Seeking Design Advice: Pro Audio Workstation

Hello L1 alumni. This is my first post, primarily aimed at establishing a build plan for my next workstation.

First some context. I built my last PC tower using an AMD Phenom 9950 (Agena 140w) 10 years ago. After that, I had acquired a first generation MacBook Pro retina in 2014. The core i7 in the MBPr ran laps around my PC tower so I have been using it as my daily driver exclusively since 2016. I have not planned a PC build since around 2008-2009, so I am out of the loop rather heavily when it comes to the current state of professional builds (not that my old tower was “pro” by any stretch of the imagination).

If you are curious, I bought the MacBook Pro to open the possibility of using Logic, as well as to try MacOS.
This is a PC build, not a Hackintosh build. I have no interest in pursuing the Apple side of things, even in approximation.
That said, if you feel I would benefit from aiming for a Hackintosh system let me know why.

What I want this system to be able to do:

  • I want the system to be professional and upgradable with emphasis on computing power, stability, and longevity.
  • I would prefer to aim for a best performance per dollar approach rather than bleeding edge tech, if possible.
  • I want to have two curved ultrawide displays (Samsung CF791) to form a wrap-around, immersive creative space.
  • I want to run Cubase Pro, Ableton Live 10, and Studio One V3 without fear of bottle-necking while working.
  • I will be using either an RME HDSPe MADI FX, an RME MADIFace XT, or a Prism Sound Orpheus for A/D, D/A conversion and processing.
  • I want to run Adobe CC video editing apps, such as Premiere Pro or After Effects, without fear of bottle-necking.
  • I want the option to run an Adobe CC video editing app along side an audio program if necessary (rendering in background).
  • Gaming is not really something I am interested in doing on this machine, but I wouldn’t rule out the possibility entirely.

That out of the way, I saw this video while looking for 2018 workstation builds:

Ultimate Mega-Workstation PC Build Guide (Linus Tech Tips)

This seems like a compelling concept for a machine. It is certainly light years ahead of my old tower and my MBPr, but I lack the knowledge necessary to vet the compatibility of this design with my intended concept.

To counter this flashy build, there is a configurable option from Silent PC that I am also considering, but my main concerns with this option are markup on parts or unnecessary add-ons, and a lack of customization; issues that have more leeway and direct control in DIY builds.

I apologize for the simplicity of this post and hope that the build in the linked video can serve as a placeholder for a parts list for now, like a “loose sketch” of the general direction I would like to go in. Currently, Threadripper seems like the best performance to cost ratio out there, but I don’t want to invest in a system this year only to have it flag far behind more affordable systems only a few months later.

Keep in mind that this build will be something that I save toward and will take time for me to accomplish.

I would like to start discussing what sort of components will prove critical and which options among those I should be selecting, such as nVME vs M.2, AMD 1950X vs similar Intel chips, or ECC vs non-ECC dimms. The 128GB dimms that the motherboard in the LTT video build is compatible with are supposedly not available yet, but these would be something I would pursue down the line.

Thank you for reading.

1 Like

Do you plan on running GNU/Linux or BSD at any point in time on this thing? Natively or in VM?

Everything that he expects to use is either for Windows or MacOS ( or both ). Don’t see any need for linux at all.

@anon5644329
I was asking just to make sure. Pretty much determines what hardware will be easiest to use.

@WhizzingGosh
In that case, I have nothing to really offer, except AMD Ryzen would be the more well rounded platform to meet your needs. Threadripper is overkill. I would suggest an R7 1800x or R7 2600X for plug and play. These will give you the ability to work on thread heavy apps and allow you to encode while recording and what not.

Graphics card is the hard part. The market seems to be getting closer to MSRP now but still it is easy to pay much more than what the card is worth.

RAM is expensive right now as well so expect that to hopefully go down in the future. There is no need for ECC RAM in your situation unless you just need.

I use GNU/Linux as a DAW so that is the most that I can help you with.

1 Like

Thunderbolt

Based on further research, I hadn’t counted on the Thunderbolt standard being an option restricted to Intel machines. Add-in support for TB3 on the Gigabyte Designare EX (sTR4) would fix that issue, but there has also been a change in my plan.

