Blackmagic Forum

Let me caution everyone that this is not or even near a scientific test. It is more of a personal experience trying out the limits of DR Studio 18 on a barebones Apple MacBook M2. The system had the base 8/8 core cpu/gpu, 8gb of RAM, and 256gb storage. When I tried this computer out, I set some basic parameters like my timelines where 2K DCI scope and that I would try out one short film that was 4K DCI scope and a 2K SCI scope feature length. The short film was 11 minutes in runtime and the feature (which can be watch on Amazon Prime) is almost 160 minutes long. Both where 23.976fps and shot in Prores 422 using the URSA Mini Pro 4.6K G2 and Sony A7iii (with Atomos Ninja). My workflow is ACES cct 1.3.

These are old archived projects so all I had to do is import them into DRS. I had to use the Studio version because the project had some effects and NR, and other features only available on the Studio version.

I am amazed that both imported with minimum hitch, like having to delink a few files and adding custom LUTS that were made for the films respectively. The only thing I set on DRS was to set the render cache to ‘Smart’ and use the lightest codecs for creating the optimized media so it can be a little faster nd small file sizes.

I kid you not that both played back on the timeline at 23.976 smoothly, scrubbing the clips on the timeline was so smooth. The only thing I notice is at the beginning of the run inside DRS, there is a slight struggle to get to 23.768fps (always green) from beginning to end. I can’t believe what I was seeing. With the feature, once in a while I get it to stagger (red) but when I get back to it again, it was it’s playing smoothly. During processing, NR tend to stress the system out but very slightly.

This is not a replacement for any workstations and I am not suggesting that this is the right computer for a serious editor and colorist, but I can say if you’re working mostly on HD/FHD and 2K projects with some temporal work, and have render caching on, it’s a nifty option. I’m impressed. I am just sharing my experiences trying out the Mac M2, especially coming from a PC user and environment for which I do my professional editing and grading.

Rending out the timeline as 2K DCI scope in H.265 took 13.10 minutes for the 11 minute short film and interestingly the feature film only took 19.20 minutes to render as 2K DCI scope in H.265. I am still doing a lot of indie features that are 80 to 160 minutes in runtime length but they are mostly for 2K DCI 23.976 or 24fps and my experiment with the M2 seem to tell me that it’s a good budget laptop that can support a no so demanding workflow.

I also run the BMD Speed Test tool on the same MacBook M2 and this is the result I have been getting:

If you’re on an Apple MacBook Pro M2, Apple Studio Max M1, and/or Apple Studio Ultra M1, if you can run the BRAW speed test software and paste the results in this post, that will be most appreciated. Please include the configuration (RAM, CPU and GPU cores, etc.).

So no stats of BRAW speed on Mac silicons.

M1 MacBook Pro 14" M1 Max Chip, 10 core CPU, 32 Core GPU, 64GB Unified Memory, 1TB drive. You didn't mention this machine in your request but seeing as nobody else has replied as yet here are the stats. Pretty impressive stuff. I guess the Studio M1 Max would be pretty similar. These are no ordinary laptops. It eats BRAW for breakfast and ProRes Raw for lunch without a hiccup. I've had it almost a year and it is a truly amazing computer.

@Michael: What version of BRAW Speed Test are you running? I would expect the current 2.7 version to show 8K results, not 4K. If you install a current Resolve 18.0.4 version, that will install the current BRAW 2.7 version.

@Ellory: Sort of surprised you are not getting more responses to this. This is what I get on my M1 Pro 14" MBP:

Dwaine Maggart wrote:@Michael: What version of BRAW Speed Test are you running? I would expect the current 2.7 version to show 8K results, not 4K. If you install a current Resolve 18.0.4 version, that will install the current BRAW 2.7 version.

