Testing/Workflow

Workflows
This page lists the typical workflow of some users who are testing the Audacity 2.0 RC.

The idea of listing the workflows is that we get to see the kind of use that is being covered by non-systematic testing. It shows, for example, that as of 18th February 2012, long recording, e.g. recording 8 hours of audio, is not being tested nor is the typical workflow testing ogg.

Mono vocals + backing to .MP3
Windows XP SP2.


 * Mono vocal, i.e spoken word, recording from microphone (USB Realtek), about 45 mins, some pause, resume, some new-track creation.
 * Noise removal.
 * Lots of cutting (mostly small chunks), lots of zooming in and out, some shunting around of audio to align different takes, use of (span rather than point) labels.
 * Mix down (1x vocal, 1x backing track).
 * Fade-ins and fade-outs,
 * saving project, and export to mp3.

Quote: "I'll also try some of the things which are recently changed and some of the things which I think likely to cause trouble."

Internet radio to .WAV
Windows XP-HE-SP3 500MB RAM Dell Dimension 2400 Audacity set to 441.kHz 32-bit flloat
 * 1) open webstream for internet radio capture (one hour show with 5 minute pre & post buffers)
 * 2) reset audacity to record from my onboard SoundMax rather than my external USB soundcard and turn off Software Playthrough
 * 3) Save the empty project to its assigned name/location
 * 4) zoom level set to show 30 minutes on screen
 * 5) mark approximate start points and end points of material I want to "keep" with point labels
 * 6) after capture, save the project
 * 7) trim the pre and post 5 minute buffers - save the project
 * 8) remove DC offset with Normalize (no amplification) - save the project
 * 9) zoom project to fit screen. select to nearly between each "end" marker and the next start marker.
 * 10) delete the unwanted stuff between markers with Ctrl+K
 * 11) repeated until all required material only remains
 * 12) Noise removal (internet feed "noise") if I can find a suitable sample
 * 13) switch to normal zoom level
 * 14) tab to "song" start
 * 15) trim cruft from front of song with Silence and Fade-In
 * 16) move label to proper start position
 * 17) Tab to "end" label marker for the song
 * 18) trim cruft from end of song with silence and Fde-out
 * 19) remove the "end" label
 * 20) repeated last 6 steps until all songs trimmed
 * 21) Amplify to -3.0dB
 * 22) Export Multiple as 44.1kHz 16-bit PCM stereo WAVs
 * 23) Delete Project and backup WAV files

LP to .WAV
My LP digitization workfow is very similar except that immediately after capture:
 * 1) I export a 32-bit WAV file of the whole project process that through Brian Davies' ClickRepair application and
 * 2) re-import the de-clicked WAV file into Audacity, uing "faster".
 * 3) Then delete the original capture track from the project and
 * 4) move the label track down below the imported audio track
 * When transcribing LPs I never bother to Save an Audacity project at all (the exported 32-bit WAV makes a good capture backup).

Testing user-experienced scenarios

 * launch and quit HEAD (and other versions of Audacity) dozens of times a day. At the moment HEAD is the 2.0 RC.
 * Lots of quick recordings and short tone generations;
 * lots of quick cuts and pastes, in the course of checking some user's specific report.
 * Also testing of a bug under active consideration on bugzilla.
 * Frequently changing language to test committed .po files or translation bugs.
 * (sometimes) looking at some very long possibly buggy project submitted by a user.

Stereo LP to .MP3
Mac 10.7.2 Intel and 10.5.8 PPC


 * Audacity set for 16 bit 44100
 * save empty project to a folder named for the artist
 * Capture an LP through my USB interface
 * amplify to -1 dB
 * export each LP side as aiff for processing in ClickRepair
 * import processed aiffs
 * amplify to -1 dB (now that clicks are gone can amplify more)
 * editing typical of James and Peter - zoom in and out, remove small bits of audio, fade out and silence between tracks
 * add point labels to mark songs
 * edit metadata
 * export multiple as aiff

Quote: "I typically go through similar for editing live recordings, except that the recordings are imported, not recorded."