Currently I'm on a MBP M1 Max and Topaz PhotoAI is taking somewhere around 30-40 seconds per each Sony A1 50Mb file.
Who is using an M4 Max MBP and PhotoAI and what improvement in speed have you seen using Topaz?
I don't use Topaz, but if this helps: a 40mp raw file converted and output to a full-sized 16-bit psd from CaptureOne and opened up ready to work in PS, takes my M4 max MBP 2-3 seconds.
Crop in Photomechanic, import to LR, hit edit (via tiff copy) in Topaz PhotoAI
Inside Topaz PhotoAI:
- using denoise at standard
- using sharpen at strong
Though I don't have an M4 Max, I can give you benchmarks for my M4 Mini Pro with 48 GB of RAM
For a RAW Nikon 58 MB file, it took less than a 4 sec for NR to a DNG.
For a .tiff Nikon file of about the same size for denoise/standard, sharpen/strong it took less than 2.5 sec for a single file.
What time period are you measuring? I'm particularly interested in the time interval defined as the beginning of pushing the Save to LR button until the tiff is re-displayed in LR.
p.1 #10 · M4 MBP Max and Topaz PhotoAI performance
schlotz wrote:
So far nobody with a MBP M4 MAX running LR and Topaz PhotoAi ?
Well, a few of us tried to help and you obviously shirked the information instead of trying to draw logical and reasonable conclusions from it. Your attitude towards this information then seemed kind of arrogant instead of appreciative; like seriously, nobody’s using the exact combo I want information on? 🤷🏼♂️ But know the M4 is FAST for image processing, significantly faster than an M1, I would estimate on the order of 10x or even 20x for the imaging I do. With that, I’ll just say good luck to you in your endeavor.
p.1 #12 · M4 MBP Max and Topaz PhotoAI performance
Far from it and definitely no arrogance involved what so ever. Simply looking for experiences with this configuration vs previous equipment in order to help decide if moving from a MBP M1 Max is worth the effort. Appreciate those who have responded. Hoping to hear from others.
Jack Flesher wrote:
Well, a few of us tried to help and you obviously shirked the information instead of trying to draw logical and reasonable conclusions from it. Your attitude towards this information then seemed kind of arrogant instead of appreciative; like seriously, nobody’s using the exact combo I want information on? 🤷🏼♂️ But know the M4 is FAST for image processing, significantly faster than an M1, I would estimate on the order of 10x or even 20x for the imaging I do. With that, I’ll just say good luck to you in your endeavor.
p.1 #13 · M4 MBP Max and Topaz PhotoAI performance
Agreed most likely it is. Using it because of the need to quickly produce acceptable results for client while on deadline during a soccer match. Topaz isn’t the best for sure and don’t use it for other work. Speed in this particular case here is the priority
RoamingScott wrote:
Things I know:
- The M4 shreds through editing in LR and Photoshop
- Topaz PhotoAI is a non-serious program for non-serious post processing
p.1 #14 · M4 MBP Max and Topaz PhotoAI performance
Is your workflow jpeg based?
If not, would use of Denoise in Lightroom be an option to reduce the workflow stages since you're already moving images to Lightroom after initial editing in Photo Mechanic?
I'm not familiar with the file options from Sony's cameras but for me with Canon, I switched to Canon's lossy compressed RAW (CRAW) option for most work. Its smaller storage footprint isn't significantly larger than low compression jpegs and it's compatible with Denoise, whereas jpegs are not.
With the usual in-camera tagging during an event and use of fast cards, reader and computer, I haven't found the compressed RAW workflow overly taxing compared to jpegs.
That said, Denoise relies heavily on GPU count and seems less affected by generational M chip improvements and more by core count. I'm not familiar with Topaz, but if it's highly CPU dependent, then for example the various tests from Art is Right suggest that strictly based on CPU, the M4 is potentially ~3x faster than M1 for certain tasks. What @armd indicated with the Mini M4 Pro is substantially faster than what you're seeing with the M1 Max. So if you were to stay with Topaz, then it would seem M4 would significantly cut down on this aspect of your workflow.
The question about upgrade benefits from the M1 generation to the M4 definitely interests me. I'm on a lowly M1 Pro and have recently introduced 45MP images to my previously strictly 24MP workflow. The M1 Pro is noticeably bogging down a lot more and exports are taking considerably longer (no surprise). For me, I'm mostly interested in Denoise and AI masking performance and can be working on hundreds or thousands of images, so with the 45MP files taking at least 2x longer to denoise, mask, etc., it's really adding up. Based on Art is Right it looks like for Denoise, going from M1 Pro 10/16 cores to M4 Max with 16/40 cores might be 3x faster. Added to this, at least for me, I do most of this work at my home office and going with a Studio instead of MBP could potentially benefit from additional performance due to improved thermal control.
