Finding the Island of Source Stability

I figured this is a sort of neglected, sort of interesting topic that deserves it’s own thread, and worth talking about (continued from my spur of the moment post). This will not be comprehensive, nor specific, but I think it’s good to get the discussion going and get some general ideas of where to start and what to look for. No guarantees that everything here is 100% safe or suggested by experts, because it’s not; I am a careful person myself, and these are tradeoffs with my system that I’m willing to make for my obsession. Please point out anything that’s too stupid if you see it though. None of this is for free, every change is a tradeoff, and it’s trading a bit of system usability, security, or stability for the sake of sound quality. Some will trade off more than others, and it’s up to you to figure out what tradeoffs you want to make through your own testing.

While going for a dedicated, purpose built, only audio streamer from a proper hifi brand may be the best way forward for most chasing ultimate digital performance, there’s plenty of reasons why that’s not always the most reasonable solution. Reasonable solutions aren’t the highest profile nor highest performing, but they can offer some serious satisfaction if you put the effort and thought into. Most use their computer for a variety of different things, from gaming to media consumption to work to just plain fucking around, and sometimes you either need or want your music playback coupled to that same device. In my case, my journey was started by the need to both play and record music using this computer, and a dedicated streamer would just not fit my equipment and workflow well. Others may want to find a way to reuse what they already have to create something better without spending a ton and keep their current way of life. A select mentally ill few may just enjoy tweaking so much that an uphill battle project like this is just appealing to them. Those freaks. But I’m all 3 at the moment, so I can’t really criticize.

This is not on how to fully optimize a home built audio exclusive streamer, but it’s also not only one or two small tweaks to make computer playback a bit more passable. These are some general guidelines and tweaks that when properly experimented on and combined will allow you to have your cake and eat it too when it comes to music playback from a general use computer, without sacrificing usability and general computing performance. This feels like element hunting, smashing computer tweaks and hardware together until you get a stable and effective result, hoping for something both significant and actually practical/useful. Sometimes you end up with a great sounding result, that unfortunately disappears in minutes after you realize you can’t use your computer like a normal computer. Sometimes you end up with a perfectly stable and usable computer, but one that’s a bit heavy and poisonous to musical enjoyment. Some tweaks and tests are cheap to try but time consuming, others are more expensive but higher probability in giving the result you desire. Either way, if you keep shooting, you may land near the island of usable yet high performing musical source that lasts long enough to enjoy and note down, close to the level of a purpose built streamer without the usability limitations. You also may not, the odds are against you, but it’s not impossible.

Bold statement, I know, but for myself, I feel like I was able to optimize my system from both a hardware and software level, using both consumer and hifi components, to a level that I really didn’t think was possible from a general computer. Your results will vary, greatly, depending on your exact setup, use case, and mentality/effort. I can’t cover such a wide swath of potential configuration in hardware, software, and use, so I’ll just touch on changes that are general enough to be applicable to most, yet still meaningful in most.

This is focused on getting a balance of high stability/low latency with moderate overall power draw and system noise, rather than lowest noise but most variable latency, or highest performance + low latency at the cost of pushing system noise and stability. I want to use my personal computer as a personal computer, yet still extract as much audio performance out of it as well during local file playback and audio editing. I do not stream from wan or lan, so I can’t offer much there, but tuning your network stack would realize larger gains I would imagine more than I focus on here. I am not looking for a car I can only drive on a track in best conditions, I want a daily driver that can largely crush it on a track when needed, but can still drive to the grocery store or work without too much stress or punishment. There are plenty of varying methodologies when it comes to getting this result, and there’s a lot of debate by those way more dedicated than I, so this will just be a small collection of what I prioritized for my own setup. I’m not going to flat out tell you how to get the exact same build I have, I’m too lazy for that, but the goal is also to encourage people to experiment with their own systems first and hopefully learn more than I have already in this endeavor on their next build.

Not in any real order.

