Photography question M3

[Chucks4me]

Check out this video on YT, it answers many questions in regards to performance and impact of RAM etc.
The tests are performed with a specific software but I think it still is very valuable.

He has many other videos comparing the performance if various M based systems.

 

[Miltz]

Check out this video on YT, it answers many questions in regards to performance and impact of RAM etc.
The tests are performed with a specific software but I think it still is very valuable.

He has many other videos comparing the performance if various M based systems.

Adobe products don’t behave this way and these programs are for things you do to a few images not general editing in a professional workflow.

 

[Allen_Wentz]

Check out this video on YT, it answers many questions in regards to performance and impact of RAM etc.
The tests are performed with a specific software but I think it still is very valuable.

He has many other videos comparing the performance if various M based systems.

Interesting guy but not really helpful IMO. In my experience it is not the 25-seconds versus 13-seconds comparisons (the comparisons that show well on UTube videos) that matter the most, because on those long activities your brain is off somewhere else whether it takes 12 seconds or 30. What matters is what happens as one makes some adjustment that may take 0.1 second versus 2 seconds depending upon RAM/CPU/GPU. Those comparisons show poorly in UTube videos seeking clicks.

Some easily measured things do matter, like how fast 200 pix can load from a CFExpress card, but we can see those in specs.

 

[Allen_Wentz]

Just offering my perspective and experience here, for whatever it's worth..... I have a 16" M1 MBP with 64 GB RAM and 4 TB storage. Why? Because, although I am not a professional photographer, I do tend to shoot in a variety of situations using either a Sony A1 or a Sony A7R V, and both of these have a generous number of megapixels (I think it's 51 on the A1 and 61 on the A7R V). I shoot RAW Uncompressed, and yes, these files take up a lot of space!

It's not a big issue when I'm spending an hour or two some afternoon just shooting a few macros or closeups, but it definitely is when I am out shooting wildlife and fire off the A1 set at Continuous High+. Doesn't take long for the rapid accumulation of frames on the memory card! I don't use Photoshop or LightRoom, I prefer DXO PhotoLab for my editing and I have a couple of other software programs for culling those massive high-speed-accumulated files prior to even taking them into that program.

I shoot pretty much every day so that's one reason for the larger SSD internal drive space, and of course as well I depend a lot on external drives, too, to archive and/or temporarily stash files when I don't have the time right away to work on them. I'll admit, I have a lot more fun shooting the images than I do processing them later!

For me, 64GB RAM seems to be the sweet spot in terms of handling those large files. Storage (archival and temporary) and RAM are really the most important factors as far as I'm concerned. Yes, of course power and speed of the processor is nice, too, and I definitely was surprised at the difference when I replaced my 2018 Intel MBP (2 TB storage, 32 GB RAM) with the M1 machine I have now -- wow!

For my particular situation I made the right choices, although I'll admit at the time I was buying the M1 MBP just short of a year ago now, I wondered if I were overreaching. My bank account was squealing in protest. Seems that I wasn't, after all, and this has worked out very well for me. (And, yes, it was expensive, but in the long run, worth it to me.) Everyone has his or her own preferences, priorities and financial concerns......

I too (D850) find that 64 GB RAM is the RAM sweet spot for 2023 still images work. But I will use the 2023 M2 Mac until 2028-2029 probably, so I put 96 GB RAM in it. There is no way that 2028 RAM needs will be the same as 2023.

 

[HawkTheHusky1902]

We disagree. IMO 16 GB RAM is not just fine for anyone dealing with images.

Oh well

 

[HawkTheHusky1902]

I too (D850) find that 64 GB RAM is the RAM sweet spot for 2023 still images work. But I will use the 2023 M2 Mac until 2028-2029 probably, so I put 96 GB RAM in it. There is no way that 2028 RAM needs will be the same as 2023.

Well true i guess but it depends on the app...must be painful for you having bought m2 max for so much money:/ when m3 max completely smokes it..

