[Dave Haynie]

If you can’t pony up enough for an SSD boot drive, I can definitely recommend a hybrid drive. Last summer I built a system for my son, gaming oriented with a four core i7 and a decent nVidia GPU, along with a 750GB hybrid boot drive. At the time, this was the fastest PC I had made, and the hybrid drive made it plenty fast. That drive, with something like 8GB fast flash cache, was close to $200 last year, but this year you can find them for only a little more than straight HDDs.

You can add on a small SSD to cache a large HDD, but there’s kind of a jungle of software options too enable this. Unless some of the folks here have found a great system, I couldn’t recommend this, though only because I have no experience setting up a successful system.

-Dave

[Dave Haynie]

[Stephen Mann] “It’s just experience. I made Karaoke discs for my daughter with this trick. On most music CD’s the music is stereo, but the vocals are almost always mono.”

In fact, for more professional studio recordings, each sound source is essentially recorded in mono. I would use one mic per singer, one or two per guitar cabinet or acoustic guitar (not for stereo effect, but for tweaking the character of the sound when mixing), three on a piano (88 sets of strings don’t get picked up properly with a single mic), three+ on a drum kit, many more if possible. I’ve even done this on “live in the studio” type recordings — whole band playing at once.

All this goes into a mixer, where the mixing engineer is placing each of these sound sources in a stereo field. And the lead singer is more often than not placed right in the middle of the mix. So that’s easy to cancel out.

You can sometimes play the trick with other instruments. Very simple mixing does nothing more than varying the level of a source left-to-right, so you can reverse that process and “pan out” other sources. Sometimes. More sophisticated stereo mixing uses levels and small delays… that’s hard to undo.

-Dave

[Dave Haynie]

[Rich Kutnick] “Dave, excellent explanation of LTH, HTL and the Watershield technology! So what brand of BR-R do you personally use, and do you have any after-the-fact moisture issues with the disk surfaces?”

Thanks. I originally used TDK HTL discs (TDKBLDRBB), which was a safe buy early on, because TDK only sold the one disc. But lately, they seem to have added an LTH disc under the TDK lable, though I don’t think they make it themselves. And many of the manufacturers are not very good about labeling little details like the disc type or the manufacturing code. You’ll find that the same “model” as sold in the stores can change over time, which is a Very Bad Thing.

These days, I like Panasonic (MEI T02 or RA1) or Sony HTL (SONY NS2, NN2, or NN3); Panasonic only makes HTL discs, and they match my camcorders :-). Sony also sells LTH, so you have to be careful. I used to buy Taiyo Youden and Ritek DVDs, but TY only make LTH discs, and Ritek makes both.

This is a useful list: https://www.hughsnews.ca/faqs/authoritative-blu-ray-disc-bd-faq/10-compatibility#10.5

I haven’t had any huge problem with moisture and disc printing. The Epson disc printing solution works great, no smudges or anything, ever (actually works better than paper — their feeders aren’t all that great, even with gravity there to assist). I thought about using some kind of spray-on finish, but haven’t so far… you have to be careful about solvents that would damage the ink, printing surface, or the disc itself, and also about spraying evenly enough to avoid unbalancing the disc.

-Dave

[Dave Haynie]

I’d personally avoid the Dell. I don’t buy off-the-shelf desktops, but there were a number of Dell’s at my previous day job, and I don’t recall a single one of them not having problems. Not a super significant example, but good enough for me to look elsewhere.

You certainly can edit on a system with integrated graphics. That editing with be slower than editing on a system with dedicated graphics. The problem is, a dedicated graphics chip is sharing memory with the main CPU. It has to fetch that 2-megapixels or so every 1/60th of a second just to display video. So there’s going to be memory bus contention with the CPU, and the GPU always wins. Move to a GPU with dedicated memory — pretty much any GPU with dedicated memory, and you’ll go faster. Some very new system architectures are eliminating most of this contention (moving the CPU to a GPU memory bus, caching graphics memory, etc) but that’s not moving to mainstream PCs yet.

Now, to the specifics. The nVidia GeForce GT620 is a very low-end graphics card… they run under $50 at NewEgg. Now, low-end these days means 96 GPU cores… that’s enough to accelerate AVC playback on the desktop (my old laptop had something like 32 cores and could play back video not even possible to play on the CPU alone). This isn’t enough of a GPU to use for GPU acceleration in Vegas, far as I know, but it’s also pretty low-power. But power is a concern here… what does the basic PC ship with, PSU-wise. Sometimes bundling a graphics card also gets you a better power supply… that may be part of what you’re paying for. I’d recommend at least a 300-350W supply before adding on that level of GPU.