I think bottom line is M4 will be faster, without doubt. How much will be workflow dependent. If you're already set to potentially spend the money, it might be worthwhile making the purchase to coincide with a period when you will be doing this work and use Apple's 2-week return period to determine if the configuration actually meets your needs.
p.1 #15 · M4 MBP Max and Topaz PhotoAI performance
Thanks for the input Ron. Workflow here is RAW. I do like the results of LR's Denoise Ai but the time it takes is far too slow on my M1 Max for what I need field side under a time constraint. Sad thing is I really am not that impressed with Topaz Photo AI either. It's speed is not great and the output is sometimes funky which I have to catch and make sure those are not sent. Looks like I need to spend some more time evaluating what the M4 Max with top gpu core count can do to shorten LR's denoise.
p.1 #16 · M4 MBP Max and Topaz PhotoAI performance
What specific results are you looking to achieve quickly? Noise reduction, or perhaps skin smoothing, or is it more just wanting a unified look across a broad range of images?
p.1 #17 · M4 MBP Max and Topaz PhotoAI performance
Matt, have you tried denoising and sharpening with DxO Photolab or PureRaw?
I have used Topaz Photo AI since the early days and own the latest version of this app. Upscale is the best tool there. Remove can be useful occasionally. Adjust Lighting works sometimes but cannot be trusted, the same as Balance color. However, I totally avoid Denoising on tiff files and Sharpening can produce ugly artifacts. Topaz Photo AI has been justly heavily criticized. It seems that the program developers at Topaz are trying to come up with new tools, that actually don't work reliably or don't work at all, however the existing tools see no improvement and I feel increasingly disappointed with Photo AI. The "artificial intelligence" in this app is still as intelligent as a small child. Only Upscale is the tool that (surprisingly!) works reliably and quite well.
If you haven't tried denoising and sharpening in the DxO apps, give them a try. DeepPRIME 3 should be reasonably fast on your laptop and the quality is good.
p.1 #18 · M4 MBP Max and Topaz PhotoAI performance
Jack Flesher wrote:
What specific results are you looking to achieve quickly? Noise reduction, or perhaps skin smoothing, or is it more just wanting a unified look across a broad range of images?
Photos are peak soccer action to be submitted to the team and USL league right prior to the end of the first half and by the end of the second half. Given they are night matches that start at 7pm, raw captures range in ISO level that require denoising followed by sharpening.The worst of the lot are the second half photos especially early in the year and late autumn where the sun goes down earlier. Starting about now thru mid September they are not so bad and in fact I have a set of automatically applied presets during the import to LR which cleans them up quite well. It's those two extreme ends of the season where extra effort is required to produce acceptable images for transmitting. Then there is the USL Women's League matches that are all held indoors. Nice facility but the LED lighting is far from adequate when it comes to action photography. This dungeon requires ISO 25,600 and higher. All of these photos require a significant amount of editing to produce acceptable results.
p.1 #19 · M4 MBP Max and Topaz PhotoAI performance
ruthenium wrote:
Matt, have you tried denoising and sharpening with DxO Photolab or PureRaw?
I have used Topaz Photo AI since the early days and own the latest version of this app. Upscale is the best tool there. Remove can be useful occasionally. Adjust Lighting works sometimes but cannot be trusted, the same as Balance color. However, I totally avoid Denoising on tiff files and Sharpening can produce ugly artifacts. Topaz Photo AI has been justly heavily criticized. It seems that the program developers at Topaz are trying to come up with new tools, that actually don't work reliably or don't work at all, however the existing tools see no improvement and I feel increasingly disappointed with Photo AI. The "artificial intelligence" in this app is still as intelligent as a small child. Only Upscale is the tool that (surprisingly!) works reliably and quite well.
If you haven't tried denoising and sharpening in the DxO apps, give them a try. DeepPRIME 3 should be reasonably fast on your laptop and the quality is good. ...Show more →
Actually I haven't tried DXO software. Guess it's time to do so. You think DeepPRIME3 should be my first try, correct?
p.1 #20 · M4 MBP Max and Topaz PhotoAI performance
schlotz wrote:
Photos are peak soccer action to be submitted to the team and USL league right prior to the end of the first half and by the end of the second half. Given they are night matches that start at 7pm, raw captures range in ISO level that require denoising followed by sharpening.The worst of the lot are the second half photos especially early in the year and late autumn where the sun goes down earlier. Starting about now thru mid September they are not so bad and in fact I have a set of automatically applied presets during the import to LR which cleans them up quite well. It's those two extreme ends of the season where extra effort is required to produce acceptable images for transmitting. Then there is the USL Women's League matches that are all held indoors. Nice facility but the LED lighting is far from adequate when it comes to action photography. This dungeon requires ISO 25,600 and higher. All of these photos require a significant amount of editing to produce acceptable results.
Ah, now makes total sense why the need for the heavy lifting on ISO noise. Not that this will help, but when I run heavy NR on my system via C1, the output/process time for full-sized 16-bit PSD is still in the 2-3 second range, meaning I can't see any difference with or without. Unfortunately I don't use LR or Topaz, but I also suspect NR is NR across these platforms. Perhaps if you made one of your raw files available with proposed adjustments somebody with an M4 could process it for you.