I'm still on 17.4.6 Dwaine. I've been hanging on a bit before moving to v18 mainly because I need to go back through some of the BMD tutorials (mainly color) and I want to minimise potential confusion where changes have been made since the tutorials were created. I worked through all the color tutorials when I got Resolve back in May but it is a complex program and I feel I need to revise some stuff. I will upgrade soon for sure.

Interestingly, our results are very similar for the 14" MacBook Pros.

I’m really impressed with how it can handle BRAW. For your M1’s to be able to handle up to 8K @24/25fps is astonishing. I don’t know if this tool is stressing things out BRAW with effects and NR type load by still impressive. With that said, I hope to see others on the Max and Ultra. I plan to invest on a Max or Ultra as these numbers are looking good. I hope there will be others who would post their numbers and that there will be some kind of stress test (if this tool is not it) from BMD that will benchmark with high loads (effects, NRs, many nodes, etc.). Thanks Mick and Dwaine.

there is something wrong with the M2
here the result of my 12 year old cheese grater macpro 5.1

Mark Foster wrote:there is something wrong with the M2

What is wrong with the M2? I run the speed test over and over again and the results where around those numbers. I’m using the version 2.7 of the tool and 2.7 of BRAW.

I think what is important are the checks on what works and the numbers on the how fast sections. That shows what BRAW compression it can support, as I understand the tool.

Ellory Yu wrote:I don’t know if this tool is stressing things out BRAW with effects and NR type load by still impressive.

This tool shows the performance of decoding BRAW clips, there's no stress test in it to simulate DaVinci Resolve processing.

I'd post my results but Dwaine already did with the same laptop.

Ellory Yu wrote:

Mark Foster wrote:there is something wrong with the M2

What is wrong with the M2? I run the speed test over and over again and the results where around those numbers. I’m using the version 2.7 of the tool and 2.7 of BRAW.

I think what is important are the checks on what works and the numbers on the how fast sections. That shows what BRAW compression it can support, as I understand the tool.

I think your problem comes from the 8-core CPU/GPU and especially the 8GB RAM,
because here the RAM is also part of the VRAM that is available.

this computer is a good writing machine, but not sufficiently equipped for video work.


.

Mark Foster wrote:

Ellory Yu wrote:

Mark Foster wrote:there is something wrong with the M2

What is wrong with the M2? I run the speed test over and over again and the results where around those numbers. I’m using the version 2.7 of the tool and 2.7 of BRAW.

I think what is important are the checks on what works and the numbers on the how fast sections. That shows what BRAW compression it can support, as I understand the tool.

I think your problem comes from the 8-core CPU/GPU and especially the 8GB RAM,
because here the RAM is also part of the VRAM that is available.

this computer is a good writing machine, but not sufficiently equipped for video work.


.

I agree with you that it is not sufficiently equipped for heavy video work but it had impressed me because, if you read my (OP) 1st and 3rd paragraphs, I talked about how well it performed on editing, some color grading, and rendering of 23.976fps 2K footage - one was a 10 min short film and the other a very lengthy feature film, the same project that I used when I worked on it and now is streaming on Amazon Prime. So I am not saying it is the right laptop in terms of what it has, but for what it has it was impressive enough to tackle some FHD and low frame rate footage with ease. The BRAW speed results also shows numbers that backs up the premise of my experience. Anyway, I am not trying to sell anyone on using a 8 core cpu/gpu with 8gb ram. I’m just pointing out that the silicon architecture is rather capable even with minimum core and ram specs.

After all this discussion, I now wonder if there is an advantage of getting the Mac Studio Ultra with more Cpu/gpu cores and RAM over a Studio Max. This is a decision that’s hard without benchmarks because Mac Studios are not upgradable - can’t just upgrade the cpu or gpu, or Add more RAM.

If I’m just editing/grading 2/4/6K at most feature length indie films shot in BRAW and Prores most of the time with frame rates < 60fps, with an ACES cct workflow, some effects and NR, will the Max be sufficient?

