I read quite a few articles on minidsp flex. To my specific case, would a ddrc-24 be enough? What could I lose? if I put Dirac on flex, ddrc-24 is about half the price. I read that Dac is different, but difference is really audible?
1. The measurable differences between decent DAC's are below the threshold of audibility. Some "audiophile" DAC's which were designed with misguided notions may have much worse performance to the point that differences are audible (and maybe even preferred by some). But even those tiny differences will be swamped by DSP. ANY form of DSP will have a much more profound impact on the sound than any miniscule measurable differences that may exist between DAC's. By several orders of magnitude. Take home message: don't worry about the quality of the DAC, worry about the quality of the DSP.
2. PC based DSP vs. hardware based (e.g MiniDSP). The difference is that PC's have vastly more computational power than those SHARC DSP chips. What this means is that you can use linear-phase FIR filters with high tap counts, e.g. 65536 taps or more. Because of its limitations, MiniDSP must use mixed-phase DSP - in other words, minimum-phase IIR with a cascaded FIR filter (typically with only 1024 taps per channel).
I realise that you may not understand what all this means. If you are interested, you can google the difference between linear-phase FIR and minimum-phase IIR if you want more detail. In a nutshell, linear-phase FIR does not introduce additional phase rotation so it is easier to design. It arguably also sounds better since there is no phase distortion. However it comes at the cost of requiring more computing power, additional latency, and potential pre-ringing. There are many more differences between them that I can not cover in a forum post.
The most important aspect of any DSP is the user. If you take bad measurements, or use the wrong DSP strategy, bad results will follow. It does not matter if you have the most high-end DSP system with the world's fastest PC or best DAC in the world. DSP software simply does what it's told. If you don't know what you are doing, you can potentially ruin the sound.
There are many DSP software packages on the market. Dirac is probably the easiest to use, but it achieves its ease of use by removing controls from the user. It is also the most expensive and the least flexible. And there are many complaints that it is a buggy nightmare that sends corrections the wrong way, fails to correct more than one subwoofer in a multi-subwoofer setup, causes bass to "go missing", etc. I personally would not use Dirac, but there really isn't a better option for beginners.
The best DSP requires human eyes and human brains making decisions. But this means there is a massive learning curve if you want to do it yourself.
Another option is to use a DSP calibration service such as
Accurate Sound by
@mitchco or
Home Audio Fidelity. You pay them for their time, and they walk you through the measurement process and make a set of filters for you. This will result in the highest quality filters possible. Since it's a human and not an algorithm which is making your filters, you can complain to them if you don't like the filters
tried many times. I am always more happy when I turn it off 
