Ok…
It seems most of the obvious glitches have been addressed. Definitely more usable at first sight. Now I will have to run heavier tests and especially with louder sounds (as it’s much easier to reveal sound glitches…). I tested the record (obviously), re-record, next part while recording… everything seems to behave correctly, although I didn’t yet test extensively all the flows.
I have not been able to reproduce what is reported by @LordRadish
I’ve been trying with endless single note and with repeating single note… but I didn’t notice anything. Will retry later when I’ll be able to push a bit the volume.
I found that it really takes timing everything just right, when to fade, when to press overdub, play, etc. That’s why I think it needs a bit more xfade time. Interesting that you couldn’t get that glitch to happen.
This is not reproducible with these steps, the particularity of the case makes it hard to pinpoint a means of fixing, as of now we don’t see this as a reason to not release, please let us know if there is a specific step by step you notice, but on following your general gist, the pop is not audible.
Could you let us know the settings you have on when this happens? It could be even more particular than just timing and source material.
The most important feature is still missing. I am a live performer and it is critical to be able to get in and out of record mode as fast as possible.
It should be a simple fix: In record mode, tap the stop button (to stop the recording only, not the loop) to immediately cancel record mode and allow a new recording to begin right away.
This is a basic function that has been overlooked in an otherwise stellar product IMHO.
The loops are now really smooth, no more audio pop ! Wow ! Thank you guys
I can’t wait the next features (auto quantize, backing track importation, pan, fade out, locked tracks…)
Still didn’t perform any extensive tests, especially on UI and flows, but regarding my main concern which is how the sound is processed when pressing buttons, how loop seam is handled, this looks very good. I have not (yet) been able to defeat it !
This sounds like a major milestone and finally something that can become the foundation of all the crazy ideas exposed in this forum ! @DavidPackouz and all team: congrats !
Thanks for providing this upgrade. After installing it I tested twice today, all in 6x6 free form. Unfortuntely, I stumbled across two behaviors never seen before.
Issue 1 (drop outs)
Recording a beat box rhythm on track 1 (2 bars). Then pressing play
Recording a bass line on track 2 (4 bars). Then pressing play and … track 1 was “muted” for exactly 2 bars (=exactly one track 1 cycle) , then “unmuted” again. It looked like a dropout for exactly one cycle.
This occured not every time, however, I could reproduce it at least 10 times and it occured at different points of my song. So I left my house and came back to this in the evening.
Unfortunately, I could reproduce this and tried to collect some more details. Obviously, Aeros jumps temporarily into part 2 (part 2 indicator is on) even though I never touched the part switch button and song was reset before. And it comes back to part 1 automatically afterwards. I’ve video taped this.
It looked like “locked tracks” a bit Have you already started implementing code for this?
While listening and analyzing again and again I felt that something was wrong with the quantization. So I ran some tests with unproper end points of the recording cycle and came to this conclusion:
Issue 2 (quantization)
(Same scenario, same song as above)
Recording a beat box rhythm on track 1. Then pressing play
Recording a bass line on track 2. Then pressing play … obviously a bit too late (let’s say 100 ms)
Track 2 is played back with wrong timing for exactly one cycle. Seems like playback cycle 1 is delayed by these 100 ms. From the second cycle on it’s ok again (apparently as played while recording). Quantization seems not to work within the first cycle. I’ve never faced this with the previous release (or I was not aware at least). I’ve also video taped this.
Would you like me to upload the videos (high res mov). I’ve got one more for song and device setting. Where is the right place?
As to #1, we are not able to reproduce your issue with the information you provided, you could upload those videos to google photos and create a shareable link, that tends to be the easiest.
In #2 I see what you mean, this is really a corner case issue, we can definitely fix this, but ideally users will adopt the behavior of pressing before the events. We understand for users that are used to “on the beat” pressing that the loop forgiveness is useful, but it really is just a trick to perfect imperfect timing. That being said we have all the intention of fixing it! Thank you for bringing this up, we will plan on including this in the dev schedule.
Thank you for your kind words @LaurentB and @Eidknab - we’ve been obsessing for quite a while over getting this version as perfect as we can, so it’s very gratifying to hear you like it.
@popa we will investigate these issues you have reported.
In the meantime, we have decided that since most testers have found no serious issues, to release version 3.1.10 as an official non-beta firmware for download over WIFI. I have edited the original post to reflect this.
congrats ! 