Problems with Topaz A.I. Clear and Topaz Studio Preview and noise reduction bug

[samueljohnchia]

TLDR:

As requested, here is a summary of the various issues:

Issue #1
The preview in Topaz Studio does not match the image or its output, even when there are no edits applied to the image. When A.I. Clear is applied, the mismatch is even worse.

There are three related issues with the preview in Topaz Studio which causes it to be a mismatch to the output TIFF, and it is not related to an application colour management issue, nor is it a tonal or colour mismatch.
1. There is a cross-hatching pattern artifact visible at zoom percentages of 100% or more.
2. The preview of the rendering of the A.I. Clear adjustment once applied is blurry.
3. In addition, Studio preview zoom magnifications of more than 100% results in an even more blurry and inaccurate preview.

For further details, look here.

Issue #2
There is a bug in Topaz Studio which is that it applies additional noise reduction without user's knowledge or consent to all imported images regardless if any adjustment is or is not applied.

For further details, look here.

Issue #3
A.I. Clear produces *relatively* significantly worse results than Topaz Denoise. This is tangential to the claims of Topaz that it is "much better", which I feel was misleading to me. Users who are interested to produce high quality, halo-free, without adding more artifacts to their master images cannot rely on A.I. Clear yet.

For examples, look here for a tough astrophotography example, and further descriptions of the issues.

This post shows another example, a less revealing and demanding image, more typical of most user's files, but the problem persists and more importantly, is visible.

This third example is a regular landscape image which has the combination of both issues and reveals the issues with A.I. Clear very well.



-------------------------------------------------------------------------------------------------------------------------------

Hello everyone, I apologise for being late to the A.I. party. I wasn't sure if I should add on to plugsnpixels thread but since this is going to be a lengthy one with new issues, I thought it best to open another thread. I had only just been able to activate my trial of A.I. Clear after reaching out to Topaz for technical support over three months ago. After their initial reply from their Head of Support for further information which I provided within a day, I did not receive any further responses for over three months despite making multiple attempts to reach out to them. So I took to posting my issues publicly on their forums and thankfully, it got their attention. If you wish to follow the discussion there, here are the links. However, I am reproducing my report here in a more chronological manner with more accurate descriptions. I hope they will be useful to the Topaz user community or anyone considering A.I. Clear. Experienced users are encouraged to look for similar issues with their own setups before assuming that my issues are hardware-related. I am on an i5-4690K, Firepro W4100 running Windows 10. I have not noticed any hardware-related glitches running Studio or A.I. Clear, not even a single crash when attempting to process large TIFFs with A.I. Clear. I am currently on Studio V1.13.1.

Topaz Studio zoom percentages, cross-hatching grain artifact, noise reduction without any edits
Poor results with A.I. Clear

There are three related issues with the preview in Topaz Studio which causes it to be a mismatch to the output TIFF, and it is not related to an application colour management issue, nor is it a tonal or colour mismatch.
1. There is a cross-hatching pattern artifact visible at zoom percentages of 100% or more (more visible at higher magnifications). This is not present (thankfully!) in output images.
2. The preview of the rendering of the A.I. Clear adjustment once applied is blurry. Please see attachment "Studio-preview-problems-sharp-stars-100mag.jpg".
3. In addition, preview zoom magnifications of more than 100% results in an even more blurry and inaccurate preview at high viewing magnifications in Studio.

These are pernicious issues which prevent accurate judgement of the edits one chooses to apply. The "Recover Detail" slider for AI Clear has a resolution of 1% steps, and it influences the sharpness of the final result significantly. To make a precise and informed decision based on the GUI preview in Studio is nearly impossible.

The zoom percentage steps in Topaz Studio do not allow me to specify exact steps of 200%, 300%, 400% etc. unlike Topaz's older plug-ins. There is a shortcut button for 100% zoom, but then the next interval is 125%, 156.2%, 195.3%, 244.1% etc. The zoom slider's resolution allows for smaller varying steps, but the slider's resolution is not fine enough to set something like 300%, rather I can get 299.8% and that's it. It is vital that the zoom intervals are in whole steps so that the preview isn't interpolated and thus would cause difficulty with judging sharpness. Whole steps of 200%, 300% etc should only use nearest neighbour enlargement to preserve the original pixels' values for the optimal preview effect.