Even my machine is sufficient. BRAW and ProRes are really easy to compute.
And then, Apple's top of the line is always pretty much overpriced and aimed at those who want to show off. One step below is usually the better price/performance ratio. The Ultra might be interesting if you need that many TB ports, but not as much for computing power.

Ellory Yu wrote:After all this discussion, I now wonder if there is an advantage of getting the Mac Studio Ultra with more Cpu/gpu cores and RAM over a Studio Max. This is a decision that’s hard without benchmarks because Mac Studios are not upgradable - can’t just upgrade the cpu or gpu, or Add more RAM.

If I’m just editing/grading 2/4/6K at most feature length indie films shot in BRAW and Prores most of the time with frame rates < 60fps, with an ACES cct workflow, some effects and NR, will the Max be sufficient?

I don't think one should be thinking just about what is sufficient now when buying one of these non-upgradeable Macs. Think about will it still be more than sufficient in a few years time. That is why I pretty much maxed out when getting my M1 Max MacBook Pro.

Basically the M1 Max Studio is the same computer as the M1 Max MacBook Pro in a different enclosure so the benchmarks would be very similar. The Ultra is basically two of these in one and the performance apparently reflects this depending on what one is doing.

It really depends on what you can afford to spend. One thing I would suggest is to prioritise RAM and CPU/GPU cores over internal storage, which is very expensive once you go above 1 or 2 TB. External SSDs are a lot cheaper and the differences in speed are not noticeable.

There are rumours of M2 MacPros next year but I expect these will be very expensive.

I second your suggestion regarding RAM and cores. But from most tests I've seen until now, the Ultra does not really show double performance over the single chip. We don't really know yet if this is due to thermal throttling, or the link being not as efficient as Apple says, or software not yet optimised for it.

Only in the latter case it means future proofing.

mickspixels wrote:I don't think one should be thinking just about what is sufficient now when buying one of these non-upgradeable Macs. Think about will it still be more than sufficient in a few years time. That is why I pretty much maxed out when getting my M1 Max MacBook Pro.

Ideally that is the case with most folks and I agree. For me, I based it on ROI. With something in the 3-5K range, a couple of projects will pay it off so I prefer not to be stuck with a model that in a years time a new one will be in place, like an M2.

mickspixels wrote:Basically the M1 Max Studio is the same computer as the M1 Max MacBook Pro in a different enclosure so the benchmarks would be very similar. The Ultra is basically two of these in one and the performance apparently reflects this depending on what one is doing.

Further research has it that the M1 Max Studio is going to be faster that the M1 Max MBP because it is over clocked. I’m no hardware wiz so just sharing what I read on Apple forums. The Ultra is definitely different. Again, based on my use cases, and from what Uli said, the Max is sufficient. Yes, totally agree on more cores and RAM over storage.

Mark Foster wrote:I think your problem comes from the 8-core CPU/GPU and especially the 8GB RAM,
because here the RAM is also part of the VRAM that is available. this computer is a good writing machine, but not sufficiently equipped for video work.

That's my opinion as well. I think 16GB would be OK, and 32GB would be better still. 8GB ain't gonna work for the kind of work I do.

Marc Wielage wrote:

Mark Foster wrote:I think your problem comes from the 8-core CPU/GPU and especially the 8GB RAM,
because here the RAM is also part of the VRAM that is available. this computer is a good writing machine, but not sufficiently equipped for video work.

That's my opinion as well. I think 16GB would be OK, and 32GB would be better still. 8GB ain't gonna work for the kind of work I do.

I agree. Not the kind of work most of us do. However for low resolution and low frame rates with codecs that are easy to process, not so much NR and effects, or just straight wise editing (no grading), might be sufficient for. Use case would be YT deliveries is one I can think of. My interest got perk because of how the architecture of the system makes it possible to deliver some performance with little resources.

The M2 Air also has a 10core GPU and 24gb RAM option which I feel would be pretty capable for Offline/Proxy editing.

Should be plenty for proxies, yes.