Physical build considerations (without boutique hifi hardware):

  • No RGB/lighting for case and components
    • generates undesirable electrical noise and interference, begone vile gamer slop.
  • High quality, oversized PSU
    • you want to reduce the amount of electrical noise and improve quality of power for all components, go for platinum or titanium rated with at minimum a 25% oversize for optimal efficiency/performance.
  • High quality barebones/non bloated motherboard
    • better quality parts like caps, chipsets, more thought toward electrical isolation and thicker pcbs, more performance and stability focused, bonus points for non gamer. I feel like a smaller form factor might have the potential for less latency given shorter traces, however the closer packed nature theoretically may be less advantageous than something full size, dunno as I haven’t experimented with this personally yet.
  • Tightest timing ram for your platform, at least 16gb but 32gb in 2 sticks if possible
    • trying to balance speed and latency, with an audio focus low latency will generally take priority, along with having the ability to use it as either a ram disk or ram playback from your music player and avoid having swap if needed.
  • CPU with single chiplet design focused on single core performance, lots of cache
    • most audio is single threaded and will benefit from extra cache so it touches ram less, along with an integrated io controller. Try for 6-8 cores if you plan to disable SMT and mess with process/core affinity. I clearly have a preference toward modern ryzen (as long as it’s only 1 chiplet, if you have multiple try to force everything audio related onto one)
  • Single MLC/TLC NVME drive
    • will generally provide the lowest latency and fastest access over QLC and beyond and potentially lower electrical noise.
  • Passive CPU cooler if possible, if not then use air tower, never water unless pump is powered externally
    • reduce electrical noise in system and physical vibration, and physically quiet is good too.
  • Powering case fans with external PSU
    • reduce electrical noise in system
  • Use integrated GPU if possible, if not then lower power dedicated if possible
    • reduce electrical noise in system and reduce workload of system, I know this is hard if you play games, but pick a high quality card if you do plan to game, and undervolt.
  • Avoid wifi and use wired internet if possible, with respected chipset on motherboard
    • reduce electrical noise and interference in system, and reduce workload on system, NO POWERLINE, EVER. No bluetooth either, same with wireless keyboard and mice (maybe a little is ok, as a treat).
  • Spacious, separated case layout
    • the more distance you can get between potentially electrically noisy components, the better, and more space for passive cooling and expansion in the future.
  • A separate isolated stand for your computer
    • farther away from audio equipment without having to resort to a long digital cable, vibrationally coupled.
  • EMI/EMF reduction techniques
    • finding a way to reduce incoming and outgoing interference and noise can be benificial through various means like tape/coating/sheets/etc.
  • Selecting the USB closest to your CPU for your playback device
    • Make sure your music playback device is on it’s own dedicated USB controller with the shortest path to the CPU. A dedicated USB card will elevate things further if it’s quality, especially audiophile focused, will touch on that later.
  • Isolating any connected cables and peripherals
    • limit the amount of things connected to the computer if possible, use USB isolation devices for your input and non audio usb devices, reduce the amount of noise that could flow into the computer.
  • Using quality displays
    • I’ve had alot of shitty monitors over time that buzz and whine and negatively impact anything around them, and while hard to test, make sure if you get anything seemingly dirty to send it back and try for something else.
  • Careful internal cable routing
    • Doing whatever you can to get everything away from the power supply cables, maybe using a case with multiple sections or running it as far away from everything else as you can, paying careful attention to keep it away from PCIe, USB, and storage devices. Nicer power supplies will also have thicker more high quality cables, which are more cumbersome but worthwhile to deal with.

BIOS Tweaks (that generally won’t noticeably negatively impact general computing performance):

