AI cuts reflections in glasses

This is a follow-up to our last post about PortraitPro. Using a self-portrait taken for the purpose with bad reflections in uncoated reading specs, I went through the options of the reflection removal process. Mid-May a 15% discount was authorised for code CC524 at www.anthropics.com which applies to all 50% discounted program downloads.

It was taken on the Sony A7IV with 85mm f/1.8, tripod, ISO 400, lens at to f/8 and control through iPhone 15 Pro Max using Sony Creators’ App remote viewing and control. The screen on the A7IV was vertical and facing me, so I could also look at the camera and see the reflections move as I changed my head angle. Setting this up showed me some problems with the A7IV articulated screen design I had not realised – it can only face the self-portrait subject when folded out at the left-hand end of the camera, which with a Arca-Swiss L-plate means hanging down… obscured by the tripod head! So no L-plate but standard Arca small plate, and camera upside down compared to normal hand holding.

This is the result using PortraitPro V24. Read on to learn more, and don’t forget if you decide to get this program use Cameracraft’s additional 10% discount code, CCV245.

PortraitPro has come a long way in a few years. At the top end, the Studio Max version is a £308 program which costs £154 with the 50% download discount that Anthropics have offered ever since the days of CDs in packaging. Since no-one now buys a CD, the real price is £154 (with 10% off for Cameracraft’s code, CV245 in the latest May/June issue).

You may not need Studio Max with its 48-bit file capability, workflow from raw to exported finals, multiple image batch processing intended to auto retouch complete portrait sessions, handling of wedding groups and granular control fine-tuning its effects. The basic V24 includes this function and costs £99 less 50% download only less our 10% – so £44.55.

It is now very fast indeed on Apple Silicon and integrates with Adobe’s photo programs. Under the hood it uses some of Adobe’s functions, without venturing into Generative Fill AI to change a digital capture beyond the scope of many competitions. It uses AI, but does not rely on on stolen images or ones licensed for almost nothing in bulk from the big picture libraries. Anthropics built their platform on measurements of the human face and body, research into what people like or dislike, and many years of coding. When it uses image-based AI it draws that from your photo and its bank of facial features modelling data.

The Reflections in Glasses problem

Recently we came across a question in a professional photo organisation Facebook group asking how it was possible to remove reflections from glasses. It’s very difficult, and when it happens in a set of pictures where the photographer is unable to prevent it, it can ruin groups and presentation shots. Many battery studio-location flash heads now have very low power modelling and it’s all too easy to light your subject and fail to spot that your octa-box is reflecting in specs.

PortraitPro’s specimen example might just be good luck, so I decided to test Version 24. My studio room has shutters when blackout is needed. Two pure white plain blinds 110 x 220cm cover the tall south facing windows to prevent furniture, fabrics, art and photographs fading or warping in direct heat. They make a wonderful giant dual light source in daytime sun even in midwinter but reflect in glasses when the camera angle is not just right.

Removing reflections from specs does not come under the Eye menu – it’s under the “Inpainting menu” along with Mouth & Teeth and Remove Stray Hairs.

This is a crop from the original file.

The Reduce Reflections in Glasses view above shows other retouching functions too (notice some reductions in skin blemishes and wrinkles) but has the reflections reduction set to Off. When you select Remove Reflections in Glasses, you see choices for Off (the start position) then Options 1 to 5. Each is a different AI generated restructuring of what should be visible through the reduced reflection. My eyes are old enough to be slightly difficult and it was interesting to see the five choices.

Option 1

Option 2 (note the left eye eyelid in all these and how it changes).

Option 3 which I felt got the eye almost right, though further retouching would be needed for a portrait. It would be good enough for a PR or informal shot.

Option 4 rather odd mismatched detail.

Option 5 eyelid droop…

Option 3 got the upper eyelid almost perfect (not quite but acceptable) and the Strength slider did allow the reflection to be eliminated to the degree shown above. However, it looked better with 85% effect or even the 50% of the earlier example, a faint reflection remaining without obscuring the eye.

The time taken on my Mac M2 Studio Max was next to nothing, I didn’t bother to time it as everything happens in real time include the export from the starting 33MP JPEG to a same size with all PortraitPro’s very subtle modification of the portrait. The defaults were just right but I increased fine wrinkle reduction out of vanity!