The HD7570 is a bit of an upgrade; these usually run around $100 stand-alone. The GPU is a little bit faster, nothing most people would notice, a percent or two here and there… GPU Boss says “too close to call”. These run a little cooler (lower power) than the nVidia, and usually ship with 1GB video ram, rather than 512MB on most GT620 boards.

The HD7770 is the best of the lot, about $120 stand-alone. Maybe 20-25% faster on most benchmarks than either of the others, and lower power than the nVidia. Also usually comes with 1GB memory.

You can technically do GPU computing on any of these, but I doubt any of them are powerful enough to help out your CPU… much that’s based as well on how slow the CPU is. The i5 is a decent enough mid-range CPU for general computing these days, not something you’d choose, however, for a media PC. Then again, the PC you can afford today gets you much more work done than the one you won’t have until next year.

-Dave

[Dave Haynie]

You do have to qualify “budget” here 🙂 If you’re thinking hex-core, you’re already well ahead of nearly every consumer PC.

1) AMD vs. Intel. I’ve been doing a little bit of ping-ponging. I had an Intel Q6600@2.4GHz (quad core) some years back. I switched to an AMD 1090T@3.2GHz (hex core), which was pretty nice upgrade for relatively low amounts of cash (just a main board upgrade).

But as you get a faster core, and more of them, memory performance does suffer in any of the standard memory systems (2×64-bit, which is pretty much what you find for all AMD and everything from Intel except the high-end). So if you can, get an Intel Socket 2011 main board. Socket LGA2011 also gets you 40 PCI Express links, versus only 16 on the standard LGA1150 socket (just over half the pins, something has to give). Could be overkill, but hey, you’ll never run out! If some day you had a legit video use for 4 GPU boards, you could run that.

I built such a system recently, based on the Gigabyte GA-X79-UP4 motherboard. Pretty happy with it so far. This board included four PCIe x 16 slots — way more than I’ll ever need, but also 12 SATA ports (2x eSATA, 10x internal SATA), which really comes in handy (my old system had an add-in SATA card, as the main board only had 5 SATA ports). I chose the i7-3930K over the i7-3960X… I couldn’t see paying nearly $500 more for an extra 100MHz 🙂

AMD’s CPU micro-architecture has definitely fallen behind Intel’s. It’s interesting — they have a standard core that’s actually two CPU cores sharing a single FPU core. They kind of suggested this as a better answer to Intel’s hardware multithreading (Hyperthreading), and they do actually offer large core count chips, but you have to go to Opteron to get 12 or 16 cores. Those run much slower, and seem more likely to accelerate web applications on a server, applications not usually very FPU intensive anyway. I looked into those, and was not convinced they’d be a good deal for a media workstation. In their desktop processor family, the top chip is the AMD FX-9590, which runs at 4.7GHz and sports 8 cores (4 piledriver modules)… not really competitive against Intel. Part that may well be the problem of keeping the cores fed from the 2-way memory bus, even given a slight advantage in cache size (smaller L3 cache, larger L2 caches, 2MB per piledriver module… but of course, they’re 2-way shared L2 caches, which is probably a contention issue from time to time).

Intel currently has a process advantage, too. The FX-9590 runs a TDP of 220W, the i7-3930K only 120W. My newly revised PC is the quietest PC I’ve had, well, since CPUs needed cooling fans. Part of that’s also in the build. I have a heat sink with a 120mm fan (Cooler Master Hyper T4 — lots of cooling for little cash), very quiet. And three other larger, slower fans on the case, all designed to speed up based on heat. I used a Cooler Master HAF XM case, which, along with room for a crazy number of fans, includes this modern idea of routing pretty much all of the cables behind the main board. So there’s nothing in the way of airflow, anywhere in the case.

The APU idea is kind of a cool one. If you look at lower-end consumer chips and gaming performance, the AMD APU chips are slaughtering Intel on performance. Part of that’s a better GPU, and I’m convinced part of that is the tighter coupling of the GPU and CPU in these chips. Unfortunately, none of these chips are using GPUs fast enough to make any difference in video rendering, far as I can tell. Maybe something for the future, but not useful yet.