Composer Workstation

I am now shifting my focus toward designing a composer’s workstation in the long-term, rather than aiming for a single system that ticks all the boxes. The ultimate goal is to have a master system that will then control three slave systems. I feel that the smartest way to proceed would be to build the master system to begin with and then add in the slave systems one at a time thereafter.

The complete system design will go something like this:

  • Intel Master System
  • AMD X399 Slave Systems (x3)
  • Systems networked via Ethernet using Vienna Ensemble Pro

This networked systems approach really shines when a composer needs a large number of detailed samples available in their entirety without losing system performance when arranging in the sequencer. This leads me to a key point in my overall plan.

I want the slave systems to be built using the X399 platform because it will be upgradable to a max of 1TB of RAM once higher densities become available, something that samples will benefit from. Also, the cost is much less than building Intel systems. A huge number of cores is not necessary as far as I’m aware, but I’ll return to that point later.

What is not so clear to me is how to strategize the same approach when incorporating entire packages of VSTi that involve realtime calculations.

Based on all available research I could find, a balance between per core speed and core count with high floating point throughput and an efficient chipset on the motherboard are all critical in this context. Beyond a certain threshold of core count, more latency could be added due to unnecessary complexity in the processor compared to the intended task. Meaning, it makes no sense to get a 7980XE or a dual socket motherboard, for example, because while this would be considered “powerful” by a certain metric, it is not going to be of benefit in this case.

Unfortunately, Threadripper, especially the high core count chips, act as two physical sockets jammed into one and testing for pro audio purposes shows that they cannot achieve good performance at 64 samples of latency. The best they can manage is 256 samples, which is far too much latency for my purposes. This is damning because of the seemingly greater potential found in the 64 PCI-e lanes on sTR4 boards. But ultimately this means all calculations must be done on an Intel system due to their overwhelmingly superior floating point throughput and low latency performance.

X299

Here’s where things start getting murky for me.

From what my further research has revealed, the X299 platform is a much less compelling option because of considerable cost and even base features being removed in favor of paid upgrades, such as RAID support beyond mode 0. I do not want my system to fall behind even faster due to building on an already outmoded socket, but there are also other things to consider, such as greater driver support on Core i7 as opposed to Core i9.

I am torn between the greater performance on Core i9 and the greater savings of Core i7. It just feels like X299, as a platform, is a bit of a money trap to get involved with. My primary concern is to find that best ratio of per core speed, core count, float throughput, and chipset efficiency. But my secondary concern is to future proof it against rapid obsolescence.

I have also considered building two much more powerful Intel based machines with two more modestly built AMD X399 systems. One of these Intel machines would be my dedicated system for audio work, while the other could have a dual purpose as a content creation machine for Adobe CC (2D/3D assets, video editing).

Gaming Not Important

Notice that I haven’t said the word “gaming” once so far. I am not interested in playing games on a system design intended for audio work. If I want to game, I would do that on a separate system to something like this just to make sure everything was optimized toward that one workflow and set of tasks. That said, I will need to power at least two monitors and possibly three to work comfortably on this setup, meaning at least one suitable graphics card on the main system (if only one Intel machine).

I welcome any and all input regarding this concept as it currently stands.
Is it sensible? Stupid? Modern computing makes it irrelevant? If so, why does Hans Zimmer use setups like this? Etc.
Thank you all for reading and helping educate me on how to start.

based off prior research/testing and some new testing done by Gamers Nexus if you are going to use Adobe CC then just get a 8700k processor and 64gigs of ram as it does not fully utilize the cores of the processor unless you are doing multiple concurrent renders.

Thanatopsis:

Thanks for the input. I have seen the 8700K mentioned a lot even in terms of pro audio use as a best overall value when choosing Intel CPUs. What I meant by using Adobe CC on the possible second Intel system is that I figured I could have two user profiles, whereby one was intended for VSTi calculations and the other was intended for 2D/Video focus. Do you think this would still be a good pick of processor given that clarification?