  • Disable Spread Spectrum
    • disable undesirable distribution of electrical noise in system which also increases jitter, this setting exists for compliance, not quality, will not impact compute performance.
  • Disable C States/CPU Sleep
    • disable core sleeping and varying performance states which negatively impact latency, will also improve compute performance at cost of increased power use and heat.
  • Disable EIST (speedstep, speedshift)/CoolnQuiet (CBS)
    • disable variable CPU control that negatively impacts latency, will also improve compute performance at cost of increased power use and heat.
  • Disable EPU/Power saving chips
    • disable variable motherboard/system power control that impacts latency, will also improve compute performance at cost of increased power use and heat.
  • Disable virtualization/iommu (if unused)
    • reduces overall system overhead which will reduce latency.
  • Disable QPI/UPI Power Management
    • keeps link at max performance and non variable which reduces latency and increased compute performance.
  • Set LLC to middle/balanced setting
    • keeps voltage consistent and stable, improves potential performance if not at extremes at the cost of a bit more power and heat.
  • Set CPU & DRAM Power Phase Control to middle/balanced
    • keeps VRM consistent and stable power, improves potential performance if not at extremes at the cost of a bit more power and heat.
  • Disable Power Decay Mode
    • keeps CPU voltage regulator from moving to low power, improves potential performance at cost of bit more power and heat.
  • Set Idle Power-In Response to Typical
    • keep system from reducing power usage at idle increasing overall consistency.
  • Disable UNUSED features/devices (motherboard audio, networking, sata ports, pci slots, thunderbolt/USB controllers, etc)
    • reduces system workload and electrical noise, improving latency
  • Disable Memory Scrubbing
    • improves memory latency at cost of less error correction but not a big concern.
  • Disable ASPM (pci power management)
    • increases PCIe power stability/performance by preventing lower power state, improve latency.
  • Disable any integrated remote management chips/features
    • reduces system workload and noise
  • Disable fastboot
    • ensures hardware and software boots cleanly each time, improving stability
  • Disable EDAC/ECC memory (if you have it, if you can)
    • removes extra latency and overhead if applicable (this is also part of software tweaking), I’m not convinced ECC is good to have here.
  • Set Uncore Frequency to max
    • improves overall system latency by keeping frequency high for i/o, memory, and cache.
  • Disable SMI/SMM and ME if applicable
    • removes system overhead and background tasks, this is a bit more convoluted but can be done on some systems and also part of software tweaking.
  • Enable HPET
    • increases accuracy of system timing, improving latency in some situations.
  • Disable any RGB or lighting control
    • reduces electrical noise and overhead.
  • Enable resizable BAR
    • On some systems this might be a bad thing, but I think if you’re using a PCIe USB card it could actually work in it’s favor, though I haven’t fully tested this myself.
  • Disable any link between the OS for mfg software
    • reduces unnecessary overhead and latency in some cases if applicable.
  • Disable any local AI models, tau boost, other weird mfg specific things
    • see above, it could be anything.

More Painful BIOS Tweaks (that will cause a bit to a decent compute performance hit, but still be usable in most cases):

  • Disable SMT/Hyperthreading
    • improve latency at the cost of compute performance, potentially drastic performance hit in some workloads.
  • Disable any turbo boost, variable speed, P states for CPU, manually set voltage as well as base rated clocks, a slight undervolt could be beneficial in some cases
    • improve stability and latency at cost of performance, noticeable performance hit depending on CPU boost config.
  • Set ram voltage and speed manually to base XMP, then focus on timing but don’t neglect overall speed, a slight undervolt could be beneficial in some cases
    • improve latency at the cost of speed, smaller but meaningful difference, depends on the platform a lot.
  • Drop PCIe speeds to 4.0 or even 3.0 on storage and experiment what sounds better
    • reduce overall system electrical noise and workload at the cost of system and storage speed, I feel most mixed on this one.

I would say these are generally applicable to most systems, there’s a lot more fine tuning you can do with each platform, but as mentioned that’s too much effort and scope to really cover, but it’s worth poking around and experimenting yourself. There’s so many things that have been added or changed in modern BIOS and system config that I’m sure there’s some things that are skipped over that could make a noticeable difference, it just requires people who are willing to waste time to find them lol. I could mention more extreme things like disabling networking entirely, removing things from motherboards, and so on, but that’s approaching more streamer than usable computer. Please let me know what you find here, and if something needs to be changed. A lot of this is from direct experimentation on my own systems, but what works for me may not work for you, and vice versa. I’m also sure that some of this deserves a bit better explaining or is a little sloppy, but it’s good enough for now, I’ll update when I have more mental clarity.


This separator is so I can talk about some of the more audiophile hardware that also makes a drastic difference to the experience, that’s worth investing in. Not to the realm of converting this to be boutique like a streamer, but things easily integrated in a system like this to take it to the next level. I will keep this brief, because there’s only a few things that I think are needed, before it makes more sense to just go for full streamer and not general use PC.

  • Dedicated audiophile USB card
    • something from JCAT, Pink Faun, Taiko, Afterdark, etc will really do wonders for the quality of your output, make sure to use the PCIe closest to your CPU and that the rest of your system is optimized properly, the most important thing to consider.
  • Dedicated audiophile network card (if you stream only)
    • I don’t think this is worth it unless you’re streaming your music from the network, but otherwise potentially similar impact to adding the USB cards.
  • Dedicated audiophile PSU for that USB card
    • basically a requirement for one of these cards to be worthwhile, it’s not that the PSU is doing all the heavy lifting, it’s just that the cards don’t shine without it either. Pick the best PSU mfg of your choice, the second most important thing to consider.
  • Dedicated audiophile Master Clock for that USB card (if applicable)
    • if they have the option to take a clock in, it can be a good way to further boost performance down the line, but it can be really pricey to get a good clock and starts to border into “might as well streamer range”
  • Dedicated audiophile ATX psu replacement
    • a few select companies like hdplex or jcat make native options for this, but a custom psu is also an option, but once again cost wise this borders on just get a dedicated streamer complexity and price range.
  • (This space intentionally left blank)
  • MAYBE dedicated audiophile SSD + PSU
    • I really want to try this since there’s a few out there, and it’s a part of the system that is critical, so the returns just may be worth it, but because I can’t test, I can’t say.

