Enjoying M43 more with Topaz Photo Ai

I'd be interested to try DXO today because last time I loaded up the trial was the end of 2022. And if there were graphics out of focus behind the subject it would sometimes try to make portions of them really sharp. So that would look pretty off.
Yes, in the last versions of DXO Lab and Pure Raw, the level of sharpness are low. The images are a bit soft. So then you can apply sharpness as you wish.
 
Last edited:
I like to at least add a tiny bit of grain back in from Capture One. This looks more like a computer aided drawing then a photograph.

Yes, the Noise REDUCTION is great. But removing all of it so it looks fake, and is too much for my eyes.
 
When I switched from years of APSC to M/4/3, I discovered the well known issues. More noise at higher ISO and often softer images with consumer grade long telephoto lenses. But, these issues have largely disappeared with Topaz Ai. Noise no longer is an issue at ISO 3200 and slight soft tele images can be greatly improved. However, I am not a pro. I do not print larger than 13 x 19 and images with fur, feathers or similar detail shot at higher ISO and consumer grade zooms are not printed larger than 8 x 10 or 11x 14 and look fine for me. Also, I realize that flaws visible at 100% monitor viewing do not show up in the way I view them.

Greg
 
I like to at least add a tiny bit of grain back in from Capture One. This looks more like a computer aided drawing then a photograph.

Yes, the Noise REDUCTION is great. But removing all of it so it looks fake, and is too much for my eyes.
Definitely a little grain still there, at least by whatever DPR does to it when I uploaded it.

Can you see it?

If some people want more grain some want less, well I guess we will have to all get little sliders on our computer monitors to dial in what you want. 🙃

8b6d60b664c94ea296ce464bc1e96dc4.jpg


--
Thanks,
Mike
https://travel-curious.com/home/north-america/united-states/utah/
 
Last edited:
I was talking aboutthis you uploaded for us.

Uncanny valley for a photograph. There is zero grain or noise.
 
Last edited:
I was talking aboutthis you uploaded for us.

Uncanny valley for a photograph. There is zero grain or noise.
I see. I assumed that when you're critiquing an image in a thread you're actually critiquing the image in the thread.

You hit on one of the nicest features of software like this, it give you the ability to control the final output based on whatever your personal preferences happen to be.
 
Last edited:
As a Linux only user I have found Neat Image to be the best solution for noise reduction.
Neat Image is good and I use it on my Win 10 PC when processing/restoring old photos.

However, when processing raw files, DXO Photolab is much better.

Allan
In the context of being available to a Linux only user.
 
I literally critiqued the image YOU UPLOADED in this thread. And I linked to it in what you replied to. Do you not know what you even post?
 
I literally critiqued the image YOU UPLOADED in this thread. And I linked to it in what you replied to. Do you not know what you even post?
Your link is to the magnified comparison image not the landscape. Is there a point to any of this?
 
Last edited:
I have been thinking a lot about this recently, not least because I am a dedicated Linux user, and have been for 25 years and the New AI based post processing has become pretty effective. My thought is that extended pixels, and larger sensors, are no longer necessary because the dynamic range, Noise at High ISO, and various other conundrums that M43 users have had to tackle over the years are made redundant with AI Post production.

...

So if this is true, then perhaps M43's is the sweet spot in regards quality and convenience, now that Post production pretty much negates any benefit larger sensors might have accrued in previous years. Not only that, but the increased DOF characteristics that M43 offers and the 2X natural crop factors might actually count very much in M43's favour, along with the portability of the equipment and lenses because of this.

..

So the question then is what is the actual sweet spot, given the revelation in AI post production?... is 16MP actually better than higher pixel sensors in M43 camera, and if so could we begin to see a reversal of the trend towards Mega Pixel OCD, and a return to more modestly priced and compact varieties, with the emphasis turning to AI post production and in camera AI optimization?

...

Perhaps even the TG-7, with it's modest 12MP sensor might actually surpass it's greater Pixelated brethren given the trend in Post production. I do recall some technical articles some time ago discussing why Olympus downgraded the TG series to 12MP from 16MP, and the conclusion was that they actually offered better resolution. Could we see the same happen to non-fixed lens cameras, now that AI post and in camera production has improved to the point at which the benefits and costings of mega pixelation are no longer cost effective either for companies or consumers alike?
 
