That's a complicated question. All pumps emit some noise, no matter what. Depending on the size of your case and how much ventilation (perforations) it has, it can be completely muted (Fractal R5), fairly quiet (TJ08E) or always present (M1). Beyond this, many different things can be a factor to noise. Too much flow/pressure relative to the components and restriction in a loop can cause a cyclone effect. Air in the loop can make a pump sound much louder than a properly bled loop - the difference can be drastic. Same problem if the pump is vibrating against metal vs if it is properly decoupled so those vibrations can be absorbed.
The other noise factor is pump speed. When it comes to the DDC pump, you have two main types - a lesser performing/slower operating MCP 350, and the higher performance/higher rpm MCP 355. The latter will perform better, but emit more noise, naturally. It's running at higher RPM. You've also got the PWM version which will allow you to more easily tailor the speed to your particular loop. You may only need your pump running at around 1400-1500 rpm if you've just cooling a cpu and single gpu, and none of the components are particularly restrictive. Adjusting RPM can be crucial to managing noise levels.
The D5 is a larger pump with a larger motor. Running at lower RPM, it can provide similar performance to the DDC (flow), but at a lower dB. The DDC traditionally wins on head pressure compared to it's larger sibling, though from what I recall it's relatively close. The D5 Vario has a 5 speed analog switch that will allow you to set the speed and manage the flow/sound balance. D5 PWM (outside of the Aquacomputer one) is relatively new and from my understanding if plugged in incorrectly the tacho meter will burn out. All D5's are affected, including the Aquacomputer ones that are for Aquabus/Aquaero use. Have to be careful with them.
With respect to reliability, DDC (MCP 350/355) and D5 (MCP 665) pumps are both considered reliable, and many would argue the best both in terms of reliability and performance. This is why you see them so widespread. I have noted that some people consider the MCP 355 as less reliable due to running at a consistently higher speed (thus wattage and generation of heat is higher), however if you ensure the pump remains cool (heatsink, airflow over it, etc) it should not be any worse than the slower running variant.
The Eheim pumps I have only done a small amount of research on. With widespread availability of the Laing options, they have fallen out of favour both for reliability and performance reasons. They are also larger, and require specific connectors as they are not G 1/4.
I forget who makes the EK DCP pumps, but couldn't really tell you how reliable they anyway. Swiftech makes use of DDC and D5 pumps, but they have made a custom pump solution for the H220X.
I quite like Dazmode's store section for pumps as he takes great effort to clearly lay out which pumps are what model. Please note "Darkside" is his own branding/product line, in case you were wondering.
https://www.dazmode.com/store/category/bare_pumps/