Notice how I’ve focused on USB, I know there are PCIe cards that can output SPDIF, but to be honest I’d rather leave that to a higher quality DDC down the line and use one of these USB cards in conjunction with one, as I think there’s more flexibility and potential for improvement. You could use a USB reclocker/isolator out of a motherboard port, but I genuinely think if you’ve optimized the rest of your system, a card + psu will do more for less.

14 Likes

WIP, OS and Software Optimization:

yada yada i don’t feel like doing this for windows, so here’s linux, sorry. If I am super bored one day, I’ll pull out my notes on windows tweaks. I prefer linux for audio here because there’s less going on by default, there’s more optimization potential, and you can customize the os to do what you actually want it to. Other reasons not relevant to audio. The goal of OS optimizations is to reduce the amount of background processes aside from playing audio, increase overall system stability/consistency, and give the most direct path for audio from both a program process/playback perspective and digital audio path perspective.

THIS BEING SAID: my tweaks here are a mix of decreasing latency, and reducing noise. For purely audio playback, not all of these tweaks will be positive because at a certain point I think it’s better to prioritize low system noise over latency, but since for myself I’m doing more realtime recording tasks alongside playback, I value both. So this would be different from a purely listening perspective

Linux Base Tweaks (for most general systems):

I’m not an expert, not anywhere close, and some of this stuff I don’t fully understand myself. Some of it is redundant depending on the system and config, so double check everything. yada yada

  • Picking, or creating, the right kernel
    • If you have the know how, modifying your kernel for your system can be really rewarding from a performance standpoint, but can be a hassle. I don’t have the energy to go into how to do your own kernel, but if you’re lazy I’d go with a Liquorix or Real-Time kernel, as those will out of the box handle low latency well.
  • Picking, or creating, the right distro
    • Personally, I have always lived with Debian base, and for my system I’ve settled on MX AHS Linux with my own kernel and system tweaks, picking the most barebones and building from scratch would give you the least bloat and overhead, but I understand that sometimes something premade that’s a good balance is just easier even if it’s not as immediately ideal for sound. Arch is also on my list to try for utmost performance.
  • Disable Spectre/Meltdown mitigations
    • These mitigations add extra system overhead and decreases performance, though it will reduce security so think about how much you care about that. It’s more impactful on older systems than brand newer ones, still important though.
  • Add user to an audio group with rtprio 100, memlock unlimited, nice -19
    • Important for realtime performance to force highest priority for audio performance
  • Set CPU governor to performance
    • Increases performance and consistency of system, but may result in increased noise if you didn’t lock cpu performance in bios
  • Disable any USB & network power saving/autosuspend
    • Fixes problems that can occur after gaps in playback, and makes sure usb and network isn’t being handicapped
  • Run in User mode and not root
    • Audio playback isn’t intended in root.
  • Enable high resolution timers
    • This shouldn’t have an impact on most systems for audio playback (especially with other tuning options in place), but it will reduce jitter for some driver/process types.
  • Enable tickless kernel
    • It shoulddd be that way by default on some distros, but you want to enable it for sure.
  • Use Prempt RT/threaded IRQs
    • This is only if you’re using non realtime kernels, but allows realtime scheduling and threaded IRQ to prioritize IRQ threads over everything else. Setting the kernel to preempt full also should be done.
  • Allow user space programs to set realtime priority
    • Allows apps to force themselves priority if they are smart enough to do so, otherwise this may need to be done manually.
  • Disable swap or change swapiness to 10
    • If you don’t have to touch the disk for swap, don’t, if you do then reduce the amount it does it.
  • Use EXT4 or XFS for your filesystem
    • There’s a lot to go over here, a lot of potential performance tuning to improve ext4 performance and reduce disk activity by disabling journaling or safeguards but I don’t want to suggest that here if you’re not comfortable, this is a risky one if there’s data you care about on this system.
  • Ensure IRQ are not shared for sound devices
    • Unshared IRQ will result in more stable and better performing audio, and sometimes you may have to unbind ports if there’s interference.
  • Allow power management to be controlled from user space
    • This only matters if you don’t lock your system in a performance state and if the app you use has the capability to do so.
  • Use a lightweight DE and disable compositing, run at lower res if possible
    • Just reducing the overall system graphical and overall load.
  • Disable unusued services/daemons and prevent unused drivers, devices, and modules from loading
    • See above.
  • Don’t use disk encryption/luks
    • Adds extra disk activity and overhead, undesirable in audio playback.
  • Set process niceness and affinity to isolate music related tasks to other side of CPU die
    • Manually configure audio playback processes to run on isolated cores on the other side of the processor from everything else if you can, and make sure they take priority.
  • Disable system sounds
    • If you’re running exclusive this won’t impact things, but it will help to reduce interruptions or weird behavior in resampling if not.
  • Optimize your tcp/network stack
    • Another really long one, if I was not streaming I would focus on reducing overall network activity, if you are streaming then I might consider focusing on prioritizing it, it’s more than I’d like to get into here.
  • Enable proactive memory compaction
    • Improves latency performance with huge pages.
  • Disable kernel samepage merging (KSM)
    • Memory deduplication that causes tlb shootdowns which increases memory access latency, undesirable.
  • Disable transparent huge pages (THP)
    • THP causes latency spikes when pages are promoted and when kernel attempts to force them into existence, don’t use this.
  • Disable NUMA balancing
    • If this is used it increases memory access time variability which in my mind is undesirable for latency and consistency.
  • Disable cstates (again)
    • sometimes even if it’s disabled in bios its still bad if it’s on
  • Disable ASPM (again)
    • see above
  • Pre fault, lock, and pre allocate memory
    • Reduce overall page faults and reduce interrupts and disk activity.
  • Change I/O scheduler to none
    • For a hdd use deadline, others will increase latency.
  • Change dirty ratio/background ratio to 0 or 5 depending on system
    • Try to prevent stuttering or variability during IO activity.
  • Offload RCU callbacks + lazy
    • Try and reduce the amount of callbacks to prevent unneeded waking.
  • Disable audit framework
    • Increases latency in low latency or realtime kernels.