I have been thinking a lot about this recently, not least because I am a dedicated Linux user, and have been for 25 years and the New AI based post processing has become pretty effective. My thought is that extended pixels, and larger sensors, are no longer necessary because the dynamic range, Noise at High ISO, and various other conundrums that M43 users have had to tackle over the years are made redundant with AI Post production.
...
So if this is true, then perhaps M43's is the sweet spot in regards quality and convenience, now that Post production pretty much negates any benefit larger sensors might have accrued in previous years. Not only that, but the increased DOF characteristics that M43 offers and the 2X natural crop factors might actually count very much in M43's favour, along with the portability of the equipment and lenses because of this.
..
So the question then is what is the actual sweet spot, given the revelation in AI post production?... is 16MP actually better than higher pixel sensors in M43 camera, and if so could we begin to see a reversal of the trend towards Mega Pixel OCD, and a return to more modestly priced and compact varieties, with the emphasis turning to AI post production and in camera AI optimization?
...
Perhaps even the TG-7, with it's modest 12MP sensor might actually surpass it's greater Pixelated brethren given the trend in Post production. I do recall some technical articles some time ago discussing why Olympus downgraded the TG series to 12MP from 16MP, and the conclusion was that they actually offered better resolution. Could we see the same happen to non-fixed lens cameras, now that AI post and in camera production has improved to the point at which the benefits and costings of mega pixelation are no longer cost effective either for companies or consumers alike?
Your starting point still matters and with current resolutions and performance I'd say APS-C is the sweet spot if you're going all in on AI processing just to allow for the best size/price/performance balance. Plus Fuji/Nikon/Sony have a huge selection of 3rd party lenses coming out, Nikon just needs a new sensor.

That being said m43 is still enough for most people, it's really pushing into certain niches where it has issues and even then you can sometimes make it work. And personally it looks like I'm just going to have a m43 and full frame system to be on either side of it. Though I was also using the Panasonic F1.7 zooms which goes back to the fact that your sensor only can do so much. The quality of the lens in front of it matters and those 2 have great rendering and are sharp wide open.
 
Last edited:
My thought is that extended pixels, and larger sensors, are no longer necessary because the dynamic range, Noise at High ISO, and various other conundrums that M43 users have had to tackle over the years are made redundant with AI Post production.
...
So if this is true, then perhaps M43's is the sweet spot in regards quality and convenience, now that Post production pretty much negates any benefit larger sensors might have accrued in previous years. Not only that, but the increased DOF characteristics that M43 offers and the 2X natural crop factors might actually count very much in M43's favour, along with the portability of the equipment and lenses because of this.
The way I see it M43 has benefited from the perfect storm of three circumstances meeting at a cross roads:

1) One is the increase in quality of post processing software as you mentioned.

2) The second is the rise of the internet as a viewing platform which reduces the image quality requirements needed in social media

3) The third is the dominance of the cell phone as the main viewing platform has lowered the image quality requirements even more
..
So the question then is what is the actual sweet spot, given the revelation in AI post production?... is 16MP actually better than higher pixel sensors in M43 camera, and if so could we begin to see a reversal of the trend towards Mega Pixel OCD, and a return to more modestly priced and compact varieties, with the emphasis turning to AI post production and in camera AI optimization?
One thing maybe not being recognized is all the solutions applied to getting the most from M43 are work arounds that most people would gladly forego.
  • Shooting RAW
  • Post processing in multiple software packages to get to the final result, sharpening, upscaling, noise reduction, etc..
  • Not to mention the other work arounds employed such as getting around the lower dynamic range of the m43 sensor with shooting multiple exposures and blending them in HDR software, or shooting stacked panos and having to combine in software...
It's actually a lot of time, work and financial investment to milk the most out of the M43 format.

Much of this would be reduced or removed through not reducing the sensor mega pixels, but by a nice bump from 20mp to 30mp. That would be a game changer, increase dynamic range, reduce noise, increase image quality a few steps closer to Full Frame sensors while retaining the small weight and size of M43 cameras and lenses and that would be a real revolutionary reduction in the work-arounds and amount of time spent over-coming the short-comings of the current sensor.
 
Last edited:
My thought is that extended pixels, and larger sensors, are no longer necessary because the dynamic range, Noise at High ISO, and various other conundrums that M43 users have had to tackle over the years are made redundant with AI Post production.
...
So if this is true, then perhaps M43's is the sweet spot in regards quality and convenience, now that Post production pretty much negates any benefit larger sensors might have accrued in previous years. Not only that, but the increased DOF characteristics that M43 offers and the 2X natural crop factors might actually count very much in M43's favour, along with the portability of the equipment and lenses because of this.
The way I see it M43 has benefited from the perfect storm of three circumstances meeting at a cross roads:

1) One is the increase in quality of post processing software as you mentioned.