I suspect users may not have noticed the blurry preview issue since most images we would like to test on would be largely very noisy and blurry anyway. My stars example is quite extreme in that it is noisy but also exceedingly sharp and of high resolution, so it reveals the problem quite well. If you look at the attachment "Studio preview problems lake example.jpg" which shows 100% magnification views and a 400% enlarged crop, and compare to the screenshot from the application (attachment "studio-preview--398.6%.jpg", you will see that it is harder to spot the blurry preview bug, but discerning viewers will notice they are indeed different, and the cross-hatching pattern bug persists.

Addendum: On the Topaz forum, I asserted that the Studio's preview shows more pronounced halo-ing, but in fact it is just a result of a blurry and thus lower resolution preview sometimes causing the ugly halos from A.I. Clear's detail restoration routines to become even more grungy. This would be a more accurate description of how I am perceiving the issue now.

[samueljohnchia]

Also, there is a bug in Studio which is that it applies noise reduction to imported images even when no adjustment is applied and the original image is saved out.

The procedure to investigate this bug is as such.
1. Import a new image into Studio. Note that the "Save" button is greyed out since no adjustments are applied yet.
2. Apply any adjustment by clicking on anything in the Adjusments dropdown menu.
3. Delete that adjustment, so now nothing is being done to the image. However, the "Save" button is now no longer greyed out.
4. Save the image.

Now when comparing the saved unedited image from Studio, you will notice that noise reduction was applied even though no adjustments were made (left vs middle crop). I have tried this with several different images and they all exhibit exactly the same issue, even after uninstalling and reinstalling Topaz Studio.

The attached example includes a screenshot from Studio's application preview which adds another comparison to show how far off it is from the actual output TIFF. Also note the cross-hatching patterning again. We are looking at 300% nearest neighbour magnification of the original TIFF and TIFF output from Studio with the above steps applied (basically nothing was done but open and then saved). The noise reduction issue should be apparent to any viewer. This issue should be something that anyone can attempt to reproduce, and I would be interested to know if you see it on your end too. I am making this request as I seem to be getting some resistance from Topaz support if such a bug even exists at all, or if it is isolated to my setup only.

Note that Studio's preview does not become even more blurry when using something like the "Basic" adjustment - the tonality and colour of the preview are always an exact match to the output TIFF. You can use Studio to make tonal and colour adjustments with confidence. Only detail-related/resolution adjustments are problematic, but it is significant since we are also highly interested in the power of A.I. Clear.

Edit: I would like to add another observation. Even when applying a 'Basic' adjustment in Topaz Studio, and no noise reduction adjustments were applied, noise reduction is still occuring. Refer to attachment 'noise-reduction-bug-persists,jpg', which shows the result from a 'Basic' adjustment layer being applied with +0.2 exposure and nothing else, and still noise reduction is being applied without the user's consent to the output. I cannot ascertain if this noise reduction bug applies to an A.I. Clear adjustment since it also does a form of noise reduction when activated. It is annoying.

[samueljohnchia]

Now to share some feedback on A.I. Clear, in particular with some astro photos that I would like to process. If anything, it would seem to be the hardest noise reduction test. Small super sharp stars but low signal to noise, with flat regions of sky and smooth gradients prone to banding and revealing artifacts more glaringly.

