View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0002485 | ardour | bugs | public | 2008-12-08 18:30 | 2020-04-19 20:13 |
| Reporter | peppot | Assigned To | paul | ||
| Priority | normal | Severity | major | Reproducibility | sometimes |
| Status | closed | Resolution | fixed | ||
| Product Version | SVN/2.0-ongoing | ||||
| Summary | 0002485: Transpose and TimeFX results in clipped waveform | ||||
| Description | This is not absolutely reproducible. It happens often in a session that has quite a few tracks and regions. When Transposing (F5) or doing TimeFX on a region, the resulting audio is severely clipped. | ||||
| Tags | No tags attached. | ||||
|
|
When enabling "Preserve Formants", the recording becomes entirely/completely clipped, and results in the deadly and horrible super-klick when starting playback. Calling cannam -- what's up here? |
|
|
Is the 'perserve format' option making it consistently reproducable for you on a specific session? Seablade |
|
|
Incidently when trying to reproduce it here I got some very cool results completely accidentally. Sadly no bug, but happily new ways to morph sounds;) Seablade |
|
|
It's reproducible in a test session I've now made. Recipe: normalized region, do a tranpose of 1 half tone up. Result: totally clipped region. Attaching video and session. |
|
2009-01-28 08:38
|
|
|
2009-01-28 08:39
|
|
|
|
does it have to do with the prior normalization ? |
|
|
Would seem so |
|
|
I can reproduce it here, but it is definitely not a bug in Rubber Band -- if you look at the audio files in the session directory, which reflect Rubber Band's output, none of them is clipped (and their levels are all broadly the same). I also checked the input/output RMS levels to the stretcher object, and they're fine there. Something in Ardour must be fiddling with the region gain. I'm looking around to see if I can spot what it is. |
|
|
Ah, hang on. The source region has gain and/or a gain curve. RBEffect reads from the source region, so it has the gain already applied to its input data. It then saves out a new file which is used as the source file for a new region copied from the old one. The new region therefore inherits the gain &c from the old region -- so the gain is now applied twice, once baked-in to the audio file, and once by the new region object. (And, of course, all of the audio files in the test session apart from the one original have all been through rb_effect one way or another already. So the fact that they have similar amplitude profiles to the region shown on-screen just shows that they have the gain applied already...?) Presumably the filter object needs to use a region constructor that discards the gain values and uses a fixed gain of 1 -- meaning gain changes can no longer be reverted after a stretch or transpose, but perhaps that's reasonable anyway. I don't know whether this would be appropriate for all filters however. Definitely one for Paul, or someone else who understands all this better than I do, however. And of course, I might be totally wrong about all this. |
|
|
2.0-ongoing rev 4888 has a fix for this based on Chris' observation/guess - when reading from the master sources we now disable any use of existing gain (including fade in/out). confirmed fixed (as best as we can determine for now) by reporter. |
|
|
Issue has been closed automatically, by Trigger Close Plugin. Feel free to re-open with additional information if you think the issue is not resolved. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2008-12-08 18:30 | peppot | New Issue | |
| 2008-12-08 18:31 | peppot | Note Added: 0005506 | |
| 2009-01-28 01:59 | seablade | Note Added: 0005649 | |
| 2009-01-28 01:59 | seablade | Status | new => acknowledged |
| 2009-01-28 02:02 | seablade | Note Added: 0005650 | |
| 2009-01-28 08:38 | peppot | Note Added: 0005669 | |
| 2009-01-28 08:38 | peppot | File Added: transposebug.ogv | |
| 2009-01-28 08:39 | peppot | File Added: transposebug.tar.bz2 | |
| 2009-01-28 11:35 | thorgal | Note Added: 0005671 | |
| 2009-01-28 11:54 | peppot | Note Added: 0005672 | |
| 2009-01-28 13:49 | cannam | Note Added: 0005674 | |
| 2009-01-28 13:50 | cannam | Note Edited: 0005674 | |
| 2009-01-28 13:51 | cannam | Note Edited: 0005674 | |
| 2009-01-28 13:59 | cannam | Note Added: 0005675 | |
| 2009-01-28 14:02 | cannam | Note Edited: 0005675 | |
| 2009-03-24 14:04 | paul | cost | => 0.00 |
| 2009-03-24 14:04 | paul | Status | acknowledged => resolved |
| 2009-03-24 14:04 | paul | Resolution | open => fixed |
| 2009-03-24 14:04 | paul | Assigned To | => paul |
| 2009-03-24 14:04 | paul | Note Added: 0005843 | |
| 2010-04-24 10:28 | cth103 | Category | bugs => bugs2 |
| 2010-04-24 10:31 | cth103 | Category | bugs2 => bugs |
| 2020-04-19 20:13 | system | Note Added: 0021837 | |
| 2020-04-19 20:13 | system | Status | resolved => closed |