What’s wrong with modern flash storage? like Samsung nvme drives and Intel optane drives have low latencies and are cheap as chips compared to ram.

Also, if you’re going to go with higher core count Intel, they’re also “multiple CPUs glued” design… aka. NUMA.

@khaudio maybe?.. not sure though…

@risk Thanks for the suggestions. I took your advice and implemented both nVME and optane in the first draft design of my sound module PCs. These components are not replacements for DIMMs but can greatly improve performance in this sort of situation.

Since my sound module PCs will more or less always load the same things every time, the ability to cache that on the optane is incredibly desirable to reduce possible delays.

That said, it will only be of benefit when powering the systems on. What the systems will load will pretty much just stay loaded until they are powered down again.

Here is the PC Part Picker list for my sound module PCs:
https://pcpartpicker.com/user/WhizzingGosh/saved/w2RZ8d

I feel that with the 8700K as the best per-core performer of all currently available Intel chips, this overall design is the best option to pursue for these systems. It’s even a bit overkill considering these subordinate machines will be running sample libraries as opposed to calculation heavy VSTi. But I have a reason behind that.

When running Vienna Ensemble Pro 6 (VEP6), you are forced to make changes to parameters on the system where the plugin is hosted. This means I either need to be running a remote desktop setup for easy KVM (cpu consumed by doing this), or I could simply load whatever I want on the master PC and offload it by hooking the main machine into a Waves DiGiGrid IOS for acceleration of any VSTi that I need quick access to without any mucking about with KVM, etc.

An added incentive for incorporating a DiGiGrid IOS is that it provides an 8 in, 8 out interface with preamps along with the entry level DSP network for offloading. Plus, if I ever need to boost that with the other systems, I can do that as needed.

At this point my main conflict is as follows:

  • Build all systems in the network using 8700K for greatest performance per dollar
  • Build an entry level Core i9 master PC for future proofing
  • Build a secondary master PC to whatever Intel one I build using sTR4 for a “best of” both chip makers
  • Build a sTR4 master PC to reduce cost, considering the DiGiGrid IOS will compensate for any downsides to Threadripper’s NUMA (and is scalable)

I also plan on running two Samsung C34F791 curved displays on this thing and wouldn’t rule out the possibility of having some sort of third monitor of some kind. I figure one graphics card will be able to drive both screens with no trouble?

What do you think would be the best option with respect to designing my master PC?
I am leaning toward an all Intel network at this point, mainly because of how much better Intel chips perform.

If you’re loading into ram and it’s sticking there, optane doesn’t matter, 960evo or similar is fine. Optane 900p should have <20uS random read latency (e.g. 4k blocks) . For comparison, ddr4 ram is around 65-80ns these days (for a 64 byte read).

Where I was going was, 20uS = 1s/50k – which is kind of at the level of your typical CD quality output sampling rates, so with any luck, your software won’t have to prefetch a lot of stuff into ram, you can grab the data from “disk”.

Also, 32G optanes are an entirely different tech, don’t bother with those.

@risk After checking into it a bit more, I think that Optane won’t actually suit my purposes. It primary benefits users of legacy drives that take time to spin up. I’m only going to be using M.2 2280 SSDs in these Mini ITX units, so it actually won’t be doing me any good to include this module.

I have created a X299 variant of more or less the same VE Pro Sound Module build, linked here:
https://pcpartpicker.com/user/WhizzingGosh/saved/W346XL

The price difference between the two builds is roughly $293.24, leading me to prefer going with the X299 option over the LGA-1151 parts list due to future proofing and negligible cost difference. Scaling this design up by three, assuming I build one master system and have three slave systems as originally planned, this means the cost difference in the overall network will be $879.72 in total.

I feel that this is a more than acceptable risk considering the ability to upgrade down the line when prices for Core i9 will become more reasonable in the long-term. Also, while overkill for the moment, the greater strength in the X299 specification allows for greater flexibility overall as opposed to comparative limitations on the older socket.

