#171
- Created on2011-03-06 22:21
- Reported by
- StatusFixed
- PriorityNormal (3)
- Assigned to
- Milestone
- Plan Level-
- Due Date-
Mac has paInputOverflowed set on most every callback
The Mac Code implementation is setting paInputOverflowed for almost every callback even though the input is clean. This seems like a false alarm.
The exception is when run at 44100 Hz where it only happens once or none at all.
It also happens with blocking Pa_ReadStream().
This can be seen easily in the paloopback test.
The exception is when run at 44100 Hz where it only happens once or none at all.
It also happens with blocking Pa_ReadStream().
This can be seen easily in the paloopback test.
************ Mode = Two Streams (Half Duplex) ************
|- requested -|- measured ------------------------------
|-sRate-|-fr/buf-|- over/under/total for in,out -|- frm/buf -|-latency-|- channel results -
| 8000 | 0 | 150/ 0/ 150, 0/ 0/ 152 | 64- 64 | 19.38 | OK
| 11025 | 0 | 207/ 0/ 207, 0/ 0/ 208 | 64- 64 | 9.43 | OK
| 16000 | 0 | 300/ 0/ 300, 0/ 0/ 238 | 64- 81 | 8.75 | OK
| 22050 | 0 | 353/ 0/ 353, 0/ 0/ 238 | 75- 112 | 9.34 | OK
| 32000 | 0 | 356/ 0/ 356, 0/ 0/ 237 | 108- 163 | 10.88 | OK
| 44100 | 0 | 0/ 0/ 353, 0/ 0/ 237 | 150- 224 | 12.11 | OK
| 48000 | 0 | 325/ 0/ 354, 0/ 0/ 238 | 163- 244 | 14.08 | OK
| 96000 | 0 | 163/ 0/ 354, 0/ 0/ 237 | 326- 489 | 24.72 | OK
| 44100 | 0 | 0/ 0/ 353, 0/ 0/ 237 | 150- 224 | 12.09 | OK
| 44100 | 16 | 1/ 0/3308, 0/ 0/3315 | 16- 16 | 0.18 | OK
| 44100 | 32 | 0/ 0/1654, 0/ 0/1654 | 32- 32 | 3.54 | OK
| 44100 | 40 | 1/ 0/1323, 0/ 0/1326 | 40- 40 | 1.52 | OK
| 44100 | 64 | 0/ 0/ 827, 0/ 0/ 827 | 64- 64 | 4.83 | OK
| 44100 | 100 | 1/ 0/ 530, 0/ 0/ 531 | 100- 100 | 4.10 | OK
| 44100 | 128 | 0/ 0/ 414, 0/ 0/ 414 | 128- 128 | 7.78 | OK
| 44100 | 256 | 0/ 0/ 207, 0/ 0/ 207 | 256- 256 | 13.63 | OK
| 44100 | 512 | 0/ 0/ 104, 0/ 0/ 105 | 512- 512 | 25.24 | OK
| 44100 | 1024 | 0/ 0/ 52, 0/ 0/ 53 | 1024-1024 | 48.41 | OK
************ Mode = One Stream (Full Duplex) ************
|- requested -|- measured ------------------------------
|-sRate-|-fr/buf-|- over/under/total for in,out -|- frm/buf -|-latency-|- channel results -
| 8000 | 0 | 150/ 0/ 150, 0/ 0/ 150 | 64- 64 | 20.38 | OK
| 11025 | 0 | 207/ 0/ 207, 0/ 0/ 207 | 64- 64 | 16.60 | OK
| 16000 | 0 | 238/ 0/ 238, 0/ 0/ 238 | 81- 81 | 16.75 | OK
| 22050 | 0 | 237/ 0/ 237, 0/ 0/ 237 | 112- 112 | 17.41 | OK
| 32000 | 0 | 236/ 0/ 236, 0/ 0/ 236 | 163- 163 | 21.81 | OK
| 44100 | 0 | 232/ 0/ 237, 0/ 0/ 237 | 224- 224 | 23.06 | OK
| 48000 | 0 | 236/ 1/ 237, 0/ 0/ 237 | 244- 244 | 25.83 | OK
| 96000 | 0 | 234/ 2/ 236, 0/ 0/ 236 | 489- 489 | 42.08 | OK
| 44100 | 0 | 220/ 0/ 237, 0/ 0/ 237 | 224- 224 | 22.99 | OK
| 44100 | 16 | 3306/ 0/3308, 0/ 0/3308 | 16- 16 | 8.39 | OK
| 44100 | 32 | 1571/ 0/1654, 0/ 0/1654 | 32- 32 | 7.55 | OK
| 44100 | 40 | 1316/ 0/1323, 0/ 0/1323 | 40- 40 | 8.25 | OK
| 44100 | 64 | 825/ 0/ 827, 0/ 0/ 827 | 64- 64 | 10.02 | OK
| 44100 | 100 | 524/ 0/ 530, 0/ 0/ 530 | 100- 100 | 14.58 | OK
| 44100 | 128 | 410/ 0/ 414, 0/ 0/ 414 | 128- 128 | 16.44 | OK
| 44100 | 256 | 206/ 0/ 207, 0/ 0/ 207 | 256- 256 | 25.19 | OK
| 44100 | 512 | 100/ 0/ 104, 0/ 0/ 104 | 512- 512 | 48.46 | OK
| 44100 | 1024 | 48/ 0/ 52, 0/ 0/ 52 | 1024-1024 | 94.85 | OK
Leave a comment
on 2012-06-22 16:15 *
By bejayoharen
I could be wrong, but I believe this is from CoreAudio unreliably reporting the xrun status.
This turned out to be caused by code that was confused by the return value from PaUtil_WriteRingBuffer(). It returns a count in elements not bytes. But the code was comparing the result to a byte count and flagging a paInputOverflow when they were not equal.
Fixed in SVN rev 1842.
Fixed in SVN rev 1842.