Leaderboard

I'm putting this question out there to see if anyone else is doing something similar. I've been running Parsec over a modest 1Gbps home network, paired with BetterDisplay to ensure proper 4K+ 60fps signals are sent to my iMacs—while editing 6K video on very modest hardware. For those unfamiliar, Parsec is low-latency remote desktop software. With BetterDisplay, I can match the resolution to the native Retina display of my iMac, allowing me to work from two different locations in my home without needing multiple high-end machines—or a 10Gbps network. The end result? It feels like I'm using an Apple M2 Mac Mini directly on a 2019 4K iMac. Honestly, I can barely tell the difference. I’d estimate the latency is under 25ms with this setup. Now I’m curious—has anyone tried this off-site? I imagine cross-continent setups would suffer from latency, but within the same continent, it seems like remote editing could be entirely feasible. It sure beats lugging around terabytes of footage. For me, this setup has breathed new life into an older machine. I’m even considering picking up another iMac—maybe a 5K model from 2017 or later. With H.265 decoding not being too demanding and the M2 or M4 Mac Mini handling encoding with ease, it seems like a highly effective and budget-friendly workflow.

With Parsec, the key is efficient H.265 (HEVC) encoding and decoding. On my 2020 Intel MacBook Air, I’m seeing decode times around 9ms at its native resolution. On the other end, the M2 Mac Mini encodes the HEVC stream in about 8.5ms. Over Wi-Fi (on the Air), network latency hovers around 4ms. Altogether, that gives me a total latency just under 25ms—well below the 42ms frame time of 24fps video. In practice, it’s totally usable and, to my eyes, indistinguishable from working directly on an M2 MacBook Air. Most Intel Macs from 2017 onward support QuickSync, which provides solid hardware-accelerated decoding. The real game-changer, though, is the vastly improved encoding performance on Apple Silicon, which makes setups like this feel snappy and responsive. When using Ethernet—even at modest speeds—latency drops further (often below 1ms). Parsec can stream H.265 video with variable bitrates up to 50 Mbps, which is trivial for gigabit Ethernet. It also supports 10-bit 4:2:2 streams, though honestly, I can’t see a difference on my aging eyes—especially when using high-DPI iMac displays. For me, Parsec solves three major problems: Reviving modest hardware — especially for tasks like video editing. Enabling near-instant resource sharing — with Thunderbolt providing a pseudo-network connection up to 40Gbps between Mac Minis. Letting me work remotely — from the kitchen table, for instance, while the main setup runs in the basement. When you pair Parsec with BetterDisplay, the experience is nearly identical to working on the host machine—if not better. In my case, the iMac’s high-DPI Retina display actually gives me a superior viewing experience compared to using the M2 Mac Mini with a standard external monitor. Unless you're using an Apple Studio or Cinema Display, it’s hard to beat the visual quality of 4K, 4.5K, or 5K iMacs. Interestingly, the 2017 iMac Pro isn’t the best choice as a Parsec thin client. It lacks both QuickSync and the T2 chip, which hurts performance. If you're going this route, a standard 5K iMac (2017 or later) is a much better option.

You might be talking me into reinstalling Parsec and trying this out, but for different reasons. Usually I edit on my Macbook (The lower-end M2 Max) and it's totally fine, but when I'm at home, I prefer to edit off my basement NAS instead of from external drives and I use a postgres database on a server next to it. But if I don't plug into the network (which means sitting at my desk where the dock with 10ge is), editing video is not great (and editing photos in Lightroom ain't great either). But I have a windows gaming PC also set up near the desk which has a pretty decent setup with a nice GPU. I suppose I could just parsec to that while on the couch or in bed and do my editing, with the main catch being that I need to use Windows shortcut keys. That or I wait until there's a decent deal on an M4 Mini with 10ge (there have been some solid deals, but they're always for the base model with only 1ge)

I often use Splashtop Remote Desktop when editing remote and its always been very reliable. My main use-case is making an edit on a windows machine hooked up to a bunch of media servers in one city and I'm in another city on a MacBook. It handles the keyboard mapping and display scaling extremely well. It also has support for multiple monitors, so I can plug the MacBook into 2 external displays and have all 3 screens mirror the 3 Windows screens. Or assign them to spaces if I'm using the MacBook without the externals. Another advantage, in my case, is it has the option to blank the screens on the remote computer, and runs through a secure network, so feels very safe if you need that. I'm curious about Parsec, but it seems to be a similar price for similar feature set. Since Splashtop is doing the job just fine, I'm debating if its worth the time to research and switch over. For colour depth, both say they're 4:4:4 so I don't think that's much of an issue.

