count~ not always autoresetting

15 years back i tried to use multiple copies of [count~ 0 9 1 1] in conjunction with multiplexing and demultiplexing audio signals in order to lead upsampled signals from poly~ to the mother patch and back.

because it did not work, i had to switch to a master phasor~ system instead. which is all but ideal.

i did some tests now and i found out that [count~]´s autoreset message simply does not do what i thought it should.
all messages to count~ are not executed inside before the DSP is on, so things can get scrambled. from my experience this is the only audio object which does that.

linking it to dspstate~ and/or using messages to reset counters and make them start at 0 is out of question, too, as you can easily imagine, especially under SR 352800 conditions.

does anyone know about a workaround to synchronize counter~ instances at DSP start?

/ main
/ bpatcher1 / poly~ foo 1 up 8/ 1, 2, 3, 4, 5, 6, 7, 8, ...
/ bpatcher2 / poly~ foo 1 up 8 / 1, 2, 3, 4, 5, 6, 7, 8, ...