2) The second is the rise of the internet as a viewing platform which reduces the image quality requirements needed in social media

3) The third is the dominance of the cell phone as the main viewing platform has lowered the image quality requirements even more
..
So the question then is what is the actual sweet spot, given the revelation in AI post production?... is 16MP actually better than higher pixel sensors in M43 camera, and if so could we begin to see a reversal of the trend towards Mega Pixel OCD, and a return to more modestly priced and compact varieties, with the emphasis turning to AI post production and in camera AI optimization?
One thing maybe not being recognized is all the solutions applied to getting the most from M43 are work arounds that most people would gladly forego.
  • Shooting RAW
  • Post processing in multiple software packages to get to the final result, sharpening, upscaling, noise reduction, etc..
  • Not to mention the other work arounds employed such as getting around the lower dynamic range of the m43 sensor with shooting multiple exposures and blending them in HDR software, or shooting stacked panos and having to combine in software...
It's actually a lot of time, work and financial investment to milk the most out of the M43 format.

Much of this would be reduced or removed through not reducing the sensor mega pixels, but by a nice bump from 20mp to 30mp. That would be a game changer, increase dynamic range, reduce noise, increase image quality a few steps closer to Full Frame sensors while retaining the small weight and size of M43 cameras and lenses and that would be a real revolutionary reduction in the work-arounds and amount of time spent over-coming the short-comings of the current sensor.
It's a really good system to jump into especially on the cheap with used lenses but it is a kinda hard sell right now.

You're buying a mirrorless system, do you want one where there's multiple new lenses coming out for it every month? Or m43 with maybe 1-3/year?

And it's a hard thing for m43 to change since there doesn't seem to be enough of a market for Tamron/Sigma to make more lenses for it currently and OM/Panasonic have limited resources that they're using on lenses.

Then there's the fact that m43 is a pretty mature system. You not only have the normal lenses covered there's also already unique ones like those F1.7 zooms. And a lot of the current ones are already really high quality so how much of an improvement do you get for X R&D money?

Don't get me wrong, I shoot with m43 professionally so I'm hoping it sticks around and the bodies especially keep getting better. But it's just not in the same situation as the other makes. Hopefully the OM-3 can jump start that a bit.
 
Last edited:
It's a really good system to jump into especially on the cheap with used lenses but it is a kinda hard sell right now.
Lot of good comments you made.

What do you think makes it a hard sell right now?
You're buying a mirrorless system, do you want one where there's multiple new lenses coming out for it every month? Or m43 with maybe 1-3/year?

And it's a hard thing for m43 to change since there doesn't seem to be enough of a market for Tamron/Sigma to make more lenses for it currently and OM/Panasonic have limited resources that they're using on lenses.
I can't really identify with those thoughts because I just use a system that has the lenses I want, I get the lenses I want and I'm done.

It sounds like you feel there are a lot of missing lenses from the line up? I was under the impression that m43 had a pretty robust lens line up already, but I could be wrong.
Then there's the fact that m43 is a pretty mature system. You not only have the normal lenses covered there's also already unique ones like those F1.7 zooms. And a lot of the current ones are already really high quality so how much of an improvement do you get for X R&D money?

Don't get me wrong, I shoot with m43 professionally so I'm hoping it sticks around and the bodies especially keep getting better. But it's just not in the same situation as the other makes. Hopefully the OM-3 can jump start that a bit.
M43 is kind of up against the wall because of it's such a niche market now. OMD is from what I've seen only 1.5% of the new camera market so they don't have a lot to work with, hence the trying to grab some low hanging fruit with a retro camera body in the Om-3 and recycling the Om-1 guts, a new product that was proportionately cheap to bring to market in hopes of bringing in some much needed profits.

Like I said imagine OMD grabbing just 1% of the Full Frame market sales with a new 30mp sensor. That would basically double their total sales. That's very low hanging fruit.
 
It's been a fun experience seeing the differences Photo Ai is bringing to my M43 photography.

View attachment ca13f1cbd43a4d9eaefa779727b6f6ed.jpg

fe39ad8bbebd40a2b5a70fb5f54ef3c7.jpg
The newest noise reduction tools have changed the game for small sensor cameras. I no longer worry about my ISO and normally shoot on Auto ISO. Whatever the camera selects, I just clean it up if need be. This is just one more reason why large and expensive full-frame cameras are no longer needed.

--
Daniel J. Cox
www.naturalexposures.com/corkboard
LinkedIn @danieljcox
Facebook @naturalexposures
Twitter @danieljcox
Instagram @danieljcoxne
 
It's a really good system to jump into especially on the cheap with used lenses but it is a kinda hard sell right now.
Lot of good comments you made.

