladisch):

 * cc: clemens at ladisch.de (added)


Comment:

 The constraints that Kalle Rutanen used to derive these conversions are
 not actually requirements of !PulseAudio.

 In the original field, real-time graphics, integer values are usually
 quantized to 8 bits, and filters can be applied repeatedly to the same
 bitmap. Audio samples have higher precision (16 or nowadays mostly 24
 bits), and conversions are not done repeatedly, so I would estimate that
 numerical stability or bijectivity are not important at all.

 Are there specific quality problems in !PulseAudio's output that would be
 avoided by these conversions?

-- 
Ticket URL: <http://pulseaudio.org/ticket/895#comment:1>
PulseAudio <http://pulseaudio.org/>
The PulseAudio Sound Server