I am shifting my focus toward the physical layout of my workspace. I know that I will be using a Doepfer LMK4+ and that this keybed will need to be in front of me at center position. This gives me a 5’ 2" (157 cm) span in terms of minimum front edge length for my desk.

I have a crazy idea of mounting my dual curved displays, reference monitors, and the 88 key LMK4+ to an UpDesk Ultra so I can have an adjustable sit/stand workstation, but I have never seen this done before, most likely for a reason. All the same, here are my considerations for the components that would be mounted to the desk in that scenario:

  • Focal Twin6 be (pair): 61.6 lbs [30.8 lbs ea.]
  • Samsung C34F791 Curved Display (pair): 32.62 lbs [16.31 lbs ea.]
  • Humanscale M8 Monitor Arm (pair): 28.4 lbs [14.2 lbs ea.]
  • Doepfer LMK4+ 88 Keys: 52.91 lbs

This yields 175.53 lbs in total, or roughly 50% of the maximum load capacity of 350 lbs for the UpDesk Ultra models. A slight increase in weight would be introduced from desktop monitor stands, but this should be nominal.

I think the reason people haven’t done sit/stand workstations is because of using large analog mixers or the need to keep all of their speakers and desk setup in a fixed position for optimum performance. I don’t see why it would be detrimental to use a rig like this that could raise and lower the entire stereo field and configuration of the workspace with mobility in mind. But then, I’m probably missing something here if nobody’s doing it.

Hey will read more than the OP later but in summation, prioritize, in this order:

num CPU cores
RAM capacity
RAM speed
CPU speed

Thanks @khaudio.

My research has told me that balancing core count with per core performance is key when planning for audio builds, which would seem to indicate that the 8700K is still king over any other chips on the market.

That said, I would prefer to go with the 7820X (for now) because it is the best entry level LGA-2066 chip for my purposes. I am keeping in mind that this is involved in the build of my Mini ITX slave systems and not my main box.

As for my main box, I am still undecided, except I don’t want something like the 7980XE or the 1950X in there. The point is to get it built to start with and then increase its abilities as and when I can.

The main system will be connected to something like the UAD-2 Apollo 8p to allow for recording and tracking using that plugin format, while the other systems will be running things like EastWest, Komplete, Waves, and others over VE Pro 6.

For starters, I don’t see a need to go further than 4 systems for quite a while.

I have just discovered this three part tutorial by Cory A. Robbins.

In that video at the beginning, he describes using (more or less) this server: Dell PowerEdge R710 Server
His had two 6 core Xeons in it, 128GB RAM, and 4 x 450GB SAS. The one linked has 8 x 300 GB SAS in it.

My thoughts on using old servers went something like this:

  • Wow, Kontakt actually works on that hardware?
  • It’s cheap, but would I get burned by pricier server parts when servicing a clunky refurb unit?
  • I’m going to be stacking these inside of rack mount cases anyway and those are near $100 each, empty…
  • I don’t like having to use 4U rack mount cases to fit my i7 coolers, this looks like a space saver with less headache…
  • The videographer is a day job IT professional, so he has a more informed purchase decision than Joe Public…
  • Why not just get some of these and save that money for the main system instead?

But then, I’m sure there are reasons not to buy these. Like, Xeons (even dual socket configs) being really weak compared to single socket Core i7 8700’s. Or parts being iffy when dealing with refurb units. I’m sure there are a billion ways I could talk myself out of this, but it’s very compelling and I’m wanting to say yes because of how much of a turn key it would be by comparison. It’s over $500 in savings per system at fixed costs.

I think the case to be made here is that this option is a greater value, professionals are already using these units and it’s working very well for them, and the parts are actually rated to last longer than the consumer market. So, I guess it makes more sense to go for a small batch of these and only focus on the details of the main system design.

What do you think?

I would go watch the video Wendell did on TS about upgrading a Mac pro 4,1 or 5,1. Dual Xeon X5650’s and an assload of ram and the ability to have osx is probably the best route for what you’re doing, and it can be done on the cheap easily.