Setting up and configuring rtirq, Pipewire, ALSA, and Wireplumber:

Setting up and configuring Music Players:

7 Likes

Interesting topic. Even as someone who does own gear in the totl server/streamer category, I can see the benefit of going this route instead. Mainly the cost aspect as something like this would be cheaper by an order of magnitude although the same might not be able to be said about the subjective differences. Also just the hobby enthusiast aspect of building and tweaking something like this would be a really cool project and learning experience I’m sure.

I’ve wanted to mess around with a linux install for gaming purposes for a while now, I might extend that to an audio application in the future if I ever get around to it :wink: :face_with_tongue:

I do enjoy the ease of a turnkey solution that I don’t have to fuss with tho…

5 Likes

It’s certainly a learning experience, I sure don’t feel confident in any of it still lol (even though I should know this stuff???). I do think there’s a lot of value to be had in tuning, but as mentioned there is a point where for the sake of your time and potential performance benefit, a streamer might make more sense. For the sane person, if they don’t have specific circumstances, a streamer really may just be a smarter idea. I don’t want to come off as one of those clickbait “DITCH YOUR STREAMER???” or “Get TOTL streaming for CHEAP” and so on, because this isn’t exactly that.

Dew it

I wish there was a turnkey solution for my situation, I’d gladly spend that money over years of fucking with a computer

2 Likes

I’m definitely bookmarking this thread. Very good info, thank you!
I think I’m at a point where putting in the effort to optimize my PC makes sense. Literally ALL of my listening, outside of commute and bluetooth earbuds, is on my computer…and when I’m home, I’m always on. On top of that, I do very amateur production, so it seems like the logical way to move forward. Don’t get me wrong, I do want a streamer. Still holding out for a Sonore or Antipodes super deal. One day…

Funny thing, I was JUST looking at SoTM’s stuff today. They got hifi motherboards, pcie usb cards and noise-blocking sata connections. This hobby…you’ll always miss out on something new; relatively speaking.

I do have a question though. I know some people line their DIY project chassis with a copper foil shield or something to that effect. Have you ever though about doing that with a PC case? Actually, now that I’m thinking about it, I wonder how that would work because I think the motherboard is grounded to the case…right?