2) Cores vs. Clock speed… yes! You want both. Worst case, running a single threaded application, naturally the clock speed wins. This is built-in on i7 chips anyway… my aforementioned i7-3930K has a base speed of 3.2GHz, but will clock up to 3.8GHz if only one core is being stressed. Best case, which is usually attained in multithreaded rendering, all N cores are completely fed with work, and you get a linear scaling per core. So it’s a trade-off… you can make up for clock speed by adding cores, and cores by increasing clock speed. It’s linear, as long as you’re not cache-thrashing. Both faster clocks and more CPU cores can put additional stress on the caches, but particularly the additional cores. And of course, only within the same CPU core family. Comparing different families, like AMD vs. Intel, you have to look at the micro-architecture too.

3) GPU vs. CPU? Right now, for Vegas, I’d put the money into the CPU. Here’s the thing… GPUs are faster. MUCH faster. But there are two problems with that speed. It’s complicated to get your computing into a form that the GPU will help with, and then, well, you have to send that problem to the GPU for computation. What this means is that there’s an overhead for using the GPU. The GPU program has to be compiled (that’s what OpenCL is doing for you in Vegas, and CUDA in other programs), then the program and data have to be sent to the GPU’s memory buffer. One reason I like the APU idea is that this communication is much, much faster, and memory is shared between CPU and GPU (if you look at the design of the new Sony PS3, this is done for the first time basically making the shared bus a video memory bus, not using slower CPU memory as in the typical low-end PC).

So on my old system, the GPU helped. Quite a bit on some things. But you could see the issues. Rendering on the 1090T only, I got 95% or so CPU use. Using the GPU (AMD HD6970) I got more like 75% CPU and maybe 30-40% GPU, and things could go twice as fast, in some cases. That missing CPU time is communications overhead between CPU and GPU; the lower use of the GPU is indicating that Vegas can only translate so much work into the GPU’s language right now (I’m pretty sure Adobe’s doing this better these days, at least for edit-time rendering, not sure about the final render). There’s also some compilation overhead; the time the CPU is involved in compiling OpenGL programs and transferring data to the GPU is measured as CPU use, but it’s not CPU use directly applicable to the job at hand.

What this tells me is three things. One: I’m probably not going to see much of a performance increase going to a new GPU. Maybe a little, based on clock speed, but nothing based on more cores — GPU advances typically come from small boosts in speed and large boosts in parallelism (more cores, wider buses). This also tells me that, as the CPU can do that work better, that missing 25% of my CPU time could start to match or even exceed that GPU help, once you factor in all of the overhead inherent in using that GPU. And finally, it tells me that Vegas isn’t great at pipelining this work. If Vegas was looking far ahead and preparing new data for the GPU while the GPU was working on the previous problem, working during the communications delays, etc. better, I wouldn’t expect to see that CPU and GPU downtime. But it’s a complex problem, and not yet fully solved. On the other hand, there’s no problem using CPU resources to their fullest, as long as your system isn’t hitting other bottlenecks.

So I kept my GPU for the new system. And yeah, it’s AMD. When Sony moved to more complete GPU acceleration (eg, using it in the compositing engine, not just on specific renderers), I bought both the AMD Radeon HD6970 and the nVidia GeForce GTX570, and benchmarked them both. On my system, the Radeon did much better… the GTX570 was not faster on anything I tried, it often lost to the Radeon, particularly on edit rendering, and at the time, the OpenCL implementation on the nVidia was buggy — it actually failed to run a number of standard benchmark programs. I’m sure that’s no longer a deciding factor, and of course, other programs may only support CUDA, nVidia’s compute language, which was out before the OpenCL standard.

But I think, if you’re looking at a six core Intel, spend you money there, not on the GPU for now.

Memory-wise, I’ve honestly never found Vegas using more than maybe 10GB for any single operation… and I’m talking about 40+ compositing layers and uncompressed video files doing animations. And it was actually photography work, not video, that had me upgrade my old system to 16GB. My new system is built with 64GB of DDR-1866 DRAM… again, for Photography (50+ raw 20Mpixel photos composited into 10-20GB individual photo files… yeah, that takes some memory). Of course, if you run multiple copies of Vegas (or other tools) at the same time, it’s really nice to have it all in RAM. For video only, I wouldn’t go over 32GB. It’s also a good idea to buy matched memory modules, a 32GB set rather than a 16GB set now and another later. And of course, you want SSD for a boot drive… that’s kind of a no-brainer these days. I used a Crucial M500 960GB drive, still a bit spendy, but I’m filling this with programs for video, audio/music, photography, and electronics CAD. You know your drive capacity needs.

-Dave

[Dave Haynie]

I steer clear of any LTH BD-Rs. This is basically DVD-R technology, organic dyes, repurposed for BD-R… LTH means low to high, as in reflectivity. The dye is changed by the laser to allow light. These have the same longevity and light fading issues as DVD-R.