The image in question was shot on a Sony a7R II, continuous low mode (i.e. star eater not in effect), LENR OFF, ISO 640, f/2.8, 32 seconds. The lens was a Voigtlander 65mm APO-lanthar and the setup was mounted on a tracker, hence the round and sharp stars. Just to be clear, DRO and High ISO Noise reduction was also off, but it does not matter since I was shooting in raw+manual exposure. The capture was encoded as a lossless uncompressed raw file. The image was opened in Adobe Camera Raw and then saved as a TIFF, with only a custom camera profile applied, and no further adjustments, all settings zeroed, certainly no sharpening or noise reduction applied (which I understand can wreck havok with getting optimal results from A.I. Clear). I imported this TIFF into Studio (V1.13.1) and applied A.I. Clear on it, and also ran it through Topaz Denoise 6 to compare. Please refer to the attached JPEG, labelled to indicate the processing settings for each crop. A variety of results from different combinations of settings are shown to prove my point.

I immediately noticed that regardless of the settings, I was getting a dark contour halo around the starburst, and it is the same story for any other high contrast edges (brighter stars), sometimes getting worse or better depending on choice of 'High' or 'Low' for Enhanced Sharpness setting. The best possible result from A.I. Clear was using the 'low' Reduce Noise setting, with Recover Detail set to 1.00 (maximum), and Enchance Sharpness at Low, but it is still significantly worse in that regard than Denoise. This is most likely due to the built-in sharpening optimization that A.I. Clear adds. This isn't something the user has any control over.

In addition, colour artifacts appear around high contrast edges, so the result is a blob of colour around some stars. This is perhaps the most annoying issue to me.

In Topaz Denoise, a good (not necessarily optimal) result was obtained with the Strength set to 0.05, Color noise reduction at 0.20 and Recover Detail at 0.33. Note that the result is very nearly identical to the A.I. Clear Reduce Noise Low, Recover Detail 1.00, Enhanced Sharpness Low result in terms of overall amount of noise reduction, except it is much more uniformly applied and has almost no ugly artifacts. Overall, it is a significantly better result than anything I could get from A.I. Clear.

At the moment, the pernicious detail restoration is too aggressive at even the "Low" setting for Enhance Sharpness. I think it would be best to have a way to turn it off entirely. It is destroying all the wonderful fine detail that the A.I. noise reduction process is working to hard to reveal, thus being totally counteractive to the mission.

I wish to quote from the help page of A.I. Clear (https://help.topazlabs.com/hc/en-us/articles/115004759908-Topaz-Studio-Minimum-Requirements-and-Recommended-Hardware):

Allow Dr. Partha Acharjee, a research scientist with Topaz Labs, explains it best:

"Recent advancements in Artificial Intelligence has shown great potential in image noise reduction. We were able to develop a proprietary neural network, (a type of AI that learns by studying millions of examples) architecture and realistic noise model to perform well on real-world, high-iso photos. It not only achieves much higher quality result than existing products, but does so automatically without the need for any manual tweaking. Moreover, our custom neural network, trained on millions of photos, can run on any GPU on both MAC and PC, which ensures faster processing on everyday computers without the need to upload your images to a cloud server."

At this point, I am failing to see how A.I. Clear is achieving a "much higher quality result than existing products". If anyone has any helpful suggestions on how I can achieve that, I am most willing to learn.

The JPEG is too large to attach, so here is a link to it on my Drive.

One of the admins on the Topaz forums suggested using the default setting of "0.10" for "Recover Detail" and "Med" for noise reduction, which coincidentally is also what the automatic initial processing applies. Attached is a comparison of that. It turned out as I expected - it is among one of the least pleasing results from A.I. Clear. The giant crop array example comparison already reveals that none of the Med or High settings for Remove Noise is any good in this case. The noise reduction is just too heavy-handed, especially with a very low setting for "Recover Details". Nothing I do would completely remove the ugly halos that develop around high contrast edges.

[samueljohnchia]

Here is another example where the detail is not sharp to begin with, plus is very dark and would test our eyes a bit more to discern differences. It is an old Canon 50D shot of a sunset at a lake (I'm not including the whole picture in this sample, but I can show it if you need), which was underexposed, a double-whammy due to the low dynamic range Canon sensor which was also prone to banding artifacts.