1 Like

SoTM is a bit old at this point for most things they make, they were good in the past but I think for some things today there’s better value elsewhere. I’m extremely impressed by the JCAT USB XE EVO overall, get a good psu like an UpTone JS-4 and it’s truly awesome, and later on you can improve further with their master clock module. If I could redo everything from the start, which I basically am in the process of doing lol, that’s a no brainier for both that specific usb card and PSU compared to others I’ve tried. If you were going to put money into anything audio focused for your pc, this would be it. While doing all the tweaks above improve things further with or without, going from no jcat + psu to with is such a drastic improvement over anything else, that it would be my first step regardless of what you do

I have, and I have experimented with that too, although I found that it was a bit overkill or didn’t matter much for my entire PC, but I did end up doing that around the USB output card and also to create a barrier between the motherboard and PSU. If you just put your pc physically away from the rest of your equipment and electrically isolated, I don’t see the need to treat the entire case (at that point go for a dedicated streamer/server with more low emf/emi components broken into separated/isolated daughterboards and milled cases with that stuff built in, since the problem with consumer hardware is that the noise is going to be the system itself rather than anything coming in)

1 Like

Oh man, I’ll give that a look. I’m thinking I could pair that with a custom Plixir even. Might be overkill.

I remember digging through the forum waaaay back and coming across you getting rid of a JCAT. Was that the older model? It wouldn’t play nice with the Sinope or something, and I wanna say you used a reclocker instead.

I thought about going the reclocker route, but it seems like extra money/steps for just ONE usb out.

1 Like

That was just a USB isolator, which was a rebadged Intona. This is totally different. I did also try the original jcat USB and this is also a vast improvement of that, leaps and bounds

1 Like

You can just get EMI/RFI shielded cases:

2 Likes

It’s probably not a 0 value proposition, and it doesn’t hurt, but I’ve tried it on metal cases, and IME it has zero effect on noise. The reason it get’s used on guitars is because it produces a faraday cage around the electronics, to limit hum from external sources.

It makes sense to do it if you have a plastic or wooden case, In a metal case you already have a faraday cage.

For metal chasis’ I’ve tried covering the inside of the chasis, wrapping wires in it and surrounding noisy components with both tape and very thin sheets of copper, all with negligible difference. Put a 1/2 inch of aluminum between the offending component and the audio circuits and you will hear that though.

Multiple boxes and lots of air are the best solution.

3 Likes

Yeah and that’s where it’s just best to go with something designed to tackle that problem than the start, since doing it with a traditional modern pc will probably be more effort than it’s worth, although I think trying to treat a USB pcie output card is potentially worth it with better shielding

So much good stuff in this thread and I’ve been meaning to contribute to it but I’ve just been busy with life.

That said, sometimes life comes in and bitch slaps you around a bit and here I am for a brief touchbase on the subject. Couple of days ago I lost my cam, a cheap external usb speaker and my better than decent quality usb mic. After fiddling around a couple of days with the OS thinking it was a software issue or something that I’d inadvertently changed, I came to realize that the USB hub built into my old Samsung monitor crapped out.

Looking around I bought a cheap but very sturdy hub, from Amazon, not thinking about it much other than basically bypassing the built in hub in the monitor.

Of course I have an ifi 5v Power X supply laying around and used that to power the hub, connect the usb peripherals and I’m back in business with my mic, cam and cheap usb speaker, just a basic web meeting setup.

Well, later that day I sit down to listen to music while I browse and generally waste some late night time online and… … …hey things are sounding really good tonight, everything just has more depth and a little cleaner, and the black is blacker, blah blah.

Yeap, I guess I inadvertently removed a source of noise from the PC. One I hadn’t even thought about, but there is was.

5 Likes

Oh yeah I can imagine those built in usb hubs are not very good when it comes to isolation, was probably introducing a lot of noise and junk into the speakers. Not to mention some powered external usb hubs can sound very good compared to going straight out of your pc too. I have a powered anker hub I have been using for ages and I remember back before I made the switch to a streamer/server it would sound noticeably “blacker” with my usb dacs.

3 Likes

Still using @dB_Cooper sotm USB hub powered by a hypsos… Def getting put to good use :+1:

3 Likes

Ha! One of the things I wish I hadn’t sold off.

2 Likes

@M0N is there an option for the JCAT to take clock from an external clock? I’m sure their clock option is really good but the ADark Emperor Giesemann EVA clock I have feels like it could be pulling more duty if that thing had a clock input.

Their external master clock is great value, but I’m sure there could be better, however it’s 20mhz input/output on both the clock card and usb card, so it’s not compatible with 10mhz ref options. If you wanted 10mhz usb cards you’d want to try something from afterdark themselves or others that support 10mhz in, or rig up some frequency multiplier which would impact things in various manners to get to 20mhz

2 Likes