The backstory...

Lately I've been going through my media collection and updating what I can. I mostly stream remotely using Plex, so in the interest of disk space and easing any real-time transcoding, I'm compressing everything I can down to a max of 1080p 2Mbps bit rate. This will typically result in a 1.3Gb file per movie.

To do this, I've been using a Dell Optiplex 7010MT with an i5-3570K (stock clock speed). I'm mostly a Mac guy, but an old PC is the perfect solution for this kind of specialty work. It lives in the basement by the furnace, and I remote to it and keep it fed with tasks. Now that I'm not just doing a file here and there, the slowness of this PC has become a problem. Knowing how thread count is everything to HandBrake, I was thinking that I might get a new processor for the PC. The fastest that will work is an i7-3770, but they are still going for nearly $100 on fleaBay . Instead, I found a liquidator selling Optiplex 7040s with i7-6770 CPUs for $199. I figured, no matter how fast the new computer is, two would outperform one.

Anyway, on to the observation / question. The i7-6770 is the first official processor who's Quick Sync is supported by HandBrake. I fired up a movie to compress that I had previously processed on the i5. The i5 took nearly thirteen hours to complete, averaging around 7fps. Using the i7-6770 and Quick sync, the file looks like it will take around 50 minutes, averaging 55fps. That is an incredible speed increase, and I will be able to process twenty-six movies a day instead of two.

The question: Will the use of Quick Sync encoding result in a significant change in quality or file size over a pure CPU encode? Should I turn it off and just let the newer i7 chug through the movies the hard way? What are your observations with this type of work on older hardware? BTW, all of the new encodes are HEVC x265 in matroska containers. The Plex server transcodes anything that can't handle HEVC on the fly, so I'm going to go HEVC for everything.

The Dell 7010MT with an i5-3770k is running 73-77C per core, with all four cores pegged at 100% utilitzation.
The Dell 7040MT with an i7-6770 is running 55-60C per core, with the four cores averaging around 55-64% utilization.
I have yet to test the 7040MT with Quick Sync off.

I’m sorry, my comment will not be helpful at all, but maybe the bump will get someone else’s attention.

The last time I used handbrake was with a computer I had built that had an i5-3570k over-clocked to 4.7 GHz. I had a portable hard drive hooked to a wireless router in our minivan and the kids could watch whatever they wanted on their iPads. I quit doing it years ago when they got phones, unlimited data, and streaming services packaged with other services we pay for.

To my knowledge & experience, quicksync will not affect video quality or file size. It's basically hardware-acceleration for H.265. The only thing I've heard otherwise is Plex saying that hardware acceleration can maybe/slightly affect transcode quality - but I've never been able to tell a difference.
6ish years ago I compressed 2000+ DVD/BluRays and with 3x i5-3570, 1 x i5-6600 and AMD FX6300. The i5-6600 did over half the total & all the blue rays. I was retiring some smaller HDDs, so I was pulling the HDD, putting it in the machine (or next to the MB in the case of the case-less AMD) and then handbrake was turned loose on all the movies in the folder and output to my new server (local until finished, then copied to server). There were no I/O or network bottlenecks, so the difference was just CPU horsepower or witchcraft (quicksync). The AMD fell farther & farther behind.
I don't remember which computer processed which video, but I can't tell a difference in quality in any of them that affects viewing. There are some settings I tweaked partially through that helped quality, but that was a PEBKAC problem, not encoder or hardware. There are some fast-motion scenes where the encoder causes a stutter, but not bad enough or often enough to make me re-rip & re-encode - that was fixed as my knowledge of encoding increased.