Compare the original to the A.I. Clear version, there are now visible halos on the reeds. 100% magnification, and I am comparing using the output TIFF from Studio, not screenshots of the preview window which I have established I have issues with. Processing settings are as labelled. No combination of settings for A.I. Clear thoroughly removes the halos, though they are significanly reduced when the "Recover Details" setting is at 1.00. They are just more or less visible. This example shows what the 'auto' result is.

Now compare A.I. Clear with Topaz Denoise 6. No amount of adjustment with Denoise will ever help me achieve the beautiful clean look that A.I. Clear is giving (low noise plus superficially sharp details). Math-based processing is unable to replace detail in the original image the way A.I. processing methods can, simple as that. There is no accounting for taste, but I believe the majority of people will prefer the sharper, crisper, cleaner result from A.I. Clear if it is shown as a small web jpeg. However, the sharpening halos and the somewhat crunchy-looking reeds give the impression of digital artifacts, which would only get worse with further processing applied (contrast boosts and sharpening). There are also color halo artifacts around the reeds (just as there was around the stars), even though the local contrast of the detail here is significantly lower. Turning up the 'Recover Details' slider all the way to 1.00 goes a long way to making the halos almost not visible, but the color fringing halos still remain around the higher contrast edges. It is arguable that Denoise gives a more organic or analogue effect. Also the result can be further sharpened without fear of exaggerating any halos. It is difficult to pick a clear winner (pun not intended). In this case I would probably combine with hand-painted masks both renders in Photoshop, to get the best of both worlds. But it would be nice to just have an "OFF" switch for sharpening in A.I. Clear!

So I brightened the sunset picture before running AI Clear and Denoise on it so the difference would be more visible on screen. But here it is in its unbrightened state, which is the current choice for my editing, so I can keep most of the ugly noise dark and nearly invisible in the print (attachment: AI Clear problem with dark details.jpg). Note now that AI Clear is struggling to re-create detail in the reeds, and a significant amount of ringing artifacts are present, as well as the halos like before. AI Clear is now definitely worse than Denoise. The best final result might be a blend of the two (using the smooth regions from AI Clear), but AI Clear has also reduced the saturation of the colour of the water too much.

[samueljohnchia]

One last example towards my case for now. Attached is the full view of the image in question, which is slightly diffused sunlight reflecting off some debris on a pond's surface. That region is of exceptionally high local contrast, while the foliage backdrop is of exceptionally low contrast. It is difficult to balance the degree of noise reduction without too much loss of detail. A worthy challenge for Denoise and AI Clear.

Here are the crops of the original, best result I could achieve with Denoise, and AI Clear. They are shown at 100% magnification. I urge anyone looking at these samples to not just open them in your browser window, but to right click>save image as and download them to your machine and view them at 100% (or 200%/300% especially if you are on 'retina' monitors) magnification to be able to judge these issues fairly. AI Clear is noticeably worse off in this example. There are many colour botches which were added around the light sparkles which were not there before, and the detail restoration halos again are destroying fine detail. Increasing the Recover Detail slider from the default of 0.1 to >0.5 helps to recover only marginally better detail in the low contrast foliage, but the ugly colour blobs remain just as visible. As with the previous examples. AI Clear fails to produce a comparable result to Denoise on fine detail, and generates ugly artifacts around higher contrast edges consistently.


To add some perspective -

Joe Fedric, Topaz's Head of Support, mentioned this in response to my postings on their forum:

It is possible to see artifacts in some places when you start with a high-quality image. The models were trained using low to medium quality images, rather than exemplary photos. The result is that A.I. Clear works much better with images that actually need considerable improvement.

And my response:

Thanks for explaining that the goal of A.I. Clear was not intended for producing flawless, top-notch work, and indeed such work should not ever require such heavy-handed rescue with editing software. Nonetheless, mistakes happen (like with the Hubble), and sometimes extreme methods are needed to rescue a precious picture which cannot be made again. I am happy to lower my standards in such a situation.

