Image quality is a difficult topic to define or strive for. But I recently asked Hasselblad how they achieve their legendary IQ and got the answers below. Before jumping to them, I’d like to thank Philipe Liljenberg (Brand & Content Manager), Mathias Elmeskog (Image Quality Specialist) and Ove Bengtsson (Product Manager) for making this interview possible and taking the time to answer my questions. It is much appreciated !
• Today, our sensors have quantum efficiencies in the 80%, so it seems they are as efficient as they will ever get. Do you see sensor technology as a major source of improvement for image quality in the future? Or will new sensors simply add more functionality (better AF speed, for example) ?
Mathias: Quantum efficiency (QE) at 80% is not a potential limit for sensor technology so there is effort to increase this number. With that said, QE is one important factor for image quality but there are other qualities for a pixel array to deliver good image quality such as lower noise, larger dynamic range, better analog signal gain, fill factor and reduced fixed pattern noise to give some examples. Different applications have different criteria so functionality, such as on sensor Phase Detection Auto Focus (PDAF), is developed in conjunction with improving raw data output quality.
• How does CMOS compare to CCD? Is the higher sensitivity largely an artificial amplification factor or is there a real advantage to CMOS? Do you feel modern CMOS sensors have bridged the gap with the best CCD sensors of their time, in terms of colours and tonal subtlety?
Mathias: Historically CCD have had the upper hand when it comes to fill factor and bit depth conversion since a CCD traditionally only uses one analog to digital converter (ADC). Due to progression in fabrication processes this is necessarily no longer true. With back illuminated CMOS sensors (BSI) the fill factor is very close to CCD (approx. 100%) and ADC technology enables higher bit depth per pixel. Other techniques for example dual ADC channel is often used to lower the read noise but keeping pixel bit depth. Color in general is independent whether it is CCD or CMOS since they absorb in the same spectral range. Color rendition has more to do with on chip color filters, bit depth and color processing. In short, yes.
• In astronomy and biology, EMCCD and sCMOS are commonly used. Do you see such technologies coming to high end manufacturers such as Hasselblad?
Mathias: At Hasselblad we strive for the best image quality so we look at new technology and make decisions with regards to all parameters what sensors makes most sense for the customer. Scientific CMOS (sCMOS) and EMCCD is just sensor evolution as any other so if the parameters outweigh other competing technologies this will of course be considered. sCMOS for example is not clearly defined but is rather a term used to express an improved raw data output matrices for scientific use with for example very low read out noise with no penalties on readout speed. These types of sensors have many compelling qualities in photography applications as well. EMCCD is more limited and aimed towards more specific types of scientific use such as single photon detection and very short exposure times.
• What can you tell me about camera calibration? Does it have to be an expensive labour to work well? What is its impact on image quality?
Mathias: To optimize image quality per unit camera every sensor goes through calibration processes in production. Doing per unit processing steps will always be considered labour heavy and in our calibration several such steps are taken. Adding to this some of the calibration steps require long exposures adding to production time. Knowing the sensor characteristics compensation can be applied to optimize image quality and per unit consistency. This lowers noise, increasing low contrast detail sharpness and removes the need for black image subtraction during long exposures for example.
• I have found Hasselblad XCD lenses to be extremely neutral. But the recent 80/1.9 seems to bring more “personality”. Is there room for more ‘fun’ lenses in the Hasselblad lineup, now that the brand is entering the high-end amateur market?
Ove: Not sure what you mean with ”fun”. We always try to make our lenses as good as possible. If “fun” means that it has a large aperture and can create a nice shallow DOF, this is not new to Hasselblad. For the V System, we had the f/2-110mm and for the H System, the f/2,2-100mm lenses that produced similar images. In the high-end amateur market, there are of course a high demand for exciting lenses and I am sure we will add lenses to the line-up that will excite this type of customer as well as our professional customers. I cannot reveal what we have for the future, but I can promise that there will be more lenses.
All photographs on this page made with the Hasselblad X1D. Can you hear me pining from afar?
#871. Monday Post (24 June 2019) – Battle of the Medium Formats… or?
#864. Hasselblad XCD 3.2/90 review. Better than the OTUS 85?
#855. Hasselblad XCD 3.5/120 review – dealing in gentle absolutes!
#850. 3 days in Paris with the Hasselblad XCD 4/21 (minimalist review).
#847. Hasselblad XCD 3.5/45mm review. The best all rounder ?
#844. Monday Post (15 April 2019) – 100Mp mirrorless cameras, at long last …
Please log in again. The login page will open in a new tab. After logging in you can close it and return to this page.