Discrepancy in zoomed display of the region and of cuts

If that’s what you wanted, then I don’t see any technical reason why not, though I could also see complaints that you could not then (with “Snap To: Enabled”) place a point label at sample time (unless we have an additional user option for “Snap To: Sample Time” / “Snap To: Sample Period” )

Yes, that is the major advantage, it looks a lot neater, provided that we are dealing with just one sample rate and Snap To: is enabled.
Of course, with Snap To: disabled, there could still be doglegs because the audio track “snaps” to the mid-point of a sample period (as now marked by the dot) but the Timeline/label track is not snapping.

There would also still be doglegs between adjacent audio tracks that have different sample rates.

Yes, that is the major disadvantage. I expect that most users would assume that “the dot is the sample”.

Another possibility would be to show (when zoomed in to sample level) both the “sample time” (as a dot) and the “sample period” (as a line). This could be with or without the (fictitious) “wave line” (linear extrapolation between samples).

For clarity I’ve illustrated this below without the “wave line”.

This first image illustrates “Sample Times” and “Sample Periods” with “Snap To:” snapping to the middle of the sample period.
This second image illustrates how doglegs can still occur when “Snap To:” is off (disabled).
In this image I have also marked the “mid sample-periods” with yellow dots.

The selection in the audio track always “snaps” to the mid sample-period position (marked here with a yellow dot).

Split, Cut, Delete, occur at the edge of the selection (the yellow dot) in the audio track (as now).
There are of course questions about how efficiently this could be drawn, but given that it is only drawn when zoomed in to sample level I would not expect that to be a problem.

If we also had the “wave line”, should that be drawn through the start of the sample period, or the middle of the sample period? (the “wave line” is not “real” - it is just a graphic extrapolation between points, intended to give a rough idea that the points are converted to a continuous waveform when converted to an analogue signal by the sound card).

This is what it would look like with Snap To: enabled, and the dots marking the centre of the sample periods.
It “looks” very nice and neat, but I think it is misleading.
It definitely gives the impression that the sample times are where the dots are, but they are not.

If we also show the sample times as dots and the sample periods as dashes, then there is a lot more information to indicate what is happening, but it is perhaps a bit cluttered. How many users actually want or need this information?

Thanks, Steve.

If clicks in Audacity still snapped to sample times (i.e. in-between some representation of sample crossover periods) that may be preferable. I was guessing that may give more risk for off-by-one error? You don’t think so?

Clicks snap to the sample crossover points in CoolEdit, so there is precedent but it is not ideal unless it is visually clear the crossover points are not the sample times.

I’m not sure about having two different dots connected by a line - it does look a bit cluttered. I think we cannot use dots (however coloured) to indicate sample crossover points - they will be mistaken for sample times.

Doglegs with different track rates could be avoided by quantising to the lowest rate, I would guess.

Doglegs with Snap To off could be avoided if Timeline/label selections still snapped (in this suggested scheme) to sample crossover points. I’m not sure of the benefit of time selections being allowed outside audio selections (sample periods) unless user explicitly wants to snap time to other than samples. Can you see a benefit?


I think that an improvement on my previous post would be to mark the mid-point of sample regions with a dot, but when snapping to samples, to include the entire sample region. Here it is totally clear which samples are within the selection and which samples are not: