How to eliminate noise with PureRAW 3

DXO has just released PureRAW v3 to enhance your RAW images automatically with artificial intelligence (AI). I previously posted a detailed tutorial with PureRAW v2 showing how it can be used to enhance detail and reduce noise in high ISO images. Be sure to see that previous review, as I cover several details I’m not going to rehash here.
PureRAW is one of those rare tools that can improve most images with minimal work. It helps reduce noise, enhance detail, and correct lens distortions. That might sound like a tool only for high ISO images, but as I show in the video below, it can significantly improve the quality of even ISO 64 images for print. It’s incredibly versatile.

With the v3, DXO PureRAW offers the following improvements:
  • A new DeepPrime XD method which claims to help reduce the equivalent of 2.5 stops reduction in ISO.
    • In my experience, this holds up to their claims. The noise reduction is vastly improved and this has significant benefits to not only high ISO images, but the ability to print any image. When combined with Topaz Gigapixel, you can enlarge your prints to sizes I would never have thought possible.
  • Improved fine detail.
    • Noisy shadow detail improves significantly over the already impressive results from v2.
    • I find it adds substantial fine detail on ISO 64 images. It feels like my lenses are much sharper.
  • More control over settings.
    • You can now choose from 5 levels of lens sharpening instead of 2 (I’m counting off as an option).
    • You can choose from 3 methods of lens correction instead of 1. You can keep the original aspect ratio, keep the maximum area without any blank pixels, or keep everything to let you content-aware fill missing pixels to create the largest possible corrected image.
    • And you can selectively remove vignetting and chromatic aberration.
  • New batching processing options. If you use the standalone version, you can queue up several jobs at the same time, including with different settings. However, this capability does not appear to be in the v3 LR plugin
  • Support for Fujifilm X Series cameras.
A few things to note:
  • I find that nearly every image is improved. The one exception is starry night skies, where the results are a mix of good and bad. While I find the foregrounds of those images are improved, there are artifacts in the stars and I prefer the older DeepPRIME method (which comes with v3 and you can easily use or blend it as you like). I’m more inclined to use other noise reduction techniques for the stars (such as stacking or other denoising algorithms) than v3. But they keep making huge improvements and hopefully this is an area which will benefit as well in the future.
  • PureRAW’s legacy modes (such as the original DeepPrime) are still available and work the same as they did in v2.
  • You can also integrate this tool with Lightroom’s merge to panorama / HDR tool (you just need to run DXO before the LR merge).
  • I believe this is an excellent tool for HDR output from a single RAW, as it helps you avoid shadow noise while exposing properly for the highlights in a scene.
  • Try the demo version or check the camera compatibility list if you use an unusual camera or want to use a Smart Phone. I cannot process either standard RAW or ProRAW from an iPhone (and do not see any recent iPhone in the official list, but some very old models are supported).
  • If you already have DXO PhotoLab 6 ELITE, you don’t need PureRAW3 unless you’re seeking the Lightroom integration. You also have the same DeepPRIME XD capabilities.

 

The Lightroom integration offers a simple workflow. Just go to File / Plugin Extras / Process with DxO PureRAW 3 in LR after you’ve installed their plugin.

  • Set the denoising technology to DeepPRIME XD. This is mandatory if you want the best noise reduction. If you see artifacts (such as in stars), you might consider falling back to the older DeepPRIME method. But in general, I’d just use XD.
  • The lens sharpness works great at “standard“. This helps enhance fine detail is helpful when you intend to print. If you see any unwanted detail or artifacts, just uncheck the option to turn it off (it will still apply some detail enhancement even when completely off).
  • Turn vignetting on or off as you like. I prefer to leave vignetting off and use the controls in LR / ACR instead since they work similarly well and you can control the amount applied.
  • Turn chromatic aberration ON. The results here are better than you’ll get with LR / ACR after the fact.
  • I generally leave lens distortion off. I get good results from LR / ACR. More importantly, you’ll lose the option to blend the output with your original RAW if you want to mix different methods (which I would certainly do for night sky shots where I would treat the stars differently from the foreground).
  • I strongly recommend updating the name every time you change settings. PureRAW doesn’t name the file to remind you which settings you use and it can get confusing to compare different options if you don’t do this yourself. Hopefully, they’ll do this for us in the future.

 

Who should consider PureRAW 3?

  • Anyone making large prints. The noise reduction and fine detail are very powerful and a worthwhile upgrade from v2.
  • Anyone looking to reduce high ISO noise (other than starry skies as noted above). This is a great tool for hand-held portraits, weddings, events, and sports (if you can afford 15-30 seconds per images to process the RAW).
  • Those working with RAW files from small sensors such as drones or cropped sensors, which are inherently noisy even at low ISO.
If you’re curious to learn more about what they’re doing, check out DXO’s article on linear DNG and their official list of supported cameras and lenses. PureRAW costs $129 new, or $79 if you’re upgrading.

[Disclosure:  This post contains affiliate links.  I have purchased all the software referenced above and only endorse tools I personally use and recommend. If you purchase through these links, a small percentage of the sale will be used to help fund the content on this site, but the price you pay remains the same.  Please see my ethics statement if you have any questions.]

What’s new in Web Sharp Pro v5?

I’m happy to announce the launch of Web Sharp Pro v5, which is another free update for all existing customers.

In the video below, you’ll see a few highlights including the new ability to create powerful new templates. The new templates make it incredibly easy to create custom exports for all the various ways you need to export your images. For example: you could create presets to export to Facebook with blur border, a 4×5 Instagram portrait template with some added grain, and a larger watermarked version of the same image for your website. The options are endless. Whatever your needs, you can easily switch between the settings presets you create with just a few clicks. And coming in Web Sharp Pro v5.2 (expected in April), you’ll be able to use these new presets during bulk exports to create multiple versions of entire collections of images.

In addition to the new settings presets, v5 adds support to work with multiple watermarks, the ability to interactively update your overlay templates, add secondary borders to overlays, and more. See the release notes for full details on the changes.

Note: I made some quick improvements on v5.0.0, which used a different approach to the new interface. So you’ll need to update to v5.1.0 or later to use the new “save / load settings button” I show in the video. 

 

 

I upgraded from M1 to an M2 MacBook Pro. Was it worth it?

I’ve been using a 14″ 2021 M1 MacBook Pro (MBP) since it launched. It’s been an amazing computer and continues to amaze me. To push things further, I just upgraded to the new 2023 M2 version. Was it worth it?

Both my old and new laptop are the most fully-loaded version you can get in the 14″ screen size. That includes the “Max” version of the CPU, 8TB internal storage, and maximum RAM options

The new M2 Max includes faster CPU cores with 2 more efficiency cores to improve battery life, faster / more GPU cores (38 vs 32), and the RAM has increased from 64 to a whopping 96GB RAM (both models at 400GB/s).

There are some improvements to core features as well including Bluetooth 5.3 (vs 5.0 for better performance, stability, and security with supporting devices), Wi-Fi 6E (vs Wi-Fi 6 for improved speed / latency with a supporting router), and HDMI 2.1 (vs 2.0 for double the bandwidth to support higher refresh rates on 4k monitors).

 

Test results: M2 Max vs M1 Max

First, I need to note that I’m running benchmarks not long after migrating my data, which means that my M2 machine is still doing some background indexing for Spotlight which may mean that I’ll see greater gains when that process is done. Second, be sure to check out my article on how to test performance of your CPU, GPU, RAM, etc for more details on how I made the observations below.

Here are the results I saw when comparing my newly M2 Max to my older M1 Max on the same tasks:

  • Photoshop tasks take roughly 12% less time in G-Bench tests (scoring 50.2 vs 57.2 for typical runs, though I’ve seen the M2 get as low as 49.6 and the M1 as low as 55.9). Based on Apple claims and 3rd-party benchmarks showing ~20% performance improvements, I expected I’d see G-Bench scores in the range of 45-50. So I would say these results are in line with expectations, but not quite what I’d hoped to see. Note: the detailed M2 data will be included in G-Bench standalone as well as Lumenzia and Web Sharp Pro utilities within the next couple weeks.
  • Creating Smart Objects is one area where the M2 really shines. Converting a complex set of layers to a Smart Object takes ~55% less time. I’m guessing improved SSD write speeds play significant a role in that. However, opening a Smart Object for further editing only saves ~7%.
  • The RAM increase to 96GB offers real benefit for heavy workloads. I’ve often seen the Activity Monitor show “memory pressure” in the yellow or even red with the M1. I expect I will almost never see that with the M2. That means less use of memory compression or swap files on the disk. That helps avoid performance degradation with the M2 and helps further separate its performance from the M1. When I tried enlarging a massive image, I saw the time savings of the M2 jump to 30% vs a 20% savings on more modest files.
  • The internal drive is dramatically improved for saving data. Write speeds are 50% faster (6800 vs 4500 MB/s in Black Magic). That won’t speed up saving images in Photoshop if you’re using compressed formats, as the CPU is the bottleneck there. It also won’t accelerate copying data as APFS uses virtual copies internally and external sources would be slower. But if you’re saving uncompressed images, you’ll definitely see a benefit. And this probably accounts for a good part of the massive speed boost in creating Smart Objects.
  • SSD read speeds are the same (about 5500 MB/s on each). Yet, large files open in 20% less time on the M2 (which seems to be driven by CPU improvements decode the data more quickly).
  • Apple claims battery life should be improved (18 vs 17 hours of video playback on the 14″ laptop). I don’t have a good way to confirm this (especially given the aged status of my M1 battery, which now shows a maximum capacity of 89% after 16 months of heavy use). The M1 has only 2 efficiency cores and they are frequently fully utilized, while the M2 offers 4 efficiency cores and should be able to run more efficiently for email, web browsing, and other simple tasks. I wouldn’t expect much improvement using performance-driven apps like LR and PS though.
  • The doubling of efficiency cores may help speed up tasks like background indexing which are written to prefer efficiency cores. When I allow Spotlight to index, I notice both machines max out the efficiency cores, but the M1 doesn’t seem to use the performance cores in a significant way for the task (even when on wall power).
  • Gigapixel runs about 7% faster. Previous hardware updates have shown huge gains as the software uses multiple cores and GPU well, it may not yet be fully optimized for M2 and might get a further speed bump with some future update.
  • I’m generally seeing other non-photography tasks I use (video, transpiling software, etc) save 10-20% as well.
  • For reasons I don’t understand, I only saw a 5% gain using Handbrake to transcode video. I wonder if there is some thermal throttling involved, as it was using nearly 100% of every core for an extended period of time – whereas Final Cut Pro X doesn’t push the CPU cores nearly as hard. I notice the fans get to to loud speeds noticeably earlier on the M2 than the M1 when using all CPU cores, perhaps due to the revised cooling design. (Note that fan noise is still rare unless you’re doing something using all cores heavily, such as importing a batch of images to LR or exporting video).

On the whole, these are fairly incremental gains and less than I’d expected.

So, was it worth it?

For me, yes (just barely). This is the single-most important tool I need to run my business and I’m willing to invest in it. That’s based largely on the desire for increased RAM, improved productivity, and my business needs for video and software development – balanced against tax considerations as a business expense and the net cost after selling the old laptop.

Unless you bought and lower-end M1 and now want to get more of the upgrades you skipped the first time, you should probably stick with the M1 you have. The performance gains are incremental and the 96 RAM option will only benefit a small group of photographers for the next few years. The future “M3” update is likely to be a more significant update and I’d wait for that. None of that is a negative on the M2 at all, it’s just that most people won’t get significant value from updating any computer in less than three years.

However, the value of Apple Silicon just got more compelling if you didn’t already have the M1. If you have an older Mac or a Windows machine, the M2 (or a now cheaper M1) are excellent computers and I would highly recommend either for nearly any photographer. The 1600-nits HDR-capable XDR display on this laptop would be worth the upgrade alone in my opinion (as well as being a very compelling reason to consider switching from Windows to Mac for photography, which I haven’t advocated for a long time prior to these new HDR screens). In addition, you get an incredibly powerful laptop with excellent battery life, sound system, and overall quality.

Most photographers should consider the Pro version of the M2 MacBook Pro. The Max CPU/GPU upgrade is more power than you need if you aren’t doing video. I’d get the 14″ version if you want a lightweight device for travel (and have an external monitor at home). 32GB is ideal for most photographers. I’d get a 1TB drive and pick up a 4TB Sandisk Extreme as an external drive if you’re on a budget.

Read my original M1 Max review for more details on why I strongly recommend any of these new Apple Silicon laptops.

 

Disclosure: This article contains affiliate links. See my ethics statement for more information. When you purchase through such links, you pay the same price and help support the content on this site.

Make city lights glow with HDR in ACR

HDR (High Dynamic Range) is the biggest leap in image quality I’ve ever seen. It lets you truly show the full dynamic range of your camera without compromise. Adobe Camera RAW (ACR) v15.1+ makes it incredibly easy to process your RAW images as HDR on an HDR-capable monitor.

In this tutorial, you’ll learn how to setup and use HDR to get the most out of your images in ACR.

See my HDR page and e-book for much more detail on how to setup HDR. You’ll need to enable the tech previews shown in the video above and restart Photoshop to get support. HDR is supported in ACR for all computers at this time, but layer-based editing outside ACR in Photoshop is only supported on MacOS at this time. See my related videos on HDR in ACR and how to export HDR images for the web using the AVIF file format.

Be sure to see my tests and monitor recommendations to make sure your display supports HDR. If you’re viewing the video on an SDR (Standard Dynamic Range) monitor, it will be “tone mapped” to give you a relative sense of what’s possible. In other words, you’ll get a simulation of the effect on an SDR monitor, but need to view it on a true HDR display to really appreciate the benefit. Try viewing using Safari or Chrome on an M1 or M2 MacBook Pro to see just how truly stunning it can be. Also, I would avoid FireFox for this video if it looks strange (you may incorrectly show the wrong version of an HDR YouTube video on that browser).

Please also note that I’m still learning how to optimally render HDR video, and what you see in this tutorial isn’t quite as good as the real experience. I highly encourage you to try it for yourself.

How to soft proof easily in Photoshop

Whether you edit your photos in a huge colorspace like ProPhoto RGB or even a small one like sRGB, you’re eventually going to run into issues with “out of gamut” colors when your image is printed or viewed on another computer. For example, a vibrant Adobe RGB green may be a little more dull on a monitor that supports P3. Or a vibrant sRGB magenta may not be printable. Knowing how your images will look like on the web or as a print is extremely valuable. That’s exactly what “soft proofing” is, a preview of how your image will look will look on another monitor or as a print.

Soft proofing helps you:

  • Optimize your images to look their best when printed or shared on the web.
  • Consider various print options, such as the ideal paper to preserve the colors or shadow detail of your image.
  • Be more aware of how your image will appear. Whether you try to make changes or not, the  final result will be different if some pixels are out of gamut (by definition, out of gamut colors cannot be reproduced on the monitor/paper in question).

Other than standard color spaces like AdobeRGB, you will need to install ICC profiles for anything you wish to proof (such as your monitor or specific papers/printers). You can create your own profiles with tools like the XRite i1Studio or download them from your the printer manufacturer or your commercial print lab.

Lumenzia (v11.3+) and Web Sharp Pro (v4.0+) include options for soft proofing. They are designed to make the soft proofing process faster and easier, as well as tools for correction in Lumenzia. These panels take different approaches, as Web Sharp Pro is only intended to output images for the web.

 

How to soft proof in Web Sharp Pro:

  • Go to Settings / General and set the “proof when hovering over sharpen” to either soft proof or gamut warning. You simply choose one or the other, and “soft proof” is going to be the best option most of the time.
  • The “colorspace” you choose in Settings / General is what will be used for proofing. So you can proof as sRGB, P3 (on MacOS), or Adobe RGB. You cannot proof other colorspaces because they are irrelevant (since these are the only options for exporting the image).
  • Just hover over the “Sharpen” button in the main panel, and the soft proof or gamut warning will be shown (unless your colorspace is set to “do not convert”, as there is nothing to soft proof when the output is unchanged).
  • If you wish to change the color used for the out of gamut warning, go to PS prefs / Transparency & Gamut and click the color swatch.

If you wish to confirm the profile when using the out of gamut option, you can go to the flyout menu (top right of my panel) / tool tips & info and turn on “Document & PS info”. This will show proof/gamut info as extra text below the normal panel area (be sure to click and drag down the bottom edge of the panel to see it).

