How fast it rise depend on the Crossover and the driver speed. Both together give the rise time result.
You are quite correct when you note that, when they are combined together, the crossover and the driver produce a rise-time result. Such a combination of course produces a bandpass frequency response function. The frequency response function, for a
linear system, embodies the rise-time result, as well as the response of the given bandpass system to
different transients. Consider an example.
Say we have an 6.5-inch woofer, tuned to provide a certain bandpass frequency response function in combination with a low-pass filter network. It will produce a certain response to a transient, whose rise-time it will be possible to measure.
We now create another system comprised of an 18-inch woofer, which has exactly the same-shaped frequency response function as that of the 6.5-inch system. The moving mass of the 18-inch woofer will generally be much greater than that of the 6.5-inch woofer. However, when we input the same transient into the 18-inch system as we input into the 6.5-inch system, we will find that both systems will have the
same rise time. In the common parlance, the 18-inch woofer will be just as "fast" as the 6.5-inch woofer. Hence, this example illustrates the somewhat mythical nature of driver "speed". Any such "speed" is simply related to the frequency response of whatever
linear system is under test.
Or do you think the crossover frequency is the only part that influence the rise time and a speaker driver have no or few rise time influence?
No. See the example above of a composite woofer+filter system. Of course, any woofer in and of itself is a bandpass system, usually with a quite variable, non-flat frequency response in its passband.
Logically a mechanical device have more influence random phase errors because there need move masses.
At first glance, it all might sound very logical, but the mathematical model of the mechanical device provides us with insights into its behavior. When one device has the same frequency response as another device, then their rise times, etc., will all be the same irrespective of their different masses, stiffnesses, dampings, etc. That's where Thiele and Small made great inroads into the understanding of low-frequency response functions for woofers, and by analogy the same approaches work also when characterizing the performance of tweeters.
If there is no problem with speed of a driver system why should build 3 way systems?
It seems that the so-called "speed" issue is front and center in your interpretation of the various parameters that affect the performance of a multi-way loudspeaker system. As the example I provided above (hopefully) demonstrates, there is no "speed" associated with any given driver. It's how it is integrated into a system, which is typically made up of low-frequency woofers and their enclosure alignments and crossover filters and other (equalized) driver responses that determines the rise-time when tested with transients.
For sound quality, it doesn't help when 20% of rise time depend on crossover 80% on driver speed or 80% depend on crossover and 20% on driver speed. It is same slow and cause lots phase shift and lag.
I know that I am repeating myself here, but the rise-time depends entirely on the frequency response of the system. For example, when tested with a step input, a loudspeaker system that has a perfectly flat (ideal) bandpass response that is –3dB at 30Hz and 15kHz will have a slower rise time than will a bandpass system that is –3dB at 60Hz and 15kHz, all other things being equal.
For a multi-way loudspeaker system, be it 2-way or higher, there is phase shift being added into the summed response. Also, geometric offsets between the drivers will affect the phase response due to differential time delays between the outputs of the drivers. However, these are generally of a not very great magnitude, and the available literature indicates that they do not affect the sound quality; there are much more important design factors to consider.
I am only sure that steep crossover sound not good.
Hmmm. That's a subjective impression and it runs counter to the experience of others. Steeper crossovers bring many benefits to the listening equation when implemented in multi-driver loudspeaker systems, such as minimizing the driver interaction region through the crossover.
I do tests with DSP 24 18 12 and 6 dB. 24 dB sound not good in reverb and room.
When one changes the order of the crossover filtering, there is also a requirement to adjust the responses of the drivers to get the summed responses identical. Did you try and do that? If so, how did you accomplish it? If the summed responses were not identical, then what you were hearing would have been down to differences in the magnitude frequency response curves, no more, no less.
More as 24 I have not tested. But I guess it sound worser because of much phase shifts and lag. The JBL use 6 dB crossover.
Which JBL uses a 6dB/octave crossover. Certainly none of their powered studio monitors, I expect.
A 6dB/octave crossover is quite difficult to implement well, as there is a very large transition band through the crossover region. This means the drivers need to be very well behaved to get a nice and flat on-axis summed frequency response. These first-order crossovers are popular in some circles due to their perfect time domain response, but that comes at the expense of off-axis radiation patterns that may not be as well controlled as desired for applications in a typical listening room. The tweeter especially has to have excellent low-frequency power handling and excursion capabilities, as it is called upon to reproduce quite low frequencies at relatively high levels.
@bennybbbx Have you tried implementing a high-order crossover using FIR filters? These can attain a very steep roll-off, but with entirely linear phase response, manifesting as a pure time delay. Summing the FIR-filtered low-pass and high-pass responses should produce an almost-ideal loudspeaker, affected only by the low-frequency and high-frequency roll-off shapes of the low-frequency and high-frequency drivers that make up the overall system. However, there is likely to be a step-like response in the loudspeaker's directivity at the crossover point, as the transition region would be only tens of hertz wide.