Reason Reaper Transient Detection

[Eclectrophonic]

Reason / Reaper Guitar Experiment

I recorded a 12 bar acoustic guitar chord sequence in Reaper.

I imported it to Reason. Reason detected 66 transients after removing the silence from the beginning and moving the clip so the first transient starts at 1.1.1.0.

With the first transient at the beginning, the 2nd transient Reason detects is at 0:00:00:496 / 1.2.25.

In Reaper, after removing the silence from the beginning and moving the clip so it starts at 0.00.000 and on the first transient, the 2nd transient detected was at 0.00.502 / 1.2.04 when I reduced the splits to 66 the same as Reason within the clip. Which you can do by clicking the clip, press D for dynamic split and then generating guides but putting 66 in the reduce splits box with it ticked.

Reason 2nd Transient - 0:00:00:496 / 1.2.25.
Reaper 2nd Transient - 0.00.502 / 1.2.04

I was trying to work out how Reason detects it's transients. Reaper detects the 2nd transient slightly later than Reason which doesn't make sense. If anyone can shed any light here. I'd be grateful.

[imrae]

As I understand it, the standard algorithm for detecting transients is:

• Track the RMS level averaged over a fairly small window of samples
• Also track the RMS level with a larger window of samples
• Compare the two levels; the further apart they are, the more "transient-y" the signal
• If that measure exceeds some threshold, the signal is considered "transient-y" enough to mark a transient.

So there are at least three tunable parameters (two rms windows and a threshold). I'd be surprised if two independent implementations did agree consistently to the nearest sample.

[Eclectrophonic]

As I understand it, the standard algorithm for detecting transients is:

• Track the RMS level averaged over a fairly small window of samples
• Also track the RMS level with a larger window of samples
• Compare the two levels; the further apart they are, the more "transient-y" the signal
• If that measure exceeds some threshold, the signal is considered "transient-y" enough to mark a transient.

So there are at least three tunable parameters (two rms windows and a threshold). I'd be surprised if two independent implementations did agree consistently to the nearest sample.

Interesting.

[antic604]

Reason 2nd Transient - 0:00:00:496 / 1.2.25.
Reaper 2nd Transient - 0.00.502 / 1.2.04

This is weird. Why is Reason sooner in time, but later in beats? Are both DAWs set to the same tempo and both have time stretching disabled (or both enabled)?

[Eclectrophonic]

Reason 2nd Transient - 0:00:00:496 / 1.2.25.
Reaper 2nd Transient - 0.00.502 / 1.2.04

This is weird. Why is Reason sooner in time, but later in beats? Are both DAWs set to the same tempo and both have time stretching disabled (or both enabled)?

Time-stretching shouldn't be a factor as neither clips were time-stretched. Yes both were at the same tempo. I just recorded in Reaper then exported, set the tempo the same in Reason and imported. The difference must be what I've chopped off (the silence at the beginning of the clips) but I went right up to the start of the first transient on both but obviously detection must be slightly different.

[antic604]

Reason 2nd Transient - 0:00:00:496 / 1.2.25.
Reaper 2nd Transient - 0.00.502 / 1.2.04

This is weird. Why is Reason sooner in time, but later in beats? Are both DAWs set to the same tempo and both have time stretching disabled (or both enabled)?

Time-stretching shouldn't be a factor as neither clips were time-stretched. Yes both were at the same tempo. I just recorded in Reaper then exported, set the tempo the same in Reason and imported. The difference must be what I've chopped off (the silence at the beginning of the clips) but I went right up to the start of the first transient on both but obviously detection must be slightly different.

But that doesn't sound right. If both DAWs were set at the same tempo, then there's no way that 1.2.04 is AFTER 1.2.25 when represented in time domain.