I would like to add that my response to the average-quality image training for A.I. Clear should be considered together with the claim by Topaz of A.I. Clear achieving a "much higher quality result than existing products". This is what is causing me annoyance here. I cannot disagree more with this claim at the moment, and it is misleading to customers who have paid for the software and are not getting the better results they expect. I do also understand that A.I. processing is still very new and it will get better over time, and hopefully also because of feedback from users such as these. Also, Bart has mentioned before that Topaz supplies version upgrades for free, forever, thus far, so paying customers have much less to be complaining about but rather just be patient for things to get better.

[Arlen]

Samuel, as a Topaz Studio user I am interested in the topic you are posting about, but it is way too much to read without having some idea what the problem is and where you are headed with it. If you could edit the beginning of the first post with a brief summary of the problem(s) before diving into all the details, it would be helpful.

[samueljohnchia]

Hi Arlen, each reply post's first sentence actually is a summary of the issue at hand. But to make it easier to digest, I have added a summary to the first post, with links directing readers to each segment of interest. Hope this helps.

The problems are significant to any professional interested in producing high quality work, but would escape any casual user. If you look at plugsnpixels A.I. Clear examples, they are full of the artifacts that I show here, and yet he claims it's a nice/good result. Sometimes it is hard to tell if something is better or not without comparisons to the renderings of other good noise reduction software (in this case, both are from Topaz), so that is what I would like to bring to the discussion here.

There's also the issue of taste. Low quality processing is getting more prevalent than ever with mobile photography, though we already went through the phase of gaudy HDR and abuse of "Clarity". If we do not help developers know what kind of beautiful, refined, subtle, natural looking results we care about, things would head in the direction of where Studio seems to be - simplified single click solutions, generic presets, swapping backgrounds and adding texture and light ray effects etc. while the core basic quality of fundamental issues get neglected.

Most importantly, the preview must match the final result and no additional background processing (noise reduction in this case) should be applied secretly or not without the user's consent.

[jim t]

You've sort of hit it on the head as to why I decided not to go ahead and buy any of Topaz A.I. products. I'm not completely satisfied with the contrast and detail filters in Studio, let alone that I still can't get a 100% preview in Studio.

[Arlen]

Thank you, Samuel, for adding the beginning synopsis; it helps. I'm aware of the inaccurate preview issue, and we have discussed it somewhat in other threads here. I will come back tomorrow when I have more time and read through the details you've presented.

[samueljohnchia]

Hi Arlen, I'm glad that was useful. I've read through as many of the A.I. Clear threads here as I could find before posting, but I don't recall anyone actually identifying the same preview mismatch issues I have mentioned. What I've read about were the presence of gross artifacts which were likely hardware issues (edit: apart from one white specks wormy artifacts report), none of them describe what I have been observing on my setup.

[Arlen]

It was in an A.I. Gigapixel thread that went off topic. Starting with post #94 in this thread:
https://forum.luminous-landscape.com/index.php?topic=126381.80

[samueljohnchia]

Awesome, thanks, I had missed that. Interesting how the preview issue doesn't affect plugsandpixels as much on a Mac? Edit: Earlier comment was hastily made from my phone. After a closer study on my main machine, I take that back. It looks just as awful on plugsandpixel's mac. It isn't platform related.

[Zen8]

Thanks for the info. I have stopped using it.     

[Arlen]

Samuel, concerning your 3 major points:

> I can confirm that "The preview in Topaz Studio does not match the image or its output, even when there are no edits applied to the image. When A.I. Clear is applied, the mismatch is even worse." At least on PC computers running Windows 10. As I pointed out to you above, this was discussed here in a previous thread. I was hoping that there might be some follow up from MauriceRR or someone else who filed a ticket with Topaz. Or reports by other Mac users, since so far I have seen only one example where 100% preview looked acceptable, and that was by PlugsNPixels using a Mac. I regard this as an important issue, one that seriously degrades the experience of working within Studio.