The original formulation, HTL, starts out highly reflective. The material is non-organic, a sandwich of copper and silicon. When the laser writes, that section melts copper and silicon together, creating a low reflectance part. No problems with light exposure, and way more BD compatible, no software tweaks necessary.

Never used a watershield disc, I used normal inkjet printable HTL discs, with an Epson Photo 1400 printer. I think watershield is a JVC thing, and given that they had a bunch of the CD-R and DVD-R organic dye patents, it seems likely they are deep into LTH.

Not sure I trust the concept, anyway… printing water-based inks on a water-resistant surface seems sketchy.

-Dave

[Dave Haynie]

Hey Stephen… just what would you recommend in nVidia for ultimate performance these days? I took a look at current nVidia and AMD GPUs, just to get a sense, and nothing seems a dramatic improvement over the HD6970 I have today. Having the new system nearly all together now, did make me think on the GPU thing.

I’m not looking for “pro” cards… I have a Quadro at work. Probably a big win if you’re doing mechanical CAD, but kind of pointless otherwise, particularly at $2K or so. And yeah, there seem to be intentional slow-downs in some OpenGL ops on the GeForce cards, to make the Quadro seem much better for very specific CAD work. But for video and particularly OpenCL, I don’t see a win here in Quadro, particularly for the money.

Also decided to make the data drive RAID5. I have a controller, if the on-board isn’t sufficiently fast. For most of the things I’m doing, particular video rendering, I value read performance and reliability, don’t care so much about write performance (can always render to the SSD, an external drive, a project drive, or even an SD card, if this was a real issue…).

-Dave

[Dave Haynie]

[Dave Osbun] “Xeons are designed mainly for server use, but high-end workstation use as well.”

No. Xeons are designed entirely for server use. They’re applicable as well to certain kinds of workstation-class work, but I rekon Intel doesn’t spend ten minutes worrying about workstation-class performance when designing the next Xeon. Just too small a market, and it’s inherently going to be served by the intersection of server and “enthusiast” processors.

[Dave Osbun] “In regards to render times, Xeon workstations will easily outperform any Intel i7 processor.”

I don’t deny that there exists a dual socket Xeon workstation that’ll outperform the top i7 on renders. But you’re paying dearly for it. Take the Xeon E5-2687W 8 core, runs around $1900. Per core, this is slower than the i7-3930K 6 core (about $500). Sure, I could get a little more performance out of pure SMP with that chip in my main board, and it’s one of the Xeons (intended for blade servers) that works with commodity DDR3, rather than slower, way more expensive server registered for FBDIMM RAM (which is also more reliable, buffered to allow huge memories… but I’m at 64GB, enough for now).

And yeah, the Xeon has 20MB L3 cache, versus the i7’s 12MB L3 cache… but you know, media applications don’t bloody care about all that L3 cache. In a server, you may have hundreds of relatively minor threads, each running different code. In the main use of a media PC, you have one bit of code, optimized to fit in the L2 if not L1 cache of a commodity processor, running the same on each core… and data that’s just there for a visit, not really needing much cache at all… stream in, stream out. So sure, stack up two of these and you’re going faster.

But you’re probably paying 4x-6x as much for that 2x-ish speedup, going into server market pricing. That’s what I was getting at above. If you look at the PC market, there’s a price-performance curve, and there’s usually a pretty distinct “knee” in that curve. Below the knee, 2x as much money gets you 2x the performance. Above the knee, performance per dollar drops off exponentially.

And then there’s the reality of the market. It may not always be so, but it’s true so far that the top performance hardware doesn’t retain value. Intel’s top enthusiast processor… the top socket 2011 i7 these days, the i7-3960K, runs around $1000. When that’s replaced by one of the i7-49xx series forthcoming, it’ll drop to $500-ish pretty quickly, as long as Intel’s upping the bar still. This happens about every six months. So when you look at Xeon, sure, if you can really justify the expense of a $4000+ system every 2 years or so, go for it. But with a much longer upgrade cycle, you could go to the higher volume enthusiast chips, and have yearly upgrades for the same money.

For that matter, you can build a 64-core Opteron system from “regular parts” (any number of Supermicro main boards, Opteron 638x/637x series)… but each core is less than half the performance on an i7 core. So yeah, you can make a much faster PC, but you’re paying exponentially more for that performance.

[Dave Osbun] “If the budget permits, and you do this for a living, Xeon CPUs and compatible motherboards are the way to go.”