 

[HawkTheHusky1902]

I read up on Blender a bit and my guess is that for non-pro work the slowdown from lack of M3 would not be overly problematic. But IMO lack of RAM would be problematic. Just guessing of course, but from images experience.

Thank you, thats what i hope. I can save a lot of money by getting M1/M2. And btw, when i say get M1, i mean mostly a M1 Max with 64 gb ram.

 

[Allen_Wentz]

Well true i guess but it depends on the app...must be painful for you having bought m2 max for so much money:/ when m3 max completely smokes it..

Not painful at all, because I do not think that way. All my life tech has advanced every year and I consider that to be a very good thing. I am just thrilled to have had the power of M2 Max with 96 GB of RAM on board at my disposal for the last ~year. It rocks.

At some point I will become pained and I will buy the latest and greatest, but certainly not in the foreseeable future; not the (excellent) M3 or even M4 unless I get into something that wants hardware ray tracing. My still images workflow absolutely flies. Any slow parts are IMO areas where app engineers have not yet learned to take advantage of the UMA RAM, not the fault of the M2 Max hardware.

 

[Allen_Wentz]

So is it best practice to still keep the RAWs backed up after we've edited and exported photos?

I'm basically just culling and doing light edits of family trips. I planned to delete the RAWs permanently but now wonder if it's wiser to keep the original RAWs, at the expense of increased storage, in case I ever need to go back and re-edit anything...maybe use new software tools on really old RAWs later?? Does everyone typically keep the RAWs forever?

I keep the original RAW images forever after deleting the bad captures. That is the way I have my workflow set up. It requires having adequate mass storage, but I am not constantly accessing those files, they are just archives.

The reason I keep all pix is because in my work I quite frequently go back to repurpose an image. Either for an entirely different purpose or because I now want to pull a different crop from an image or because I want to document a condition at a past specific time and place.

 

[Matck06]

I own the MBP 16 M3 Max 16/40 48gb 1to, yesterday I tested with Lightroom a 310mpx panaroma opened with various curved layers + photoshop /spotify / mail /safari no memory pressure always in the green no slowdown, in fact one thing is simple most here do not know how the ram is managed, that's why most get 96 or 128 see 64gb not knowing absolutely their workflow, in fact the ram more you have more the system scales,

in a video by art is rights you can see a moment when they compare the MBP m3 max binned and unbinned and unfortunately for export import or noise reduction IA 36, 96, 128gb won't give you any more speed, where ram came in handy was on a 56gb photoshop file, where 64gb ram was the ideal point, but 56gb I don't think most people handle such files every day, so if that's the case the question of ram no longer arises, and if it's occasional, is it really worth it?

we can also see that the Topaz suite uses the neural engine for AI denoising, unlike Lightroom which uses the GPU, which gives much shorter processing times on Topaz, on Lightromm I'm at 12-13sec for denoising an A7IV 33mpx file, surely I'll halve this time using topaz.

 

[Allen_Wentz]

I own the MBP 16 M3 Max 16/40 48gb 1to, yesterday I tested with Lightroom a 310mpx panaroma opened with various curved layers + photoshop /spotify / mail /safari no memory pressure always in the green no slowdown, in fact one thing is simple most here do not know how the ram is managed, that's why most get 96 or 128 see 64gb not knowing absolutely their workflow, in fact the ram more you have more the system scales,

in a video by art is rights you can see a moment when they compare the MBP m3 max binned and unbinned and unfortunately for export import or noise reduction IA 36, 96, 128gb won't give you any more speed, where ram came in handy was on a 56gb photoshop file, where 64gb ram was the ideal point, but 56gb I don't think most people handle such files every day, so if that's the case the question of ram no longer arises, and if it's occasional, is it really worth it?

we can also see that the Topaz suite uses the neural engine for AI denoising, unlike Lightroom which uses the GPU, which gives much shorter processing times on Topaz, on Lightromm I'm at 12-13sec for denoising an A7IV 33mpx file, surely I'll halve this time using topaz.