> I can also confirm that Topaz Studio "applies additional noise reduction without user's knowledge or consent to imported images regardless if any adjustment is or is not applied". I just tested 3 images, two of them low ISO and one at high ISO. I followed your protocol, bringing them into Studio, applying a Basic Adjustment, then deleting the Basic Adjustment and saving the image back to Photoshop. For the two low ISO images, noise that is noticeable at 300%-500% was modestly but definitely reduced by the trip through Studio, where nothing at all should have happened. For the high ISO (3200) image, I couldn't see a difference, but maybe that was because the noise was so high to begin with. Again, I think this is a significant issue; if you run an image through Studio without doing anything that should affect noise, then there should be no change in noise.

> Concerning A.I. Clear, I have used it but have not tested it as extensively as you have. My experience so far is that for some images it gives dramatically improved results with little work, compared to standard approaches like DeNoise which often (for me, anyway) require a lot more work to get it right. On other images, it doesn't work so well; either the noise removal is not sufficient, or more often, unacceptable artifacts are introduced. In either case, it is quick and easy to try, and if the results are disappointing one can move on to one of the other methods. I'm hoping they will improve it as time goes by, so that the disappointing results become fewer. I wholeheartedly agree with you that it would be very helpful to be able to turn off sharpening completely.

I started using Topaz products some years ago, and now have the entire suite. I find some of them quite useful, though for me the majority are not of interest. Topaz does have a history of coming up with innovative products and improving them over time. That said, so far I am not much of a fan of Studio. I understand why they are moving towards a stand-alone application, to remove their dependence on Photoshop for the plugins. But the plugins don't have the problems Studio does, and importantly for people like me who work in Photoshop anyway, Studio can't be used on a Smart Object layer, along with other applied filters. It has to be used on its own pixel-containing layer, significantly increasing the overall file size. Nevertheless they are moving away from plugins, so they had better fix the problems with Studio and make it attractive enough to use or it I will lose interest.

[smahn]

But the plugins don't have the problems Studio does, and importantly for people like me who work in Photoshop anyway, Studio can't be used on a Smart Object layer, along with other applied filters. It has to be used on its own pixel-containing layer

Not sure why for me but not you, but Studio works on Smart Objects for me.
FWIW, I'm on a Mac, but I doubt that matters.

[Arlen]

Not sure why for me but not you, but Studio works on Smart Objects for me.
FWIW, I'm on a Mac, but I doubt that matters.

Hey, thanks a lot for pointing out your experience. When I tried it quite some time ago, I couldn't get it to work on Smart Objects, and didn't try it after that. But I just tried again, and it works on Smart Objects for me now. Really glad to know that.

[Arlen]

Actually, it seems that Studio working with Smart Objects is a recent change. I checked my bookmarks and found this link where a Topaz representative confirmed that it did not, as of last June.
http://discuss.topazlabs.com/t/studio-support-for-photoshop-smart-objects/4253

[kirkt]

My succinct review of AIClear: it works well on images that don’t need AIClear.

Kirk

[samueljohnchia]

Samuel, concerning your 3 major points:

Hi Arlen, thank you for taking the time to look into them and share your thoughts, as well as to make your own investigations. It is much appreciated.

At least on PC computers running Windows 10. As I pointed out to you above, this was discussed here in a previous thread. I was hoping that there might be some follow up from MauriceRR or someone else who filed a ticket with Topaz. Or reports by other Mac users, since so far I have seen only one example where 100% preview looked acceptable, and that was by PlugsNPixels using a Mac. I regard this as an important issue, one that seriously degrades the experience of working within Studio.

I do have an Apple laptop on hand, so I can check it out myself, though I did see in plugsandpixel's Mac example the same degree of subtle blurring in Studio's preview. However, the jpeg screenshots he attached have very high compression, making it harder to discern these subtle yet pernicious issues.

BTW did you notice the cross-hatching pattern at >100% magnification?

I just tested 3 images, two of them low ISO and one at high ISO. I followed your protocol, bringing them into Studio, applying a Basic Adjustment, then deleting the Basic Adjustment and saving the image back to Photoshop. For the two low ISO images, noise that is noticeable at 300%-500% was modestly but definitely reduced by the trip through Studio, where nothing at all should have happened.

Thanks for confirming that!