Note that Web Sharp Pro does not include corrective tools, but you may use Lumenzia or add a top group layer named “soft proof” and Web Sharp Pro will toggle the group on and off with your proofing (see the Lumenzia demo in the video above to learn more).

 

How to soft proof in Lumenzia:

  • Click “Gam” to see a popup window with options. Note that the “Gam” button is disabled if there is no image open or your active image is in 32-bit mode (soft proofing is not accurate for HDR images, you should convert to 16-bits and then soft proof).
  • Choose your desired profile from the dropdown. You’ll notice that the list is much shorter than what you see in the Photoshop setup dialog, as the majority of the options are not useful and Lumenzia is automatically reducing the clutter to help you find relevant choices. Both RGB and CMYK profiles are supported. You may click the filter checkboxes to help shorten the list of options further.
  • In the hover options section, check the soft proof or out of gamut warning options as desired. You may select both.
  • When soft proofing is enabled, you may have the option to choose the perceptual rendering intent. If the option is greyed out, the currently selected profile does not support it. Most of the time, “relative colorimetric” is best. However, perceptual may be optimal for images with color gradients which are out of gamut (such as the soft color transitions of a rose petal).
  • When the gamut warning is enabled, you may also click on the color swatch to set the color used for the warning. Middle grey is often the best choice, as it clearly stands out in the middle of out of gamut colors.
  • Click “Done” once you’ve selected the options you wish to use and then just hover over the “Gam” button to see the soft proof and/or gamut warning.

Just like Web Sharp Pro, you can go to the flyout menu (top right of my panel) / tool tips & info and turn on “Document & PS info” to see red text describing any soft proof you use through the panel.

The “Gam” button also includes an option to “add soft proof corrections group“. This creates an HSL adjustment layer with a layer mask biased towards out of gamut pixels for the selected profile. This group is automatically made visible when you hover for a soft proof, so that you may directly compare your original image (without proofing) to your corrected image (with proofing). The mask utilizes density to help ensure smooth results (see this tutorial to learn more about why density is so useful and how to adjust it). The easiest way to use the proof corrections group is to hover over “Gam” and hold <shift> as you move away (to make activate the soft proof and the corrections group), adjust the HSL layer as desired or add other corrections to the group, and then hover over Gam again to compare your corrections to the original.

 

How to soft proof with Photoshop:

Soft proofing in Photoshop is setup through View / Proof Setup / Custom:

  • You can then select the color profile you wish to use for proofing (this is the “device to simulate”).
  • Rendering intent is often best as “relative colorimetric”, but some images benefit from “perceptual”. Note that while Photoshop always gives you a choice, many profiles do not support it and you’d get the same results as relative colorimetric (if the ICC profile file isn’t 1MB or larger, that’s usually a clue that it does not include the lookup tables for perceptual intent rendering).
  • I recommend leaving “black point compensation” checked and leave the other options alone unless you understand them fully.
  • Then check View / Proof Colors to soft proof or View / Gamut Warning to see which pixels are out of gamut.
  • You may change the gamut warning color via PS Prefs / Transparency & Gamut / Gamut Warning Color. Middle grey is often the best choice, as it clearly stands out in the middle of out of gamut colors.

Whenever soft proofing is active, the proof profile name will show in the document tab (ie the file name at the top of your image). This only applies for soft proofing, not gamut warnings. You can confirm the status of the gamut warnings by reviewing the View menu.

 

How to compare different profiles with MacOS ColorSync Utility:

If you use MacOS, you can use its “ColorSync Utility” to view and compare ICC profiles. This is a great way to compare gamuts for different profiles. A few quick tips:

  • Select the profile to review in the “profiles” tab.
  • Click and drag to view the gamut from different angles.
  • <option>-click and drag up to zoom out (or down to zoom in). This is helpful to see very large spaces like ProPhoto.
  • Right-click and choose Yxy to view the color space with the traditional horseshoe plot of spectral colors (these are the most saturated colors we can see).
  • Right-click and choose “hold for comparison” and then click on another color space to compare the two gamuts directly. You should hold the larger space (which will show as a colorless wireframe) to make the comparison easy to view.
Greg Benz Photography