If task2 just wants the “last” value converted, the best choice is for the data source to just put the value in a global variable and the second task to read from that.
The idea of a queue is to allow a task to wait for data, and to startup as soon as it is available, or if the queue gets full, for the generating task to be delayed, and restarted as soon as there is space to place the data.
Both ends of a queue by definition work at the same average rate, otherwise the queue overflows or underflows.