If macOS ever supported multiple simultaneous users, I’d happily run just one Mac Mini and use 3–4 iMacs around the house as thin clients for the whole family. I’m confident the performance would be more than sufficient—as long as no one is editing video at the same time! Unfortunately, that’s not the case. But there’s nothing stopping me from setting up several iMacs as thin clients, each connected to its own Mac Mini. By placing the Mac Minis next to each other and linking them via Thunderbolt, I could enable ultra-fast access to shared resources—effectively building a small, high-speed local computing cluster with shared storage and plenty of power.

I’m not 100% focused on color accuracy here—I rely on scopes for that—since calibrating two machines in different environments is never going to yield perfect results. My main reason for using BetterDisplay is to match the exact 4096x2304 resolution of my 2019 4K iMac. Retina displays don’t play nicely with standard external monitors, like the one I currently have connected to the M2 Mac Mini. I’ve heard some users connect “dummy” displays to the encoding machine to force the proper resolution, but this setup with Parsec achieves the same effect. If I send a standard 4K signal to the iMac, it doesn’t look as clean—the image is upscaled and noticeably less sharp. With this setup, I just launch Parsec and connect to the M2 Mac Mini. What I see on the iMac looks almost indistinguishable from running natively on the 2019 i3 iMac with 8GB of RAM—except, of course, I get the performance of the M2 Mac Mini. In fact, I often forget I’m even using Parsec. I’ve accidentally shut down the Mac Mini thinking I was turning off the iMac! And the biggest surprise? I’m getting Thunderbolt 4-level performance on my media drives—despite using only a 1Gbps Ethernet connection. As long as latency stays below a certain threshold, the experience is virtually seamless. The iMac just becomes a glorified thin client- exactly what I want.

I never tried editing with Parsec, but it's a good idea! How does it handle color management? Or is that done by Better Display?

I've had an initial look at the test I shot that compared shutter speeds, including longer exposure times than 180-degrees as you suggest, but didn't really notice any improvement there. Maybe there is and I'm just not sensitive to it, but it's not what I'm sensing. I absolutely agree that human movement is better, but I don't have any volunteers handy that don't mind being published on the open internet, so unfortunately that isn't something I can easily do. I edited up my other test of the four-camera setup and that didn't yield any joy either. That test included lots of camera movement as well as subject movement and so I probably would have noticed if there were significant differences, so not noticing any is probably a good thing because it means that the camera isn't the answer. I was beginning to wonder if I wasn't viewing things objectively, but then I randomly fired up The Bourne Supremacy and it auto-resumed to some random timestamp and within seconds delivered these two frames, which have it in spades. I'm now convinced I'm not chasing a ghost, so that's positive. I think I've been struggling because I've been conflating the look I've been chasing with the fact that I have mostly witnessed it on movies shot on film, so I've perhaps paid too much attention to sharpness and grain and not enough on other aspects of the image. The fact that I haven't seen anything above about 2K projected until very very recently probably also plays into it. Recently I've sourced some higher quality reference materials and have found examples that contain the look but are also higher resolution. These are all 4K so be sure to open the full resolution file, not just viewing the highly-compressed preview image the forum shows. I will be studying these as my next area of focus. At the very least, I'm eliminating things that it isn't, and that's progress of a sort. My anamorphic adapter has finally left China so I think that'll be the next round of tests. Now I've eliminated the camera body as the source of the look I can venture out and hopefully capture some footage with people in it, but it's very wet here right now so we'll have to see.