@r.liechty_SDR if that true why [msec] only refresh 10 times not 60? As in our experiments to rotate at
360/1000 which if refresh at 1ms rate should show 1000 times…
1 Like
Try this to rotate 360 degrees smoothly
(([SEC_MSEC] *360) / 60)
Ron
1 Like
Hi, what the different to [sec_msec] * 6 ?
I uses this instead. But it not smoother either
1 Like
(([SEC_MSEC] *360) / 60) removes the decimal fractions so it is smoother I tried it.
Ron
1 Like
Sadly RON this is not smooth on WFS or on my watch . (([SEC_MSEC] *360) / 60) .
Yea I added another component and it stated jerking again. 
I thought I had it
Ron
1 Like
Hey Ron It is not down to you . I know my test is severe . If you spent half your career preparing Special Effects you learn ways to show if something is not smooth .
This is (([SEC_MSEC] *360) / 60) compared to 15HZ .
.
.
1 Like
Sure, it is a problem. You recorded what I wrote earlier.
I doubt it has anything with the watch refresh rate, it was recorded on WFS on PC
1 Like
Hi, mathematically its doesnt .eg
(1.605 x 360)/60 = 9.63 (577.8/60)
1.605 x 6 = 9.63
Removing decimals fraction only be done with floor / ceil
1 Like
Hmm @russellcresser video is more correct then yours?
@russellcresser what wfs u used?
@bruce.bergdahl what wfs u used?
I think @Peter and @bruce.bergdahl u both are recording from wfs preview. Which i myself tried it display as your video, but on the watch its fine at least for the 1/10 value.
So i guess @bruce.bergdahl and @Peter you both are previewing from wfs? No the watch itself
russellcresser’s video
1 Like
I dont have a watch with wearOS, but your records on watch support what we observe already on the WFS, that the tags with milliseconds do not return real fluent milliseconds, just some discrete fraction of that.
1 Like
@peter i admit there still a issue with the milisecond but not as u described.
I not saying no issue. But just want to accurately describe the issue as best as i can.
2 Likes
I am on WFS 1.5.7 Well . whatever is the latest . Watch is Galaxy Watch 4 Classic .
@Peter To me accurately describes the issue/bug helps developer troubleshoot better… that just me.
Hope for your understanding.
Anyway explaining that wfs and watch got two different display shows samsung they need to fixed the watch and wfs too.
1 Like
@russellcresser have u seen any accurate milisecond watch faces? Or platform that can do it?
1 Like
I dont say its perfect, but on facer.io the milliseconds look way smoother. Maybe that is one of reasons why it drains battery more than GWD or WFS faces.
2 Likes
I asked because was woundering its wfs/wff issue or android wear os issue. As in if program watch face with android studio will we have the same issue.
But it seem smoother can be archived and not impossible.
It will not be perfect due to internal clock refresh rate. And also runtime of lines of code to get the milliseconds. (Processing time) sometimes takes a fraction of a milisecond or a few milliseconds, depends on what the code is doing and hardware
1 Like
@r.liechty_SDR i done a test programmatically useing c++ instead. 60hz per second doesnt display as in the videos…
The 1/10 , 1/100 , 1/1000
Appears in this rate…
1/10 slowest, 1/100 slower, 1/1000 fasters…
Correctly
Assuming 1000/60 = 1 sec breaks to 60 parts of 16millis
1 Like
@russellcresser i think i figured it out. Its seem refreshing rate is close to 1/10 of a sec…but before it print to your watch there is a few milliseconds lag. So it becomes e.g
101 ,202 ,103 ,105
And so on.
Instead of exactly 101,201,301…
That why u see it like in my video. Its somewhat looks random but its due to lag time.
With a 2ms lag…it will take close to 1 sec to see the change of 0.010
I tested it on my console program with 1/10 sec refresh rate. The behavior is as such.
For wfs preview…i think its slower like 1/20 sec.
That my testing and current findings
P.s lags are unavoidable but can be optimise code to reduce it
@matze_styles4you and @russellcresser
By the way i made a mistake in my floor formula previously, it should have been
((Floor([MSEC]/100))*100)
Or
For sec_msec
((Floor([SEC_MSEC]*10))/10)
2 Likes
Thanks for continuing to input on this . I am seeing exactly what you are talking about but use a different way to describe it . As Peter mentioned it I will join in . Facer have a method of making smooth Scrolling and Rotation . They have broken thier MSEC timer . It has a glitch losing 10 MS crossing Zero . But they have a Rotational Tag that can be used for smooth stuff . The also have a MS Wake Timer but that is another story .
I would be absolutly delighted to have the 15HZ that the smooth Seconds Hand has . I understand that Samsung / Goodle have to make Processor Bandwidth savings but we are making it Easy for the RTOS guys to muscle in .