Main Concept AVC/AAC encoder MP4 file with these specs:

3840x2160 custom frame size

30fps high profile

Variable bitrate 28mbps max, 20mbps average, progressive download

Nice! $576 for a good upgrade.

Is that 980ti being OpenCL utilized for timeline playback?

It would be interesting to see how an RX480 addition compares.

@RealityStudio What kind of memory bandwidth ("winsat mem" at a cmd prompt) do you have currently with that platform, and are their any improvements you can make on that front?

Perhaps you can tweak the memory settings using "AMD Ryzen Master?"

Nice! $576 for a good upgrade.

Is that 980ti being OpenCL utilized for timeline playback?

It would be interesting to see how an RX480 addition compares.

Plus it's looking like old old computer components will get just over $300 on ebay, so it's going to be closer to a $250 upgrade. Yeah the 980ti makes a huge difference on the timeline, otherwise Film Convert would be unusable with just cpu alone.

@RealityStudio What kind of memory bandwidth ("winsat mem" at a cmd prompt) do you have currently with that platform, and are their any improvements you can make on that front?

Perhaps you can tweak the memory settings using "AMD Ryzen Master?"

I have Ryzen Master installed just to verify things, it runs at 3.2ghz for all 8 cores during a render at around 59 degrees with the stock cooler. I haven't tried tweaking any settings yet though, mostly because I want to run it all stock for a while to make sure everything is working ok. In my old tests on my previous machine I had found memory speed didn't make much difference to render times so I'm not expecting huge gains here either when I finally do get the ram running at full spec.

Memory bandwidth won't help with rendering since the processing via the codecs is still the bottleneck. On other tasks it could make a difference, when both the input and output will fit into RAM and the code is designed to take advantage. I found some numbers elsewhere to compare with other platforms for the memory data rate.

Great information on how Ryzen performs in Vegas Pro. Thanks for taking the time to report your findings.

No problem! I figure us Vegas guys are stuck with software encoders for the near future hence I figured this info could be useful to others here.

I was reading about the Ryzen benchmarks last week and I thought to myself... "too bad the Vegas forum isn't what it used to be, bet there would be some great discussions and excitement about this Ryzen release" Lo and behold, here's a discussion about Ryzen and Vegas. This place is still awesome. Thanks for sharing.

Very interesting numbers there, what I would do with my R7 1800X? 😀😁

@RealityStudio

What are the hard drive configurations and specs for the old Intel and the AMD PC's.

John EB

Old machine was a 1tb Samsing 840 Evo ssd, new machine is a 1tb Samsung 960 Evo m.2. But ssd speed never made a speed difference to me as it's all mostly cpu bound.

Test should be done with the SCS Benchmark project. That way we all have a point of reference.

Just a quick further comment regarding RAM: Terabytes of RAM is on the horizon, and then we really will need superscalar symmetric multiprocessors to take advantage. See IBM Power8. I bet Vegas won't be running on that platform anytime soon, however.

Thank you so much for this post! As suggested by OldSmoke, any chance you can run the SCS Benchmark rendered with MainConcept (CPU only)? That'd give a great point of reference to how our own systems compare.

Test should be done with the SCS Benchmark project...

http://download.sonycreativesoftware.com/whitepapers/vp11_benchmark.zip

Thanks for the link, I downloaded it and rendered it with the MC AVC setting they used and set the project as they requested which included using full 32bit color. I usually use 8bit color since that's what all my cameras are (AX100, A7S, etc) so I rendered both cases one at 32bit project setting and one at 8bit project setting. The result to render the entire project:

32bit: 2:05

8bit: 1:24

This is all completely stock so 2133mhz memory, no cpu overclocking and stock cpu fan. I used Ryzen Master to see what was going on during the render and all 8 cores were pegged at 3.2ghz the entire time with peak temperature being 60.75 degrees.

No idea if these numbers are good or not, but there they are. I see huge gains on my personal work so I'm super happy. I've read people are able to get 3.8ghz pretty easily on this cpu but I'm keeping it stock for now/

Your numbers are good. You have a two core advantage over me, but my six cores are running a little faster. You have a street price advantage, so far.

These are my results for the i7-6850K system in my profile.

On this system there appears to be no benefit in using the OpenCL render option for this test. I created this spreadsheet to sort the various options.

Just to compare with a 6 year old PC (3930K + GTX570) - MC 1080-30p render

- GPU OFF: 3min35s / 4min12s (8/10bits)

- GPU ON: 2min01s / 2min25s (8/10bits)

- GPU ON with Cuda: 37s / 52s (8/10bits)

Only 20% performance increase for CPU over 6 years is pretty low.

GPU encoding can't be beaten yet. I hope it will be implemented soon in Vegas...

Your numbers are good. You have a two core advantage over me, but my six cores are running a little faster. You have a street price advantage, so far.

These are my results for the i7-6850K system in my profile.

On this system there appears to be no benefit in using the OpenCL render option for this test. I created this spreadsheet to sort the various options.

Here are my results:

@John: Are you running the 6850K at stock speeds?

@GJeffrey

Only 20% performance increase for CPU over 6 years is pretty low.

Compared to??

@Old Smoke: These were run with a 13% overclock done by the ASUS utility. I can undo the overclock to run a baseline test in the future.