Talk:Export Multiple

Bill 26Sep10: Proposed fifth example of export multiple, with mixed point and region labels attempting to illustrate all the behaviours. I'll integrate this into the main page after we've determined that it is OK. I'll also re-do all the other illustrations in that section to match the look of this illustration, using 1.3.13.


 * A project prepared for Export Multiple with a mix of point and region labels:


 * [[image:ExportMultipleExample05.png]]


 * The first file will be named "First Song" and will include the audio spanned by the "First Song" region label. The "Second Song" point label abuts the end of the "First Song region label - thus the second file will be named "Second Song" and will include the audio from by the "Second Song" region label to the start of the "Third Song" region label. The third file will be named "Third Song" and will include the audio spanned by the "Third Song" region label. The audio from the end of the "Third Song" region label to the "Fifth Song" point label will not be exported. The next file will be named "Fifth Song" and will include the audio from the "Fifth Song" point label to the start of the "Sixth Song" region label. The rest of the files will be exported according to this logic. If the last label is a region label any audio beyond that label will not be exported. If the last label is a point label, the last file will contain all the audio from that label to the end of the project and will be named according to the last label in the label track.

A bit long-winded since it repeats stuff from the point and region label examples, but probably necessary to avoid confusion.

Steve the Fiddle: The second song label comes very close to the third song label region. Perhaps the illustration could have a longer second song? (shorter 3rd or 4th song). It may also be easier to read if the exported sections were in a list rather than continuous text. For example:

The exported regions will be:
 * The first file will be named "First Song" and will include the audio spanned by the "First Song" region label.
 * The "Second Song" point label abuts the end of the "First Song region label - thus the second file will be named "Second Song" and will include the audio from by the "Second Song" region label to the start of the "Third Song" region label.
 * The third file will be named "Third Song" and will include the audio spanned by the "Third Song" region label.
 * The audio from the end of the "Third Song" region label to the "Fifth Song" point label will not be exported because there is no label between the right edge of the region boundary for Song 3 and the point label for Song 5.
 * The next file will be named "Fifth Song" and will include the audio from the "Fifth Song" point label to the start of the "Sixth Song" region label.
 * The rest of the files will be exported according to this logic.
 * If the last label is a region label any audio beyond that label will not be exported.
 * If the last label is a point label, the last file will contain all the audio from that label to the end of the project and will be named according to the last label in the label track.

Gale: Thanks, Bill. I largely agree with Steve. I think the difference between the second song and the gap between the third and fifth isn't clear enough because of the long label of the second song.

Although it's less obvious with songs, people who mix labels tend to do it with speech or industrial samples or similar. I think we need an example of not exporting a whole song or unit of audio.

It might be better to start with explaining "what" we want to export from the image then we can say "why" the mix of region and point labels makes it easy to do this. E.g I added a note to Steve's list explaining why no audio between songs 3 and 5 is exported.

PS I can't see the sixth song without right-scrolling even on 1024 x 768. Maybe this would actually be better as custom audio rather than the same songs as the other examples?

Bill: OK, so what would a non-music example be, where it is convenient to mix labels? Here's my suggestion (admittedly artificial - if anyone can think of something better let me know and I'll try to mock it up). An interview track with a series of question-answer pairs. In most cases we want to export each complete question/answer pair as a separate file, so point labels work fine. But in one case we cant to cut off the end of an answer - perhaps the interviewee asked for a pause. So in that case a region label would work well.

Gale: Yes. anything like that. Another example that comes to mind is a podcast. You want to keep the intro and outro music, keep all the interviews except one, but not export any ads between the interviews. Thanks!

Gale: Thanks, Bill. I made some changes mainly because I think it's easier to describe the image that is actually there as far as possible and have it relate to a complete track. Would you object to uploading a new image with the final label renamed to "Outro" ? I have also numbered all the cases for ease of reference.

PS a minor point that there seems to be a lot of noise in the recording where one or the other is not speaking?

Bill: My intention was that this was part of the interview, (similar to the LP capture examples which show part of the LP). I can see your point, though, and I see you've changed the text. So I'll upload a new image.

The "noise" is the bleed into the opposite channel. Right now it's realistic. I could silence those sections which would be unrealistic but perhaps better for the example.

In the new last paragraph, are you talking about editing the segments in place, or exporting?

Gale: Thanks. Having an album with only five tracks would be a bit unrealistic, though not for a side. An interview with only a few questions and answers is a bit artificial, but not much if part of a longer program. I think the advantage of the users being able to see what the text refers to rather than have to visualise it outweighs any lack of realism.There is a lot to take in on this page.

Yes the silence too is "unrealistic" but the distinction between question and answer is much clearer now.

I've tried to make the second sentence of the final paragraph clearer, I think the first sentence is important, but if you think the second sentence is still confusing, we can just drop it.


 * Case 5: A mix of point and region labels:
 * [[image:ExportMultipleExample05a.png]]

In the case above we have an interview with a series of questions (upper track) and answers (lower track), and a concluding statement by the interviewer. We want to export the question-answer pairs and the interviewer's conclusion as separate files.

In most cases the question-answer pairs follow one after another, so point labels are a quick and convenient way to mark them. However after the third answer, the interviewer is told by the studio manager that time is running out, and the two parties discuss this. We want to omit this exchange.


 * We mark the start of the first two question-answer pairs with point labels "QA1" and QA2".
 * We mark the third pair with a region label "QA3" since we want to omit the discussion about running out of time.
 * We mark the start of the fourth pair with point label "QA4" and the start of the interviewer's conclusion with point label "Outro".

So this is what will be exported:
 * The first file will be named "QA1" and will include the audio between the "QA1" and "QA2" point labels.
 * The second file will be named "QA2" and will include the audio between the "QA2" point label and the start of the "QA3" region label.
 * The third file will be named "QA3" and will include the audio spanned by the "QA3" region label.
 * The fourth file will be named "QA4" and will include the audio between the "QA4" and "QA5" point labels. Note that the audio between the end of the "QA3" region label and the "QA4" point label is not exported.
 * The final file will be named "Outro" and will include the audio between the "Outro" label and the end. Had we decided to cut off the conclusion half way through, we would have selected what we wanted to keep and labeled that, producing a region label "Outro" instead. Then the audio beyond the "Outro" label would not have been exported.

Note that in order to export the pairs and the conclusion excluding the interruption, we could equally have drawn region labels for each, similarly to Case 4 above. Had we needed to edit each of these regions individually before export, it might have made sense to label them as regions in any case.