Your M3 Max experience with images and RAM is consonant with my M2 Max experience. November 2023 48 GB RAM works well and is not limiting. November 2026 (approx.) and thereafter we are unlikely to still be seeing the same 11/23 RAM usage, and even though 48 GB will still be working it will IMO likely be limiting.

TBH I do not care if Denoise takes 6 seconds or 16 seconds because after 2 sec my brain has moved on anyway. What I care about, a lot, is milliseconds. Working with images and trying some adjustment (e.g. "let's try 40% opacity here"), above some speed the adjustment timing just fits unnoticed into my mental workflow, whereas a few milliseconds slower and the thought process gets interrupted; the creative flow gets stifled. Different apps, different brains, etc. all get interrupted differently of course, but IMO the concept is relevant to all visual design folks.

Unfortunately evaluating milliseconds makes for poor click-bait YouTube videos, so we do not see those analyses. As to denoise I care mostly about the results not so much about how long it takes. YMMV. Actually YM will vary, because you have that rocket ship M3 Max!

 

[NeonNights]

For me, I think I'll be fine with 36GB for many years. With iStat Menus I never went beyond 32GB used on my 64GB M1 Max MBP 16". By the time 36GB becomes a limiting factor for me, it will probably be high time for an upgrade with the latest features and ports anyways.

I really do like the snappy response of the M3 series. In DxO PhotoLab the M1 Max took about a second to render previews, which is still really quick, but the M3 Max is basically instantaneous (sometimes maybe 0.5 seconds if I've scrolled through a few dozen images). Love the M3 Max 14/30 core!

 

[Chancha]

Get the M3 Max. It is hands down the most responsive Lightroom interfacing experience I have ever had. My main desktop is a M1 Max Mac Studio, running Adobe suites including LR on it is fine, but only after getting the M3 Max MBP I realize the culling and adjustment speed actually has room to improve.

Now the OP's original question is if the M3 Pro is adequate. I have no experience with it, and together with the M3 Max it got the same generational improvements in particular the single core speed, so it may already be enough for OP. However this year Apple has separated the two tiers with a rather clear cut value proposition, where if you see yourself even slightly needing more than the Pro, you are getting a lot more by going up to the Max, instead of upgrading say just the RAM on the Pro.

 

[Matck06]

Your M3 Max experience with images and RAM is consonant with my M2 Max experience. November 2023 48 GB RAM works well and is not limiting. November 2026 (approx.) and thereafter we are unlikely to still be seeing the same 11/23 RAM usage, and even though 48 GB will still be working it will IMO likely be limiting.

TBH I do not care if Denoise takes 6 seconds or 16 seconds because after 2 sec my brain has moved on anyway. What I care about, a lot, is milliseconds. Working with images and trying some adjustment (e.g. "let's try 40% opacity here"), above some speed the adjustment timing just fits unnoticed into my mental workflow, whereas a few milliseconds slower and the thought process gets interrupted; the creative flow gets stifled. Different apps, different brains, etc. all get interrupted differently of course, but IMO the concept is relevant to all visual design folks.Yes I agree with you, I also prefer a lower latency time to adjust my sliders and opacity mask etc... in total fluidity than to gain 10 sec or 30 sec on a lightroom photo export and I must admit that coming from an already fast MBP 16 m1 pro 32gb I'm bluffed by the reaction time of the M3 max everything is really instantaneous.

Unfortunately evaluating milliseconds makes for poor click-bait YouTube videos, so we do not see those analyses. As to denoise I care mostly about the results not so much about how long it takes. YMMV. Actually YM will vary, because you have that rocket ship M3 Max!

Yes I agree with you, I also prefer a lower latency time to adjust my sliders and opacity mask etc... in total fluidity than to gain 10 sec or 30 sec on a lightroom photo export and I must admit that coming from an already fast MBP 16 m1 pro 32gb I'm bluffed by the reaction time of the M3 max everything is really instantaneous.