I actually make my living designing computer systems… embedded stuff, these days. I did desktops back in the early days of computer video… in fact, some of those kind of defined computer video back in the 80s. I mean, sure, if you can justify it for business and take a 179 deduction on the PC in a given year, that might be the right way to do it. Particularly if rendering is your only goal and your income really comes down to dollars per hour rendered.. it’s easy to see that $4000 pay for itself. But not everyone has that kind of budget.

I also have other needs: I use my system for video, photography, music, and CAD. Video drives CPU performance and HDD size, photography drives memory size, and the CAD stuff just hangs on for the ride.

-Dave

[Dave Haynie]

Ouch! I had a dual-core AMD system ages ago. In the early day of HD, it was taking something like 28 hours rendering for about hour’s worth of HD video. I knew something was up when I got a new, slower clocked Intel dual-core laptop that rendered the same thing in 24 hours. I moved to quad and then hex core after that on the desktop.

The AMD hex core wasn’t bad when it came out, for the money, but you could still do better with Intel…. of course, all without going to multiple socket. That’s an option, but a very pricey one.

The A10 is a good value; I got my son an A10 powered laptop for Christmas last year (for grad school — he’s got a i7 desktop for gaming), and it was surprisingly good for the money.

The concern I’ve always had for any of the new AMD architecture is specific to media processing — the Bulldozer/Piledriver architecture is based on core modules. Each module basically contains two integer cores and one FPU… so while your A10 has four real CPU cores, there are stalls in accessing the FPU hardware once your code gets too FPU intensive. So their high-end processors run up to 8 cores for desktop, 16 cores for server… but only half the expected floating point hardware. Initially, the octal-core chips often lost going head-to-head with the older hex cores, like my 1090T. They got past that, but mostly by going to insane speeds (single thread burst on these goes to 5GHz) and faster memory… but still no quad-bus memory.

The APU idea, on the other hand, is a pretty good one, especially for mobile. If you look at some of the A10 reviews, AMD actually beats i5s and even i7s on graphically-intensive benchmarks and Intel’s GPU. Being on-chip, there’s much less communications latency between CPU and GPU.

One of the problems with GPU computing is that, for any given GPU, there’s a fixed overhead of compiling the OpenCL, then sending compute jobs to the GPU. Increasing the amount of work you can send to the GPU each time, increasing the GPU speed, and decreasing the latency can all make your GPU help. On the other hand, if that overhead plus computation would take less time on a really fast CPU … and as long as you don’t have other work for that CPU you can pipeline (one of the main problems in rendering engines these days, I believe), then you’d be better off just using the CPU. But with that relatively weak CPU (speaking in terms of workstation-class computing power) and wicked fast interface, the GPU should be a good thing. And as I mentioned above, I’ve had very good success with the PCIe-based AMD GPUs, at least as of a few years ago. These things do tend to bounce between AMD and nVidia, who’s got the crown this week, I can’t say.

-Dave

[Dave Haynie]

[Stephen Mann] “SSD’s do have a write cycle limit, but PC Magazine did some tests and you will be upgrading a long, long time before that limit starts to bite you.”

As long as you treat them right, that is. I do expect to see the drive replaced long before it’s getting old — unlike HDDs, solid state drives really can follow Moore’s Law. It’s also very likely we’ll be seeing alternative memory technologies in the next five years, which will address the longevity issues.

As for the WD Green drives, they work fine in editing systems. They really don’t randomly shut down like people think. They’re even supported in the Drobo RAID. And keeping cool, they last pretty much forever… I had enough of Barracudas dying early due to overheating.

[Stephen Mann] “X79? That is so last year. I just built a new editing PC with the Gigabyte Z87 mobo, an i7-4770K, 32Gb of RAM”

The Z87 is limited to LGA1155 CPUs = only two memory channels and only 16 PCI Express lanes. X79 supports the Sandy Bridge E chips on Socket 2011, four memory lanes and 40 PCI Express channels. Haswell E CPUs and system chips will certainly come out, eventually (8-core CPU, X99 chipset, DDR4 memory, etc. announced for “late 2014”), but I was looking for performance today.

[Stephen Mann] “I’ve had nothing but problems when I ran Radeon display adapters”

I actually bought both when Vegas went to full GPU support, both around $300. The Radeon was faster with Vegas, and the nVidia had bugs in their OpenCL implementation. I’m sure that was a temporary thing, but never had any problems (I was using nVidia before that). I figure I’ll wait on the GPU until there’s some compelling reason to upgrade — given the CPU, it’s unlikely today’s Vegas will see much help from the GPU, unlike with the much slower AMD system.

-Dave