But, if not, I super duper recommend a ryzen 2700x build. Hell, if you wanted some mobility the Asus ROG GL702ZC is an insane machine and you can probably put a ryzen 2 chip in it. I know people put 1800X’s and 1600X’s in them. So there’s that. I’m drooling over one right now.

Been a bit of time since I posted here, had to think things over.
I don’t have any desire to upgrade an aging Mac Pro. I want out of the Apple-sphere.

Based on news that has come out of Computex, AMD is looking more and more compelling in terms of system design, especially when considering TDP and bootable NVMe RAID.

Because of the difficulties in deciding between the two platforms, I think it really comes down to driver support and system stability overall.

The biggest concerns I have regarding AMD as an option are as follows:

  • Increased Latency from NUMA
  • Forced System Reboot When CPU Stressed
  • Random Crashes When Running DAWs
  • Lack of Audio Interface Driver Support / Unstable ASIO Drivers
  • No TB3 Support / Only “Slow” Connections

What draws me to AMD is the lightning fast response that bootable NVMe RAID would give me for workflow and loading / saving projects. I think it would be a fantastic thing to have nearly zero wait time in that regard. Also, it costs less and uses less TDP, meaning running the systems also costs less, not just the parts. Obviously there are the much larger numbers of PCI-e lanes, larger max capacity for RAM (eventually), DIMM.2, and non-bottlenecked M.2 slots to consider when sizing up Intel’s clumsy and expensive looking VROC.

Let’s rewind and revisit the purpose of this build:

  • The main system should be the strongest machine, capable of doing a lot of real time processing, such as VSTi that require calculations to achieve their sound. It’s basically going to be the “brain” of everything. It’s where the actually composing / scoring is done. I would like the flexibility to also do some visual work. Basically Adobe CC with an emphasis on video editing and after effects.

  • The sound module systems are going to be primarily used for loading many instances of Kontakt and orchestral libraries. By their very nature these libraries are more or less sample based and require a lot less CPU to simply stream out audio as opposed to generating it.

While the Core i7 8700 (non OC version) is a very good value, it seems as though it’s overkill given the context of what the sound module units will actually do. If I need to spread out entire libraries of VSTs across certain systems, like a Waves bundle, then I would probably build those specific systems with the 8700 in them for better performance. But otherwise, I don’t think it’s necessary.

When speaking in terms of socket TR4, so 1950X and its successor, could the latency of NUMA be mitigated by using virtualization to effectively split up resources into “several” (say, 4) machines in one sound module, effectively quadrupling my return on investment as opposed to much more costly Intel systems?

Can I quietly water cool rack mount cases with slight modifications for running loops? Is that as much of a nightmare as it sounds? Water cooling is a requirement on socket TR4, based on everything I’ve ever seen, heard, or read.

I want to say yes to AMD, I really do. The price to performance ratio is very tempting. But I have been burned in the past, it seems to good to be true, and the DAWBench numbers still look very lean compared to what Intel chips can deliver. Still, those benchmarks are aimed at doing everything on a single machine. My use case is a bit different. It all boils down to this:

  • I need my main system to be able to run as much as I want it to, within reasonable limits, taking Cubase Pro 9.5’s “offline processing” into account for added optimization of project files.

  • I need my sound module systems to be reliable and tailored to the tasks they are meant to do. I’d rather not build something with more beef than the job requires whenever possible, but not penny pinch it to death either.

  • I require all systems to be designed with longterm use in mind. Dependable, serviceable, upgradable, salvageable / harvestable.

I think the Core i7 8700 PCPartPicker list is still a valid starting point, but I guess I need to know if AMD is even worthy of consideration given my intentions. I have seen people praise Ryzen, but there is virtually no real world user experience regarding Threadripper 1 or 2 as an audio production machine specifically for use in composing with large template files.

Help me decide how best to not be able to decide.

I cant access that http://www.dawbench.com/ forums at all, but

First I’m hearing about this benchmark, maybe you could create thread and ask for results since people have that hardware
I still have just Ryzen 7 1700, but i’ll pitch in, Cubase is not that large software to pick and run that thing, but its email subscription to download that trial is annoying