 

[Macshroomer]

So here is what a pro is doing and why...

36 year career in commercial, editorial and fine art photography, 29 years of it digital. 1.5M+ images with some film digitized, film shot now gets printed in a real darkroom. My storage is fast, diversified and large at well over 300TB.

For reasons of kicking out hundreds of 45-100MP raw files after culling on jobs and projects that can come in at 3,000 to 35,000 when ingested, I am maxing out my 16" M3 in every lane except for I won't be doing the 8TB of storage since my three OWC SSD's for travel backup are 4TB.

Some of my Hasselblad X2D files at 100MP get stitched into wall murals 15' feet wide at 300dpi for hotel and other commercial uses, this process loves all the hardware it can get.

My M2Max 96GB has been truly great, but I am using the 100MP cam a lot more so I will order the $6K 16" M3Max with 128GB of ram and a 4TB SSD.

 

[Filmx]

For me, I think I'll be fine with 36GB for many years. With iStat Menus I never went beyond 32GB used on my 64GB M1 Max MBP 16". By the time 36GB becomes a limiting factor for me, it will probably be high time for an upgrade with the latest features and ports anyways.

I really do like the snappy response of the M3 series. In DxO PhotoLab the M1 Max took about a second to render previews, which is still really quick, but the M3 Max is basically instantaneous (sometimes maybe 0.5 seconds if I've scrolled through a few dozen images). Love the M3 Max 14/30 core!

What files are you working with? Have you edited any higher mp (45+) files in Lightroom Classic? Do you notice any lag when zooming to 200% or more, using the heal brush, brushing/masking, etc? Any delay from the time you click? Have you noticed 36GB being a bottleneck in any of your workflow?

 

[Filmx]

Get the M3 Max. It is hands down the most responsive Lightroom interfacing experience I have ever had. My main desktop is a M1 Max Mac Studio, running Adobe suites including LR on it is fine, but only after getting the M3 Max MBP I realize the culling and adjustment speed actually has room to improve.

Now the OP's original question is if the M3 Pro is adequate. I have no experience with it, and together with the M3 Max it got the same generational improvements in particular the single core speed, so it may already be enough for OP. However this year Apple has separated the two tiers with a rather clear cut value proposition, where if you see yourself even slightly needing more than the Pro, you are getting a lot more by going up to the Max, instead of upgrading say just the RAM on the Pro.

Which spec do you have?

 

[Tdude96]

He has many other videos comparing the performance if various M based systems.

I agree, take a look at some of his comparisons that look interesting to you https://www.youtube.com/@ArtIsRight/videos

He has a lot of videos that do comparisons between the different generations, as well as doing testing on different configurations within the same generation, including the different processor, RAM, and drive capacity configs. He tests against a few different real-world scenarios using a variety of professional apps, including the Adobe products (Lightroom & Photoshop mostly) and the Topaz products, rather than relying on benchmarks which don't do a lot to demonstrate real world performance.

I have to disagree with the comment that his videos are lacking value because he often measures tasks in time elapsed.

"Hey, I ran 100 photos with these specs through this process, it took 41 minutes 40 seconds to run those on this machine and 25 minutes to run them on that machine, so the first machine averaged one every 25 seconds and the second averaged one every 15 seconds."

To me, that's valuable information about how the machine performs on actual similar tasks I might be doing. It gives me something to think about that's relevant to my workflow. Would saving 16 minutes and 40 seconds per 100 images make an $XXX difference in my work? Probably yes. But if it was a closer comparison, say a 1 second difference per image, saving a grand total of 100 seconds across 100 images? Probably not.

In my opinion, that's more practical information than looking at a GeekBench score and trying to judge the merits of a machine without knowing what its real world performances is at a specific task. I prefer having an idea of how much time a machine might save in my workflow or what new capabilities I will gain with a new machine.

 

[Chancha]

Which spec do you have?

