because i can not sense it in out put voice of my speaker.
This may be either a problem of speaker quality or your skills in percepting distortions. If you ask a sound engineer, he will hear it.
There's apparently a misunderstanding about the purpose of diodes in the simulation circuit. I implemented them to simulate the real behaviour of the LM358/LM324 output stage, that isn't correctly represented in the National Semiconductor (neither other manufacturers) SPICE model. With the diodes, you get class B behaviour, which is also present in the real LM358/LM324.
When you add a complementary output driver without additional bias voltage generator (just connecting both base terminals to the OP output), you get a similar behaviour, even with a class AB OP, e.g. LM741.
Finally, if we take a look at your original 3 OP amplifier, we can expect, that the worst distortions will occur at a stage, that is driven to a medium level of e.g. 100 mV. Because each stage has a different level, there's a good chance to meet this condition at least for one stage. In so far, it's reasonable to make the simulation with medium level, as I did. Then the cross-over distortions can hardly be missed in the simulation result and will be also perceptible, I fear.