What do you think makes it a hard sell right now?
You're buying a mirrorless system, do you want one where there's multiple new lenses coming out for it every month? Or m43 with maybe 1-3/year?

And it's a hard thing for m43 to change since there doesn't seem to be enough of a market for Tamron/Sigma to make more lenses for it currently and OM/Panasonic have limited resources that they're using on lenses.
I can't really identify with those thoughts because I just use a system that has the lenses I want, I get the lenses I want and I'm done.

It sounds like you feel there are a lot of missing lenses from the line up? I was under the impression that m43 had a pretty robust lens line up already, but I could be wrong.
Then there's the fact that m43 is a pretty mature system. You not only have the normal lenses covered there's also already unique ones like those F1.7 zooms. And a lot of the current ones are already really high quality so how much of an improvement do you get for X R&D money?

Don't get me wrong, I shoot with m43 professionally so I'm hoping it sticks around and the bodies especially keep getting better. But it's just not in the same situation as the other makes. Hopefully the OM-3 can jump start that a bit.
M43 is kind of up against the wall because of it's such a niche market now. OMD is from what I've seen only 1.5% of the new camera market so they don't have a lot to work with, hence the trying to grab some low hanging fruit with a retro camera body in the Om-3 and recycling the Om-1 guts, a new product that was proportionately cheap to bring to market in hopes of bringing in some much needed profits.

Like I said imagine OMD grabbing just 1% of the Full Frame market sales with a new 30mp sensor. That would basically double their total sales. That's very low hanging fruit.
I think the issue is when people are looking at their choice of cameras as a long term investment in the system. They will be with it for years or maybe even decades. So even if it has all the lenses you'd want and they still are good by today's standards what will things be like in 5 years? 10 years? m43 will have basically the same lens lineup while the other companies could be 1 or maybe 2 versions ahead. And we've already seen some pretty big jumps in optics every decade. So will m43 glass still be as good comparatively then?

I agree with you that the system today covers the needs but there's always FOMO.

I really hope that they come out with a 25-30MP body with maybe just even a half stop less noise, a full stop would be great but I'm OK with just a little jump. As of right now I'm planning to keep shooting most of my paid work with m43 in part because no other company has a body and lenses that really look like a good replacement. The closest is Sony in terms of lenses but the ergonomics of their bodies are a dealbreaker to me.

Though as I write this I realized that we are getting a bit into thread drift territory for the PC forum.
 
Last edited:
I think the issue is when people are looking at their choice of cameras as a long term investment in the system. They will be with it for years or maybe even decades. So even if it has all the lenses you'd want and they still are good by today's standards what will things be like in 5 years? 10 years? m43 will have basically the same lens lineup while the other companies could be 1 or maybe 2 versions ahead. And we've already seen some pretty big jumps in optics every decade. So will m43 glass still be as good comparatively then?

I agree with you that the system today covers the needs but there's always FOMO.
Ah, now I see what you mean.
 
I like to at least add a tiny bit of grain back in from Capture One. This looks more like a computer aided drawing then a photograph.

Yes, the Noise REDUCTION is great. But removing all of it so it looks fake, and is too much for my eyes.
Personal taste and artistic judgement in other people is always inadequate.
 
I am not a big fan of photo AI. Yes, you can achieve good results but I find that I have to tune everything down to almost zero! I don't like the over-process trend of Photo AI and I still use the old Denoise AI, which I like much more. Photo AI keeps getting better as an integrated application, but it has ways to go IMO.
so...
I too like DeNoise AI, and when Topaz first offered Photo AI, I gave it a try... Found it to be a bit too heavy-handed for my needs and abilities in it's use...
Their updates of PhotoAI vers2 have improved it well enough for me to make use of it (in conjunction with still using DeNoise AI ).
I tried to use the Latest vers 3 and had issues with the basic loading of many images - prolly an individual thing with my computer system - and being quite rooted in the current config, I'm OK with using the most recent version of PhotoAI 2.

And agreeing with @mfinley, this really does help increase the 'quality' of many images.
Given the image is already of a quality you like...
'Quality' really being 'whatever' it is you wish from an image you might create...



And PhotoAI/DeNoiseAI are certainly not the only tools available, maybe not even some significant measure of users, but they work nicely.

I'll be the 'Judge' of my own images and an 'appreciator' of the images of others, not a judge. Finding something in an image to which I can relate or appreciate, is more rewarding than 'judging'.
Learning from what others create is 'rewarding'.
PhotoAI may be very much what works for a photographer, and extends what our little M43 system can produce. Worth trying... ( I believe DeNoise has been sunset...)
Thx
Yuri
 

Keyboard shortcuts

Back
Top