The base M3 Max 16". And to add on with what you asked the other poster above, my LR library consists of mostly D800 then D850 RAWs, so 36-46MP NEFs. Generating 1:1 preview from scratch is probably the only thing I need to wait for it to finish with a few seconds, and I suspect it has to do with my RAWs being on an external NVMe SSD instead of internal. The other culling and adjustments are just smooth like butter. Haven't tried much brushing or smart masking etc but if they choke, then well you need an M3 Ultra to get any better then, or go PC.

 

[Filmx]

I agree, take a look at some of his comparisons that look interesting to you https://www.youtube.com/@ArtIsRight/videos

He has a lot of videos that do comparisons between the different generations, as well as doing testing on different configurations within the same generation, including the different processor, RAM, and drive capacity configs. He tests against a few different real-world scenarios using a variety of professional apps, including the Adobe products (Lightroom & Photoshop mostly) and the Topaz products, rather than relying on benchmarks which don't do a lot to demonstrate real world performance.

I have to disagree with the comment that his videos are lacking value because he often measures tasks in time elapsed.

"Hey, I ran 100 photos with these specs through this process, it took 41 minutes 40 seconds to run those on this machine and 25 minutes to run them on that machine, so the first machine averaged one every 25 seconds and the second averaged one every 15 seconds."

To me, that's valuable information about how the machine performs on actual similar tasks I might be doing. It gives me something to think about that's relevant to my workflow. Would saving 16 minutes and 40 seconds per 100 images make an $XXX difference in my work? Probably yes. But if it was a closer comparison, say a 1 second difference per image, saving a grand total of 100 seconds across 100 images? Probably not.

In my opinion, that's more practical information than looking at a GeekBench score and trying to judge the merits of a machine without knowing what its real world performances is at a specific task. I prefer having an idea of how much time a machine might save in my workflow or what new capabilities I will gain with a new machine.

His videos are some of the most helpful I've found and he has responded to some of my comments on them. But I do agree that they still don't really give the additional info professional photographers need, what are the actual workflow times between the different configs when editing? For example when zooming into high resolution files (40-100mp), brushing while zoomed into 400%, quickly switching between images, using the "edit in Photoshop" command then doing work in Photoshop (sometimes with LR running other tasks in the background).

These tasks significantly slow down my 2020 M1 MBP. To the point where I click an image to zoom and have to wait several seconds for it to respond and fully zoom in and render the image at 100-200%+. Then I'll use the heal tool to remove a spec of dust and wait several seconds for it to respond. This turns what could be a 30-60 second edit into several minutes. Multiply that times a few hundred or thousand images and it is a painful workflow.

If the M3 Pro is similar in speed in that regard to the M3 Max then it may be worth saving the money and going for a cheaper config. Same with the binned vs unbinned M3 Max. But AFAIK these comparisons don't exist. The tasks where I can walk away from my computer and save 5 minutes of exporting or batch denoising with a $4500 M3 Max vs a $2500 M3 Pro are far less important than when I'm actually working on each individual image with tons of zooming, masking, healing/cloning, etc.

 

[Filmx]

The base M3 Max 16". And to add on with what you asked the other poster above, my LR library consists of mostly D800 then D850 RAWs, so 36-46MP NEFs. Generating 1:1 preview from scratch is probably the only thing I need to wait for it to finish with a few seconds, and I suspect it has to do with my RAWs being on an external NVMe SSD instead of internal. The other culling and adjustments are just smooth like butter. Haven't tried much brushing or smart masking etc but if they choke, then well you need an M3 Ultra to get any better then, or go PC.

Couldn't the additional cores and memory bandwidth of the unbinned chip perform better if those tasks choke on the binned chip you have? That's what I'm trying to figure out. Is there any difference in actual editing tasks (zooming, masking, healing, etc.) in the M3 Pro vs binned M3 Max vs unbinned M3 Max (and also do thermals affect these tasks on the 14" binned/unbinned M3 Max)

If you get a chance I'd be curious to know if you experience any lag when brushing, healing, etc when zoomed into 200% or more on your D850 raws. My 45mp R5 files cause significant lag in all of these tasks on my 2020 M1 MBP 16GB.

