Today’s post is a little different: following on from the excellent reception given to my interview with Nick Brandt, and my current focus on pushing print limits, it is high time we heard from the print master himself – Wesley Wong. I can say plenty about the process, but there are a lot of areas in which is expertise greatly outstrips my own. I also strongly believe that he is an integral part of the artistic process of bringing an image to its final form, and that my print buyers – thank you – would also enjoy meeting the man, albeit virtually. It’s a lengthy discourse as there’s a lot of ground to cover, so the interview will be split into two parts. Read on…
1. How and why did you start printing? Was it motivated by photography first, or something else?
I started photography almost 10 years ago. I waited patiently for digital camera technology to mature before beginning. To me, photography has always been about an image captured in time, to be viewed in print. Back in 2004, there were no Facebook, Flickr, Instagram, etc…except maybe PBase. I used to send my images (3MP from a Fuji FinePix S5000) to the photo lab to make 4R, 6R and 8R prints. I noticed that there was no consistency of colors on the prints from the same lab. At that point I knew nothing about color management, and was entirely dependent on the inferior LCD screen on my laptop computer. Soon after, I bought an A4 desktop photo printer and started making my own; there was no calibration or profiling but did enjoy the output.
In 2005, I encountered Ilford inkjet papers in a local photography magazine; after contacting them, the distributor invited me to his shop, where I was introduced to large format printing and color management. After a month or two, I bought my first large format – A2 – printer. At that point, I started spending more and more time researching fine art printmaking, color management and workflow. The first output from my calibrated workflow was indeed a wonderful experience! Soon after, I switched to other media – especially those supplied with manufacturer ICC profiles. In 2005, I started offering commercial print services for photobooks and canvas part time – I still had a day job in IT. 2007 saw me partner up with HP; their technology excited me because it was the only printer at the time that had an integrated i1 spectrophotometer to generate automated color profiles. I am still a HP official partner.
2. Were you happy with the initial results at the time? If not, why not?
There was definitely an initial steep learning curve – I had to learn post processing in Adobe Photoshop, handling the printer, managing colors and understanding how it all worked together. There was a lot of trial and error, and even the manufacturer’s ICC profiles do not guarantee accurate color. I cannot say I was not entirely happy with the initial output; though at least there was some consistency when repeat printing. The inkjet prints were definitely better than the minilab (dye sub) prints, with greater color accuracy and wider gamut. Black and white prints had less of a color cast.
3. Given the increasing skew towards digital-only output, how important is printing in the whole photographic process?
Making a print is the final part of the photography workflow. Without a print, there is no reference on how an image is to be viewed and appreciated. There is no guarantee that what you see is what anybody else sees. Viewing an image on screen does not carry the same visual impact and experience as colors differ from one screen to another, and the image is not viewed at the final intended size with the correct aspect ratio. Although a screen uses transmitted light, the visual experience is so different from even a backlit print – let alone an incident-lit one.
Without printing, the megapixel race does not make sense – it is unnecessary if the final output is web display. Why are we chasing higher megapixels if the intension is not making prints? I am not implying that we need to make a print for every image we capture – on the contrary – since most photographers don’t print, they also do not adhere to a workflow that renders a consistent image or one that is printable at decent sizes. Pixel quantity does not necessary equate to pixel quality if the workflow is weak.
4. A lot of people are used to small minilab prints; hobbyists will print larger, but still on somebody else’s machine; serious amateurs will probably try making their own inkjet prints at some point during their photographic careers. At what stage in this process do people tend to come to you?
While most people are happy with the small prints from minilabs, photographers after higher quality generally prefer to outsource the printmaking especially when a large print is required. Most minilabs in Malaysia do not offer anything above 8×10”, and even then they are confined to chemical prints on glossy photo papers. A typical desktop printer will make an A4 or at best an A3+ print; they also lack the gamut to make a color-accurate print. Most of my customers look for larger size prints at higher quality with long permanence on a variety of papers including fine art, baryta or canvas.
5. How many of them are surprised when you tell them their file quality is the limiting factor in the output? Have you ever had anything unprintable?
Well, frankly I get that almost all the time. Again: pixel quantity or sensor size does not necessary make a good image if the workflow is weak. Workflow begins with Capture – which involves every aspect of technical execution, especially choice of aperture, lens resolving power and diffraction limits, sensor limitations, critical focus, vibration reduction, etc. Assuming all is well, the next step is Post Processing. The four major factors which may severely degrade pixel quality are RAW Conversion, Noise Reduction, Sharpening and Interpolation. Without entering into the details, it is important to select the right RAW converter for the camera and apply the right techniques to prepare a good file for print.
It is my responsibility to advise my customers on the ‘printability’ of their files at the intended sizes before making the print. It may sometimes mean turning down a print job if the result is not going to be good.
6. Have there been any big surprises either way with file quality? I.e. ‘big names’ whose files weren’t up to expectations, or unknowns who’ve produced the most incredible pixel quality you’ve ever seen?
Well, yes and no. Often, people expect their files to be printable when they say they use ‘big name’ cameras with high pixel counts. It may again come down to the quality of optics, workflow, etc. I find that most photographers do not really understand their equipment and limitations. Much to my surprise, many of them do not understand ‘diffraction’ for example…they just stop down as much as possible to gain more depth of field. We must not forget that digital cameras are sophisticated pieces of equipment. Understanding the technology behind it is important to know what it can do for one’s needs. It is also a common belief that a larger, higher megapixel sensor will always perform better than a smaller one. While this is true in most cases, this cannot be taken as a universal truth. At system level, it requires equally matched lenses to get the most out of the sensor. I have seen both good and bad quality files from medium format to FF and crop sensor-based camera systems.
7. What qualities in a file make for a good print? What qualities will look bad? What do you look for?
Good pixels. Regardless of the total pixel count, I will always trade good quality pixels over anything else. That also calls for the need of a more disciplined and demanding workflow to ensure that the final file before printing holds as much pixel quality as possible. At pixel level we look for acuity which, which defines the contrast and micro-contrast of the image. Every image has to go through some interpolation before printing; the interpolation engine uses actual data as a basis before creating more to upsize an image. If quality is compromised at some point, either during capture or post processing, this will be further magnified on a larger print.
In my experience, most of the flaws were created by the photographer during post processing. The most common mistake is oversharpening and excessive noise reduction, which almost eliminates micro contrast. Overuse of blending modes with high pass/Gaussian blur are also highly detrimental to pixel quality.
8. All other things equal, and bearing in mind that due to the linear/nonlinear nature of screen vs print means that what works on screen might not work in print and vice versa – do some cameras produce more print-friendly files than others? Which ones, and why?
Frankly, as a master printer, I can never promise my client 100% accuracy from screen to print. As you have mentioned, there are also other factors involved – color gamuts vary from screen to print depending on equipment. I use monitors which have at least 100% sRGB coverage and >75% AdobeRGB coverage, calibrated to a delta E of less than 1.5. This mitigates deviations in the perception of luminance (calibrated to 100-120 cd/m2) and color, which is essential for soft proofing. When we look at a print, we are look at reflected light – which is very different from transmitted light through a screen. Our brains perceive colors and contrast quite differently under these two viewing conditions.
Most of the digital cameras produce acceptable colors given the current technology and maturity in RAW conversion software and its algorithms. A Bayer array based camera system seems to produce a slightly more consistent color image as opposed to a Foveon sensor, as color is calculated after capture in the Bayer sensor cameras. On the other hand, the color-aware Foveon sensor will produce very high quality pixels free from any form of color interpolation with very high pixel acuity.
9. Talk us through the importance of workflow, and what a good workflow process should include and avoid.
Here you go. As discussed earlier, the four pitfalls to avoid – resulting in degradation of pixel quality – are poor RAW conversion, aggressive noise reduction, oversharpening and poor interpolation.
10. Can workflow make up for raw resolution – and to put a popular myth to rest, where do you think the sufficiency bar lies for most people?
To achieve a good large print, the strategy within the workflow is to minimize processes that will degrade the pixel quality. In fact this is even more important with high megapixel images. Retouching strategy also has to be considered for portraiture and commercial work. Any flaws in retouching will be highly visible in a large print.
I started off making prints from 3-6MP at sizes up to 24×16”. Depending on viewing distance, most current DSLRs or mirrorless cameras at ~12-16MP should not have problems making prints at those sizes provided the workflow is solid. I have even printed to 4.5ft on the longer edge from M4/3 cameras with impressive results at viewing distances under 5ft. Ultimately, larger sensors do not equate to better results – it is the system level comprises in optics, camera, technique, workflow that matter.
11. Physical size is one obvious way to push the boundaries of printing and display; there’s immediate presence and impact, but some subjects simply aren’t suited to large prints and close range viewing – human body parts are a good example. More obviously, not everybody has the wall space to display something that large. How else can we push the limits?
I’m not sure if I fully agree with this, but sizes are relative. Consider art or painting as example: depending on subject matter, some images may look graphic or abstract at larger than life size. On the contrary, some landscapes create greater visual impact at smaller print sizes when matched with the right matte and frame. I guess we have to consider the final appearance of the print in framed form. Though it’s logical to consider a small print from a lesser quality image file, it is very satisfying to make high detailed prints at small sizes. The visual appearance is almost identical to a contact print.
12. Talk a bit about the role of colour management in the print process, and how we can ensure what we see is what we get.
Color management is purely science. It’s about transferring colors from screen to print based on the defined standard in the ICC (International Color Consortium) reference. The monitor needs to be calibrated, but prior to that, the quality of the screen and its color gamut are essential. We cannot see what the monitor cannot display. There is no point calibrating a screen if the monitor is not able to produce a good percentage of working colors, at least based on the sRGB color model. Most low end monitor cannot even reproduce 70% of the sRGB gamut. Essentially, the photographer may be missing 30% of the captured color information and even know it! Dedicated inkjet photo printers like the HPs and Epsons I use have printable color gamut exceeding sRGB on almost all papers and media.
A professional grade monitor may be able to reproduce more than 90% of AdobeRGB color space. These are typically 30-bit monitors at 10-bit per channel. The total color gamut exceeds 1.06 billion colors. That said, a good IPS-grade 24-bit monitor (8-bit channel) would suffice for most photography work with at least 100% sRGB coverage.
Calibrating a monitor literally means tightening the tolerance (measured in Delta E) to the reference. This is easily done with a good calibrator from X-Rite, Monaco, Datacolor or other reputable brands.
The last step is to match the color numbers to the printer via an ICC profile. While most will rely on manufacturer’s ICC profile, in my experience, it does not guarantee accuracy. It is however, a meticulous and elaborated process to try to create a custom (user) ICC profile. There are many factors to consider, including ink density and saturation limits for each type of paper, linearization and then only ICC profile generation. The most tedious process by far is getting the right linearization. With modern photo printers, we are not just dealing with CMYK inks – we must consider when the printer will switch over to LM, LC, Grey or even Light Gray inks, which greatly complicates the linearization process. However, linearization can only be controlled with manual intervention through Raster Image Processor (RIP) software. The typical driver does not allow user override – even the HP printers with built in spectrophotometers.
Assuming the print environment is fully calibrated from screen to the media of choice, getting higher accuracy of colors becomes easier. We can use the ICC profile to soft proof the image on screen in Photoshop. For most photography work, the rendering intent can be set as Perceptual. For some, I would prefer Relative Colorimetric (with Black Point Compensation) or even Absolute Colorimetric if the captured image is well within the color space and gamut of the paper’s ICC profile.
To be continued tomorrow in part two.
Wesley Wong is the founder of Giclee Art, offering extremely high quality printing solutions at a range of sizes. In addition to exclusively printing my fine art and commercial work, he also prints for quite a few of my students and a large number of the eminent photographers in Malaysia. He can be contacted via email here, or via his website. He also ships finished work internationally.
Places left for 2014 Making Outstanding Images Workshops: Havana and London – click here for more information and to book!
Visit the Teaching Store to up your photographic game – including workshop and Photoshop Workflow videos and the customized Email School of Photography; or go mobile with the Photography Compendium for iPad. You can also get your gear from B&H and Amazon. Prices are the same as normal, however a small portion of your purchase value is referred back to me. Thanks!
Images and content copyright Ming Thein | mingthein.com 2012 onwards. All rights reserved