On 28 July 2013 18:02, Gale Andrews <gale@...> wrote:
> | From Steve the Fiddle <stevethefiddle@...>
> | Sat, 27 Jul 2013 18:43:37 +0100
> | Subject: [Audacity-devel] Append record - Poor error message.
>> On 27 July 2013 15:56, Gale Andrews <gale@...> wrote:
>> > | From James Crook <crookj@...>
>> > | Sat, 27 Jul 2013 14:22:13 +0100
>> > | Subject: [Audacity-devel] Append record - Poor error message.
>> >> On 27/07/2013 12:39, Gale Andrews wrote:
>> >> > | From James Crook <crookj@...>
>> >> > | Sat, 27 Jul 2013 11:57:13 +0100
>> >> > | Subject: [Audacity-devel] Append record - Poor error message.
>> >> >> I think the correct behaviour would be to just append record a stereo
>> >> >> track to the first stereo track.
>> >> > That was what the Sven's "Recording Profiles" patches did, recording
>> >> > to the upper stereo track and padding the lower stereo track with
>> >> > silence.
>> >> Padding with silence is wrong. I'm -1 on that. We should just add to
>> >> the channel(s) we're appending to.
>> > I don't have such a strong view on that. In the case you gave of
>> > two stereo tracks and append recording in stereo, I think you might
>> > expect not to pad the lower stereo track.
>> > If append recording in mono to two selected mono tracks, current
>> > Audacity records to both tracks. That doesn't make sense to me.
>> > I much prefer Sven's solution to record to the top channel only
>> > then pad the lower channel (so if you make stereo, you don't
>> > have unequal length tracks).
>> > Also if you can record four channels and choose to append record
>> > in stereo to two selected stereo tracks, the padding may be useful.
>> >> > I'm OK with that as long as it does "the right thing" if the device
>> >> > can record the requested number of channels.
>> >> I'm -1 on that.
>> >> We should only change the number of channels being recorded when
>> >> the user changes that in preferences->devices->recording->channels.
>> > If you have two selected stereo tracks and choose to append
>> > record four channels, you should get four channels.
>> > I'm more ambivalent than you if user has two selected stereo
>> > tracks, and can record four channels but the selector says two
>> > channels. The choice is to record only into the upper track, or
>> > record into both tracks, or even to record into the upper track
>> > and pad the lower with silence.
>> > We already disobey the number of channels selector (in one
>> > sense) if the user has a stereo track but append records in
>> > mono. The mono recording is duplicated to both channels.
>> > Perhaps the default should be as you say, but we should not
>> > rule out more exotic options if and when Audacity can allocate
>> > multiple recorded channels to specific tracks.
>> >> > I think this problem mostly happens when user has multiple tracks
>> >> > but none of them are selected. In that case Audacity is already
>> >> > interpreting the best thing to do (currently, by trying to append
>> >> > record to all the tracks).
>> >> Wrong best-thing. Better best-thing would be append to a track that
>> >> has right number of channels, or start new track if there was none. [The
>> >> clip on the new track need not start at time zero].
>> >> > Whether Record should leave the track unselected is another
>> >> > question. I would find it more useful for the track to be selected
>> >> > once it starts to record (as if you hit ENTER straight after hitting
>> >> > Record).
>> >> I think selecting a track that is recording makes sense too. Doesn't get
>> >> to the root of the problem at all, as I can still unselect after
>> >> recording and then do append-record. The underlying issue of trying to
>> >> record wrong number of channels when nothing is selected needs to be
>> >> addressed.
>> Unless this is moving toward support for channel mapping, I'd vote for
>> not doing anything too radical or too deep at this time.
> Generally, +1. Though I still think channel mapping will need
> flexibility in what is "allowed".
>> Currently we allow appending to a mono track even if Audacity is set
>> to record stereo. I'd not really want to lose that "feature".
> I think for now if all the tracks are unselected it, would be OK to
> only record the upper track(s) as needed, up to the maximum
> channels chosen in the selector, but don't discard reasonable
> "stereo" behaviours we have now:
> * append record to a mono track if Audacity is set to record
> * append record to a stereo track if Audacity is set to record
> I still think recording to both mono tracks when the selector is on
> mono is "wrong", but if it arises from the above, let it be.
>> I'm tending toward "best behaviour" being that if tracks are selected,
>> then "Record" does "Append Record" if possible, and throws a
>> meaningful error if Audacity can't do that.
> On reflection, I'll be zero on that, rather than -1. We're already not
> entirely obeying the number of channels indicator, so could just
> record to the upper stereo track if asked to append record in
> stereo to two selected stereo tracks.
> I don't think we should deselect tracks not being append recorded.
>> If no tracks are selected then new track(s) are created according to
>> the number of recording channels set. This is close to what we do for
>> generating audio
> -1 here. We have two distinct record actions, so I feel we should
> honour append record in some way if it's requested. Recording
> a new track will never be what the user intended. Recording
> to the uppermost track may be.
I'm not suggesting that we do this now, but I think that developing
channel mapping in the future may benefit from a fresh look at the
question of Record vs Append Record. In that context I'm suggesting
that perhaps Record and Append Record need not be two distinct things.
As an analogy, we do not have a separate "Append Generate".
The separation that we currently have between Record and Append Record
is something that we have become accustomed to, but given the number
of support queries regarding "how to record onto an existing track" it
is certainly not an obvious distinction to many users. I'm suggesting
that abandoning the idea of a separate "Append Record" function could
benefit users by providing the same functionality in a context
sensitive way - that is: If a track is selected when recording starts,
the selected track is appended. If no tracks are selected, then a new
track is created. There would be no need for a separate "Append
Record" function. It would certainly make appending a recording far
more discoverable than the current "two distinct recording actions"
that we currently have.
> I don't think we should even consider it unless we select tracks
> when we record them, because Append Record would be replaced
> by standard record too often in the common case of "record a
> track, then append record to it".
>> I think it would be useful if generators also used the Device Toolbar
>> "input channels" setting.
> +1 to that.
>> > Plus user error of selecting more tracks than the channels they are
>> > trying to append-record.
>> > Gale