After saving a copy of processed result I also saved a .ppx file (the Project) which is a bit like an Adobe .XML sidecar file, and re-opens your original with all the edits at the point you saved this snapshot, reversible and adjustable as needed.

A tougher test

Here’s a worse example than anything you should end up with, so I set maximum strength on this. Option 4 worked best, and despite my eyes being almost entirely obscured by double reflections in my computer reading specs, it was not a bad fix at all. My ‘proper’ specs are coated of course and don’t reflect as badly.

I’m sure I could ask Adobe AI to do something the Generative Fill after masking the reflection area, but in the time it would take me to brush a mask in place, the entire PortraitPro glasses reflection removal would be done and dusted. Is it worth £139 (after our code CCV245 discount)? That depends on what your time is valued at and whether you ever encounter an error in shooting which leaves reflections ruining a shot.

– David Kilpatrick

To see Anthropics PortraitPro Studio Max, and the other versions which start from £49.95 (right now there’s a 15% CC524 discount, update May 23rd 2024) – all include this reflection removal function alongside stacks of other tools – go to https://www.anthropics.com/portraitpro/

AI makes latest PortraitPro deal worth having

We’ve been testing, reviewing and occasionally using this software for years now – occasionally because portraits are not mainstream for your editor, most faces are in editorial contexts where no modification is acceptable (even when it’s just a royal jumper). But for those who must keep family or paying clients happy, in this era of completely modified selfies and altered perceptions of what a portrait should be, the latest AI version has real value.

You can get maximum download discount of 50% plus an extra 10% by using Cameracraft‘s code CCV245

If badly out of focus faces within a group can be recovered, reflections removed from glasses, and smiles improved without the £10k my dentist suggests is necessary to replace teeth you can’t even see… take a look at this full info. – David Kilpatrick

  • Key New Features
  • Mouth Inpainting & Teeth Replacer
  • Glasses Reflection Remover
  • Face Recovery
  • Skin and Hair masks
  • Improved workflow
  • New gender and age detector

This is Face Recovery, though the lass has become a bit long in the tooth – the AI teeth are more realistic than any amount of focus retrieval and sharpening plus retouching could achieve in a few minutes.

This is Mouth Inpainting.

And this is Glasses Reflection removal, which again is a task not to be relished in Photoshop, and the top end version of PortraitPro (Studio Max) can handle in groups, in a series of shots where it’s having a similar effect.

This pair shows a cumulative but very subtle effect from the improved workflow, and it clearly compensates for failings in the colour management and lighting. It’s impressive to note that PortraitPro started life being a very obvious process, and our advice was always to turn the default sliders down rather than up. It has matured considerably. The new workflow has improved gender and age detection, and Studio Max is well-tuned to Apple Silicon to make optimum use of CPU, GPU and RAM.

PortraitPro 24 Editions

PortraitPro Standard is standalone software for photographers working with JPG or 24-bit TIFF files.

PortraitPro Studio is for photographers who work directly with RAW files or want the higher quality of 48-bit colour files, supports conversion between different color spaces, and provides JPEG/TIFF embedded color profile support. Offers Batch Dialog.  

PortraitPro Studio Max For professional photographers or those working with a large number of images. Full Batch Mode to speed workflow greatly.

Compare the different editions: anthropics.com/portraitpro/editions

Availability and Pricing

PortraitPro 24 editions are available from: anthropics.com/portraitpro

Remember to use Cameracraft’s code CCV245 for maximum discount! This can also be used for PortraitPro Body, Landscape Pro and the user-recipe Smart Photo Editor.

Fake SSD drives hit Walmart in USA

Our good friend and epic photographer Frank Doorhof alerted his Facebook followers today to Walmart selling imaginary 30TB tiny SSD drives for almost nothing. This saga just continues – Amazon has certainly been duped by vendors, Facebook may or may not finally have stopped the fraudulent adverts.

Back in March/April edition of Cameracraft we lifted the lid on one of these fake SSDs. I don’t know if any other UK photo magazines have bothered to do the same. We just try to give readers information they need!

high speed flash two images combined

Elinchrom ELC500 TTL studio flash

A new generation from Swiss masters of studio flash Elinca SA brings multi-platform TTL, super-fast recycling and flash durations, brilliant LED modelling and many design innovations. David Kilpatrick has been trying out the twin head kit.

The second wave of any innovation in technology is often safer to invest in than the pioneering first generation. Studio flash offering IGBT duration and power control, allowing much the same TTL and high speed functions found in camera speedlights, has been in development for over a decade but whole generations have been orphaned by advances in wireless trigger and camera firmware.

Finally bringing this to their new mid-range ELC TTL heads – one rung below the ELC Pro and one above the BRX – Elinchrom has worked for maturity in the whole technology. So, when the ELC 125 and 500 TTL arrived they worked much like any head with the EL Skyport Pro. Days later new firmware for the triggers enabled TTL operation, across a range of camera platforms already proven with the portable ELB 500 TTL.

The ELC 125 TTL is a little larger than a D-Lite. The ELC 500 TTL is substantial – as expected.
Continue reading »

Nikon's D600 – FX goes Prosumer

D600 with 24-85

Nikon announced the D600 at 5am today, confirming rumours which were beaten only by Apple’s iPhone 5 leaks for accuracy.

The 24Mp entrant seems to be part of ‘full-frame fever’ undoubtedly driven by Sony’s CMOS sensor development, pricing and more crucially, packaging the definitive 35mm format to appeal to mainstream consumers.

Despite a D3X matching resolution, the D600 is a very different sensor and package. Will this be the camera to push Nikon’s DSLR market share to over 50%?

The current DLSR line up at Nikon is quite striking, not only for capability but also the positioning, with a substantial gap between the highly-specified DX-crop D7000 and the 36Mp professional D800 bodies. The middle ground retains the D300s, almost identical in price to the D7000 but qualifying for Nikon Pro User status and now one of Nikon’s oldest DSLR bodies. The D600 fits at the upper end of that gap, with an SRP of £1955.99 in the UK for the body.

For that price, you get a tightly controlled feature set, a compact, lightweight body and sensor capabilities that exceed the state of the art just 2 years ago, when the D3X was in demand, in short supply, and retailing at over twice the D600’s figure. A quick launch-venue play suggests that the specified ISO range – peaking at 6400, rather than the D3X’s 1600 – is very usable. The body weighs only 760g, using a magnesium upper and rear body and offering similar weathersealing to the D800.

Advances in processing, video and OS make themselves felt instantly. FX and DX crop HD video recording with HDMI output for uncompressed streams and sophisticated audio monitoring, a base ISO range from 100 to 6400 extendable to 50 to 25,600, and in-body raw editing are all very compelling features regardless of resolution. The D600 manages 5.5-6fps in full-frame mode, and shoots to two UHS-1 SDHC cards.

The 100% viewfinder is bright and despite using the square, without blanking filter, window rather than the round type used on previous FX bodies seems very similar to the D800. The eyepoint may be a further slight reduction, but without detailed specifications that’s a hard one to call.

A true pentaprism is used – expected, perhaps, in a full-frame high-end body but fighting an increasing trend for electronic viewfinders.

A compact body presents a few ergonomic challenges, and Nikon have tackled the control interface with the experience you’d hope for after the clear new direction shown in the D4. Gentle slopes define the shutter release area, with joystick, function buttons and the standard buttons beside the 3.2″ screen (which features a clip-on protector). A mode wheel/drive wheel combination provides consumer-style selection of scene modes, with a drive wheel below including selection of the IR remote mode, which is supported by receivers on both the front and rear of the body as per the D7000.


Nevertheless the D600 is a consumer package. It’s a high-end one, but it carries a 1/4000th shutter, horizontal axis level only, consumer interface sockets (the compact remote/GPS port rather than the screw-in port of the pro bodies, and no PC-sync socket). Unlike the D800, the D600 has USB 2. At launch, it seemed that the WT4 wireless tethering solution was not supported, but some of the launch material suggests that it is supported, alongside the low cost WU-1b introduced specifically fort he D600.

The Android remote control application for the WU-1b (below) is already available; an iOS version will follow before the end of September 2012. It offers rather less control than Camera Control Pro, but does provide a live-view relay and release function.

That WU designation has been seen before, on the similar accessory for the determinedly consumer (and best-selling) D3200. It’s a wireless broadcast unit slightly more sophisticated than using an Eye-Fi card, and at £64 is almost a tenth of the SRP of the WT4. It sacrifices many of the camera control functions (though triggering is possible), and is mainly intended to transmit and share images via Android or iOS devices. It’s a shame that this split exists in Nikon’s line, as the WT4’s full-fat networking and storage solution is a lot for many studio photographers who would probably find the basic transfer/triggering of the WU-style units very useful on the pro bodies.

Mapping the planes

Samsung has a patent and a plan for using two lenses with triangulation (image offset) depth detection between two images in what is roughly a stereo pair. Here’s a link:

http://www.photographybay.com/2011/07/19/samsung-working-on-dslr-like-bokeh-for-compact-cameras/

Pentax also have a system on the new Q range which takes more than one exposure, changes the focus point between them, and uses this to evaluate the focus map and create bokeh-like effects. Or so the pre-launch claims for this system indicate, though the process is not described. It’s almost certain to be a rapid multishot method, and it could equally well involve blending a sharp image with a defocused one.

In theory, the sweep panorama function of Sony and some other cameras can be used to do exactly the same thing – instead of creating a 3D 16:9 shot it could create a depth mapped focus effect in a single shot. 3D is possible with sweep pans by simply taking two frames from the multi-shot pan separated by a certain amount, so the lens positions for the frames are separated enough to be stereographic. 3D ‘moving’ pans (scrolling on the TV screen) can be compared to delaying the playback of the left eye view and shifting the position of subject detail to match the right. But like 16:9 pans, they are just two JPEGs.

All these methods including the Samsung concept can do something else which is not yet common – they can alter any other parameter, not just focus blur. They could for example change the colour balance or saturation so that the focused subject stands out against a monochrome scene, or so the background to a shot is made darker or lighter than the focused plane, or warmer in tone or cooler – etc. Blur is just a filter, in digital image terms. Think of all the filters available from watercolour or scraperboard effects to noise reduction, sharpening, blurring, tone mapping, masking – digital camera makers have already shown that the processors in their tiny cameras can handle such things pretty well.

Once a depth map exists there’s almost no limit to the manipulation possible. Samsung only scratches the surface by proposing this is used for the esoteric and popular bokeh enhancement (a peculiarly Japanese obsession which ended up going viral and infecting the entire world of images). I can easily image a distance-mapped filter turning your background scene into a Monet or a van Gogh, while applying a portrait skin smoothing process to your subjects.

Any camera with two lenses in stereo configuration should also, in theory, be able to focus using a completely different method to existing off-sensor AF – using the two lenses exactly like a rangefinder with two windows. So far this has not been implemented.

Way back – 40 years ago – I devised a rangefinder optical design under which you can see nothing at all at the focus point unless the lens was correctly focused. It works well enough for a single spot, the image detail being the usual double coincident effect when widely out of focus, but blacking out when nearly in focus and suddenly becoming visible only when focus is perfect. I had the idea of making a chequerboard pattern covering an entire image, so that the viewfinder would reveal the focused subject and blank out the rest of the scene, but a little work with a pencil and paper quickly shows why it wouldn’t work like that. The subject plane would have integrity, other planes would not all black out, they’d create an interestingly chaotic mess with phase-related black holes.

Samsung’s concept, in contrast, could isolate the subject entirely – almost as effectively as green screen techniques. It would be able to map the outline of a foreground subject like a newsreader by distance, instead of relying on the colour matte effect of green or blue screen technology. This would free film makers and TV studios from the restraints of chroma-keyed matting (not that you really want the newsreader wearing a green tie).

The sensitivity of the masking could be controlled by detecting the degree of matched image detail offset and its direction (the basic principle of stereographic 3D) – or perhaps more easily by detecting exactly coincident detail, in the focused plane. Photoshop’s snap-to for layers works by detecting a match and so do the stitching functions used for sweep and multi shot in-camera panorama assembly. Snap-to alignment of image data is a very mature function.

Just when you think digital photography has rung all the bells and blown all the whistles, the tones of an approaching caliope can be heard rolling down the river…

– David Kilpatrick