Do I really need an " Audio Grade Network Switch "?

I suspect the real answer to the question is like every other question in high end audio: it depends. It depends on the quality of your streamer, router, modem, interconnects and downstream equipment. For me the answer is no, I needed no special networking equipment. I have a flagship streamer that takes care of any upstream issues and completely isolates the bit stream it releases to my DAC. But, probably other folks with other streamers are more effected by the packets and physical connection, You can probably put a lot of money in the network side and get much better sound out of a less expensive streamer… or visa versa. The effect on cables will depend on the equipment on both side.

But one thing I have found for sure is streaming can perform as good or in many cases better than dedicated CD players and with the right equipment equal vinyl. How you get there will be different for everyone, but it is a very worthy pursuit.

So both the Bonn N8 (which I own) and the SW-8 use customized TCXO clocks.  Some report even better results using switches with OCXO clocks, like this one and this one.  

To your question, do you really need to spend the money for an "audio grade" network switch - not in my experience.  Setting aside the technical reasons that the clock in the switch shouldn't make a difference when feeding a competent server, and also setting aside what it really means to have an "audio grade" switch (i.e., better clock, more expensive, what makes it "audio grade"?), I simply have not heard a difference.  I run 45 feet of Ethernet cable from my router to my server.  I have heard no improvement or difference between a bunch of stuff I have tried and compared such as:

• Ethernet cable vs. fiber optic cable feeding my server
• Bonn N8 vs. Netgear switch, and 
• Small boxes by UpTone Audio and W4S vs. nothing

There are things I am curious about trying such as the Intona Isolators, ENO Ethernet filter, and the DJM Electronics GigaFOILv4–INLINE Ethernet filter, but based on my experience I just cannot make myself spend the money when I cannot hear any problems with my current set-up - if it ain't broke, don't fix it.

I'd agree I've heard little difference with devices pre server. Its post server where all the meaningful improvements come from. I'm listening to just installed Sonore opticalrendu (powered by Uptone JS2 LPS), wonderful improvement upon SOTM SMS200. I'm now full optical from first FMC which is direct out from server, to usb out to Dac. Prior had two FMC and ethernet cable to go through. Usb out of any server is highly unlikely to be best streaming solution. Optical post server is optimal setup, absolute isolation between server and dac is a must, dont need all these add on filters with this method.

I haven't tried these products - mostly because the premise of improved "clocking", "timing", and/or "less jitter" improving SQ doesn't seem likely given the error correction in the TCP protocol - as well as where these switches are to be installed (before the streamer/server).

In my experience, "noise" in the digital front end plumbing can have downstream effects on SQ, and I have had good results using products that aim to deal with that:  linear power supplies, Gigafoil, ENO ethernet filter.  FWIW, I have what I consider a decent streamer - the DCS Network Bridge.

Looking at a few of the audiophile switches mentioned here, most only market the improved timing, decreased jitter or "data integrity" as the stated benefit.  A few do mention techniques to reduce clock and other "noise", and/or using improved external power supplies to that affect.  I could see these possibly having a benefit.

As with other proposed improvements maybe somewhat to do with your self-confidence. Do you trust your own ears or others ears?