Sony Q&A @ CP+ 2017: Insights from the design team behind the bokehlicious 100mm STF apodization lens
posted Saturday, February 25, 2017 at 6:00 AM EST
With the annual CP+ tradeshow now under way in Yokohama, Japan's Pacific Convention Plaza, Imaging Resource founder and publisher Dave Etchells had the opportunity to sit down with executives from Sony Corp. to discuss the company's E-mount and A-mount optics, including the incredibly impressive FE 100mm f/2.8 STF GM OSS lens, an optic for which we just recently posted an in-depth field test. (If you've not already read our coverage of this interesting, apodizing optic which aims to provide luscious, creamy bokeh in out-of-focus areas, you'll want to read over that article first to familiarize yourself with the lens and its capabilities, and then come back to our interview!)
On hand to offer their insights into the company's products were Yasuyuki Nagata, Senior General Manager of Business Unit 2 in the company's Digital Imaging Group, Imaging Products and Solutions Sector, and Takashi Kondo, the Chief Marketing Manager of Sony Corp.'s Marketing Division, Digital Imaging Business Group. While regular readers may remember Kondo-san from our interview with Sony at the Photokina tradeshow late last year, this was Dave's first opportunity to sit down for an interview with Nagata-san. Given his close involvement with development of the 100mm f/2.8 STF lens -- not to mention our own keen interest in the optic -- the conversation predominantly focused on its design and creation.
Other topics of interest included the company's lens development strategy to date and its feeling for the needs of Sony shooters in the future, as well as some insight into what's necessary to create optics that are well-suited to both still and video capture. Without any further ado, let's get down to the interview itself!
Dave Etchells/Imaging Resource: The first question from one of our editors is that Sony has put a lot of emphasis on medium focal-length lenses in the last couple of years. You have a number in the 50mm and 35mm ranges, and it seems like you have the range from 24mm through 100mm very well covered, sometimes even with multiple lenses at different focal lengths. But we haven't seen much in the super-telephoto, or in the super-fast, super-wide area. And the G Master series has seen only limited introductions since last year. What's your philosophy on optics, and where do you think that Sony shooters are most wanting additional options?
Senior General Manager
Business Unit 2
Digital Imaging Group
Imaging Products and Solutions Sector
Sony Corp.
Yasuyuki Nagata/Sony Corp.: [That's a] long and very deep question. <laughs> OK, so first of all, after we launched the Alpha 7-series, we started with three lenses including the 24-70mm f/4, right? So we thought that a compact size is really important for the Alpha 7-series initially. But after we launched the Alpha 7R Mark II, though, a lot of professionals started using the [Alpha 7 series]. [They told us that] the 24-70mm f/2.8 G Master type of lens is definitely needed, so they are requesting [it]. That's why we decided to [develop] that kind of a middle zoom lens, the 24-70mm f/2.8 G Master.
DE: The mid-tele zoom, yeah.
YN: We focused on the current situation. But of course there are a lot of professionals that [are asking for] a telephoto or wide-zoom [lens], but at this moment, maybe most of our professionals are basically [shooting] portraits.
DE: Ah, I see.
YN: That's why they always requested the middle lens, you know, a 50mm or 85mm prime lens, that's the situation we focused on initially.
DE: Ah, I see. So the people that really were demanding the higher-end lenses were primarily portrait photographers?
YN: Yes.
DE: What do you think is the next-biggest request? What are people most asking for?
YN: <laughs>
Chief Marketing Manager
Marketing Division
Digital Imaging Business Group
Sony Corp.
Takashi Kondo/Sony Corp.: It is almost the same question as "Which lens are you going to introduce?"
YN: <laughter>
DE: You're too smart for me! <laughs> I thought I was being so clever. "What are people asking for?" <laughs> Oh well! Next question.
YN: Yeah, I cannot comment about future development products, but honestly speaking the landscape photographer is increasing [in Sony's customer base].
DE: They're increasing, yeah. I could see that, especially with the A7R Mark II, that when you have the very high resolution...
YN: Yes, I think so. Maybe the high resolution.
DE: Yeah. So I would think landscape photographers would like wide-angle probably. Interesting.
This is probably the answer to the previous question but we're curious if you have any interest in pursuing sort of ultra-fast, very-wide angle like the Sigma 18-35mm f/1.8. That's been a very, very popular lens on our site, and I think with consumers in general. Where does that kind of lens fit into your plans?
YN: At this moment we don't have any feedback or any requests to make those kind of APS-C [lenses].
DE: You're not really feeling a request for that kind of lens.
YN: Maybe they are just using the Sigma lens.
DE: Yeah, because their 18-35mm is an APS-C lens, and you really are having demand for full-frame lenses, that's where the emphasis is.
YN: If they request the APS-C type of lenses, basically they want to use it for video. [They] ask more about a long zoom.
DE: Ah, I see, so the APS-C people want a longer lens?
YN: That's why we've established the new 18-110mm f/4 lens to record movies.
DE: Interesting. Because you obviously have a very large userbase with the professional video cameras.
YN: Yes, yes.
DE: On the one hand, you would think that if there's a lens on the market that's selling well, that you should make one too, but on the other hand if someone already has a very successful product, maybe you should spend your effort making something different.
YN: Mm-hmm!
DE: Is part of your thinking that you want to make something where there isn't [competition]?
YN: Of course, we are always [monitoring] the market. But basically our product strategy is based on our customer feedback, our customers' voices.
DE: Ah, and so it's what your customers are asking for, that's something you should make. Yes.