 

[NeonNights]

What files are you working with? Have you edited any higher mp (45+) files in Lightroom Classic? Do you notice any lag when zooming to 200% or more, using the heal brush, brushing/masking, etc? Any delay from the time you click? Have you noticed 36GB being a bottleneck in any of your workflow?

My Sony RX10 IV is only 20MP, and I'm an amateur when it comes to editing. The M1 Max was more than enough power for my workflow and I didn't feel encumbered by it at all. Suffice it to say the M3 Max is even more power than I need. The 64GB M1 Max only seems slow now that I'm using the 36GB M3 Max.

I haven't used Lightroom in over a year and no longer have a license. I switched over to Affinity Photo 2 and it consumes 12.5 GB with five images open. Zooming, tone sliders, and healing brush all work instantly with 36GB. It isn't until I tried to load 10 images that I felt the system slow down when switching between open files, but I normally wouldn't open that many for editing purposes. For me, 36GB is still more than plenty.

When I had the 64GB M1 Max I never went over 50% RAM usage in four months so a lot of it went underutilized in my use case, but it was a killer deal for 64GB and 4TB over the Summer. I just saw the same system is now $200 cheaper from B&H, a great machine if you really need the extra RAM and storage.

 

[Matck06]

Which spec do you have?

I tested the 14/30 36gb configuration of the m3 max with 61mpx files (A7rv) is a 300mpx panorama and no worries with local masks or brushes it was perfectly fluid no lag despite grain and adjusted cursor. On Art is right's YT video channel, he compares an M1 pro 10/16 16gb vs m1 pro 8/14 32go the 32gb model handles a 200mpx file without a hitch with several adjustments and no lag with the brush product, so imagine with the basic m3 max 14/30 here's the video (

)

 

[Chancha]

Couldn't the additional cores and memory bandwidth of the unbinned chip perform better if those tasks choke on the binned chip you have? That's what I'm trying to figure out. Is there any difference in actual editing tasks (zooming, masking, healing, etc.) in the M3 Pro vs binned M3 Max vs unbinned M3 Max (and also do thermals affect these tasks on the 14" binned/unbinned M3 Max)

If you get a chance I'd be curious to know if you experience any lag when brushing, healing, etc when zoomed into 200% or more on your D850 raws. My 45mp R5 files cause significant lag in all of these tasks on my 2020 M1 MBP 16GB.

I can dive in to test what you listed over the weekend. However, I don’t recall doing those was terribly slow or anything on my main desktop which is a M1 Max base Studio. Matter of fact I used to be doing this on M1 Pro 14” and 16” and they weren’t choking. But I wasn’t deliberately benchmarking or anything, I am going off my memory on whether or not I faced frustration while doing anything in LR (classic).

With all Apple Silicon Macs, the only time I could (easily) hit a performance wall was when the M1 was new, I bought a base M1 Air to test. Doing something like HDR merging of dozens sets of 5 exposure bracketing, was enough to bring it to its knees, the whole computer became unresponsive not just LR Classic. I imagine with your case, even assuming your M1 MBP has 16GB, the lack of cores definitely played a role in how it choked with LR.

 

[Filmx]

I tested the 14/30 36gb configuration of the m3 max with 61mpx files (A7rv) is a 300mpx panorama and no worries with local masks or brushes it was perfectly fluid no lag despite grain and adjusted cursor. On Art is right's YT video channel, he compares an M1 pro 10/16 16gb vs m1 pro 8/14 32go the 32gb model handles a 200mpx file without a hitch with several adjustments and no lag with the brush product, so imagine with the basic m3 max 14/30 here's the video (

)

Thanks for this link. I didn't realize he did these types of tests but they are very helpful. Does he still do these? I haven't seen any like this from him with the M3 line, showing actual workflow with high resolution individual images.