For the high ISO (3200) image, I couldn't see a difference, but maybe that was because the noise was so high to begin with. Again, I think this is a significant issue; if you run an image through Studio without doing anything that should affect noise, then there should be no change in noise.

Try layering the Studio output TIFF over your original, then use the layer visibility to toggle them. Any subtle differences should be instantly visible.

> Concerning A.I. Clear, I have used it but have not tested it as extensively as you have. My experience so far is that for some images it gives dramatically improved results with little work, compared to standard approaches like DeNoise which often (for me, anyway) require a lot more work to get it right. On other images, it doesn't work so well; either the noise removal is not sufficient, or more often, unacceptable artifacts are introduced. In either case, it is quick and easy to try, and if the results are disappointing one can move on to one of the other methods. I'm hoping they will improve it as time goes by, so that the disappointing results become fewer. I wholeheartedly agree with you that it would be very helpful to be able to turn off sharpening completely.

In fact, I've only used A.I. Clear on several (though well-chosen) test images and was able to uncover enough issues, consistently, that I do not see it meaningful to proceed further. After being enamoured by how good A.I. Gigapixel was on both crummy and high quality master images, I was under the impression that A.I. Clear would be the same. Do note the quote from Joe which mentions that A.I. Clear was trained on low quality imagery, which implies the goal was to save nearly unusable images, not necessarily make stellar images even better. In any case, even with poorer quality originals, I haven't found A.I. Clear to be a sure win over other methods, though I do agree that it's easy - just press a few buttons. But then you have to accept or not accept whatever results you get.

I started using Topaz products some years ago, and now have the entire suite. I find some of them quite useful, though for me the majority are not of interest. Topaz does have a history of coming up with innovative products and improving them over time. That said, so far I am not much of a fan of Studio. I understand why they are moving towards a stand-alone application, to remove their dependence on Photoshop for the plugins. But the plugins don't have the problems Studio does, and importantly for people like me who work in Photoshop anyway, Studio can't be used on a Smart Object layer, along with other applied filters. It has to be used on its own pixel-containing layer, significantly increasing the overall file size. Nevertheless, they are moving away from plugins, so they had better fix the problems with Studio and make it attractive enough to use or it I will lose interest.

I am in full agreement with you, and that has been my experience and impression of Topaz the company and Studio the software. It was not helped that their Head of Support was dismissive initially about needing to wait in turn to receive a response, which was hurtful for sure since I had been in the queue for over 3 months, and right after coming public with my issues, they responded almost immediately. What gives? It was never my intention to make things hard for them, only to help them make their product better. Why we go to lengths to investigate and give detailed feedback is precisely because we care about companies like Topaz which have been making some of our most beloved specialist editing tools. I want them to succeed. If they cannot see that, or refuse to make the necessary improvements in a timely manner, we would indeed just move on (or back to) other tools.

[samueljohnchia]

Or reports by other Mac users, since so far I have seen only one example where 100% preview looked acceptable, and that was by PlugsNPixels using a Mac.

I've just downloaded Studio onto my Macbook Pro, circa 2012 with retina display to have a look myself. I immediately noticed that the Studio preview was not drawn at the same pixel density as the rest of the application for this laptop's retina display. It was pixellated and not shown at 1:1 scale factor. Thus even the 100% preview is totally unacceptable and far worse than what I'm seeing on Windows. This is quite unexcepted for software made in 2018. If anyone else on a Mac with retina display knows a workaround to this so I can do a better comparison, please let me know.

Then I set Studio's zoom magnification as close as possible to 500% and got 504.5%. I could immediately see the cross-hatching pattern, though it is ~2X finer than the screen drawn on my 110ppi 27" desktop monitor, due to the Macbook Pro's 220ppi display. Check out the attached comparison.

I could also dig up a HDMI cable and connect my Mac to an external ~100ppi monitor, but at this point, I think Topaz has more than enough evidence to work on and if they are not showing more interest, my enthusiasm shall wane too.

I think it still stands that the preview issues are not platform related.