2800 is pretty high, a lower value, like 2400 might be better for the speaker, and might also produce less false clashes.
In theory, 2800 should be OK for a 8ohm speakerā¦
True, but too high values tends to cause clipping, which can sound bad AND cause additional vibrations in the hilt.
Iāve uploaded version 6.4 of ProffieOS. Itās the same as 6.3, but the following bugfixes:
- subblade errors now says āerror in blade arrayā instead of crashing again
- Fixed a rare type inference bug in TrConcat
- ignite_time fix for fett264 prop
- fix for DISABLE_COLOR_CHANGE in fett263 prop
- Improved comment at top of ProffieOS.ino
- Serial monitor welcome message now includes ProffieOS version.
Most of these are of the type āif you donāt have a problem, there is no need to updateā, but the ignite_time fix may be a reason to upgrade if you use the fett263 prop.
Iāll download it this evening and try it out. Thanks for addressing the ignite timing w the prop file. It was weird trying to get used to it.
What ignite timing are you referring too? The fix in 6.4 is for Edit Mode Menu tied to a pretty low occurring scenario for rotating backwards through Retraction Options and getting a 32768 value.
1 Like
Just for reference when the OS6 library is done, v6.4 will be required tied to this fix. V6.3 will have compile errors on multiple styles in the upcoming library.
5 Likes
There seems to be a little more lag between scrolling through presets with aux and the ability to activate the saber after selection. Itās not the boot/font wav playing time. Nothing odd showing in serial monitor so I doubt itās the prop file.
Check your SD card speed, run āsdtestā in Serial Monitor. Laggy performance is usually from a slow SD, especially if itās on font changes. Just make sure you remove DISABLE_DIAGNOSTIC_COMMANDS define if youāre using.
1 Like
Hello all!
I decided to try out OS6.3 and wanted to report back. Iāve installed it on my SP Saber Darkness v3 crossguard and utilized FETT263ās prop file to play around with some of his new features.
I can confirm the responsive delayed ignitions with out wavs to match (up=simultaneous;down=staggered) work well. Me and all my crossguard folks will greatly enjoy that so thank you!
I also tested the real clash and swing acceleration features and reordered/numbered my clash and swing sounds in an increasing intensity as instructed and I am happy to report that feature works well too. Thanks!
I also tested force push which works great.
Only things I noticed that I wanted to ask about was a noticeable change in my clash sensitivity increasing. I read in the comments where some else had mentioned this too. I actually corrected it by trail and error by decreasing speaker volume (SO Elite 28mm 4ohm/2w) to around 2000 and increasing my clash threshold up to 6.0. This seems to have fixes the issue. It was bad as first and I was getting a lot of false clashes from swing sounds when I would swing. This fix however has made it near impossible to trigger melt, stab and tip drag effect I believe due to the increase clash threshold. Not sure if a sensitivity sensitivity for that is somewhere that I could adjust or if a new separate one could be created?
Hope this feedback helps!
I might have to dig into this some more, because the more I think about it, the more I donāt quite understand it.
In ProffieOS 5.9, we ask the motion chip to give us 1600 accelerometer readings per second. However, we rarely actually used more than 800 per second, so we would skip some of the readouts.
In ProffieOS 6.x, I switched the readouts to use interrupts, so now we actually read out 1600 readings per second. I also switched the range from 4G to 16G, which could introduce slightly more noise into the data, but really not that much since the accuracy of the underlying readings shouldnāt change much, and a single bit still represents ~0.0005 G, which would seem to be enough accuracy.
So, my intuition is that clashes should be come more consistent, not actually more sensitive. If we had a single high reading in the past, there was a 50% chance that it would cause a clash, but now itās a 100% chance.
A few people have reported major differences in sensitivity though, which makes me think that Iām missing something, or messed something up.
Potentially the solution is to build a better clash detector, but such a detector would definitely need to use multiple readings, which would slow down clash detection a bit.
Either way, I think Iām going to have to experiment with this a bit and see what I can find.
1 Like
FWIW, in all my testing across 4 different sabers I maybe had to increase by 0.5. Wondering if itās something specific to certain fonts or sounds.
1 Like
That seems more like what I would expect.
1 Like
Is it possible to detect the ānoiseā playing and maybe account for it when calculating clash or is it too delayed? Wondering if we try to ignore clash if the sound is momentarily passed a certain level? Not sure if it would help but maybe an idea.
2 Likes
Maybe. but if it just makes clashes impossible, then itās still unhelpful.
Hey Fredrik.! Iāve adjust my settings to the OS6 and works beautifully. The reason why it was giving false clashes were from the accent swings being bassy. So I went from 2800 to 1800 for speaker value. Now have no problems. Thanks Fredrik.
One other thing is that Iām trying to reply back to the Proffie OS6 topic but gives me bar that says a limit of three reply is maxed to new users any why thatās doing that? Let me know thank you.
Hmm, that is a good point. There are some new things in ProffieOS 6, such as progressive swing/clash sounds, and slightly different volume handling that might make enough of a difference that it would trigger clashes more.
Of course, the old workarounds also applies, like small piece of electrical tape under the speaker, or between the board and the chassis.
Generally, I think reading and writing posts in other topics would resolve that, if it turns out to be a big problem I can maybe change it though.
1 Like
Just discovered:
If no in.wav in polyphonic font, hum continues when blade turns off.
Thatās as far as Iāve gotten.
Sure. If can change the settings for replyās to unlimited, that would be awesome. Yeah Iāve noticed that the audio is slightly louder than the one on OS5. Alright Fredrik thanks for the chat. Keep rocket bro.
1 Like
Hey Fernando.! Your Prop file for twist/ twist off and battle mode, Would I need to program it to through saber once defined to the config or is there no need?