YN: This time, we launched [the FE 100mm f/2.8 STF GM OSS lens], so that's a very requested [lens]. <laughs> Actually, the STF demand is very different from region to region. Japan and Asian countries have very high demand...
DE: Ah, yeah, yeah.
TK: ...unlike the US or the Europe.
DE: Well it's interesting, I mean bokeh is a Japanese term, I'm not sure I'm pronouncing it right. How do you say it, is it "Boh-keh"? "Boh-kay"?
TK: "Boh-keh", yeah, perfect. You sound perfect.
DE: "Boh-keh", okay. So bokeh is a Japanese term actually, and originally it just wasn't in American photographers' consciousness to be aware of how the background looked. Well Sony also had the 135mm f/2.8 STF.
TK: Yeah, that's A-mount.
DE: Yeah, that's an A-mount. It's an old lens I think, that came from Minolta originally?
TK: And you can't use the autofocus. Just manual focus only.
DE: Yeah, manual only. But that was a very famous lens.
TK: Mmm-hmm! Yes.
DE: Yeah. We haven't fully tested [the 100mm f/2.8 STF] in the lab, but just shooting with it we've been really, really impressed with it. Just the general quality and the bokeh in real-world images.
TK: Right.
DE: I'm kind of curious: It seems like the apodization filter could be very difficult to make. Is it done with a neutral density filter that's graduated, or is it done with optical elements so that you get light falloff? How does the apodization happen inside?
TK: That's a hard question. You are speaking to the right person.
YN: <laughter>
DE: Oh, really! <laughs> So Nagata-san's been working on this for years, and now it's like...
YN: Yeah. Maybe you're curious [how we] make the light go through the lenses to the image sensor. So basically, as you know, the apodization filter is like an ND filter, but it has a density graduation from center to edge, so it means that there are a lot of problems for light [reaching] the image sensor site, because light density and the strength of light. It is totally different from each position of the lenses. But we have lens technology and image sensor technology both, and additionally on top of that we have an LSI and software. We have a total solution, so that’s why we know where the light goes through the lenses and how the light is detected on the cell of the image sensor. We know that, and we can calculate and we can do a lot of the things in the LSI...
DE: Ah!
YN: ...so [it's a] total solution! To make it possible to [do] the STF autofocus system.
DE: Ah, that's very interesting. So it's a combination of making it STF but also work with autofocus?
YN: Yes. Maybe competitors can make STF lenses. But [STF lenses with an] autofocus system, they cannot, I think.
DE: Ah, yes. Yeah, because for autofocus, it's very difficult as you say. The light coming with different paths is going to have different...
YN: Yes, especially for the phase-detection autofocus.
DE: Yes, yeah, for phase-detection especially. And I'm curious, because you mentioned too, so the apodization filter is just a graduated neutral-density filter. The image processing, is that part of creating the density gradient also? Because I know that now with very powerful processors you can relax some of the constraints on lens designs, such as distortion or vignetting. I'm wondering, instead of working so hard to get the apodization filter exactly right, can you do some of the apodization in the processor or how does that work?
YN: <laughs> Actually we cannot say what it is exactly, but it is not so easy. Maybe the lens portion has a graduation like this. But [on] the image sensor, each point [receives] light from each [area]. So it's not so easy. All light paths should be controlled.
DE: Yeah. It seems like it must be very difficult to make an apodization filter with the smooth gradation.
YN: Yes. But sorry, we cannot talk [about it]. <laughs>
DE: I was just going to say, that's probably proprietary, yeah.
YN: It's classified! <laughs> Sorry about that.
DE: Yeah, well it's very impressive. So we have the 100mm STF. Fujifilm also has an apodization lens, but it has a much weaker filter. It's f/1.2, but it is a T-stop of T/1.7, but you're f/2.8 and T/5.6, so two stops. [Your lens] does a much better job of controlling very strong highlights. It takes more contrast to get a hard edge out of it. How did you decide the trade-off between light transmission and apodization?
YN: So I don't know what Fuji is doing exactly, but basically they just added the filter into the current lens.
DE: Yeah, they took an existing lens and added the filter in. Yeah.
YN: But our STF, we prioritized smoothness [of bokeh] at the initial stage of the development. So the lens itself is f/2.8, we are saying. But as you know, the old A-mount lenses have f/2.8 but the T-stop is T/4.5, brighter than this E-mount STF lens. This time we are saying that [this lens is] T/5.6, a little bit lower than current A-mount...
DE: Yeah.
YN: But we prioritized a curve of the graduation from center to [the edge]. We can simulate the bokeh situation totally on the computer, so we prioritized, we chose the curve to make it much better [with] the graduation.
DE: Mmm! To make sure I understand, so you could simulate the apodization on the computer, and that led you to see that a particular curve looked the best? And when you applied that curve, that ended up at T/5.6 and so you based the decision really on what looked best?
YN: Yes. You are right. The T/5.6 is just result. We just prioritized the graduation.
DE: I see. You weren't thinking "Well, we'd get even better at f/8, or you know, f/4", it's what looked the smoothest ended up being T/5.6, yes. Interesting.
YN: The lens itself has an almost f/2-level, the [physical] size.
DE: The size of the end [of the lens], which is why you don't get the cat's-eye distortion on the edges.
YN: Yes, yes! You are right, you know that.
<laughter>
DE: Yeah, that was also very impressive, that you go all the way up to the edge without seeing that. You had kind of alluded to it before that we would expect lenses with aggressive apodization filters to be difficult for phase-detect AF systems, although most modern AF systems work fine at f/5.6. The press release said that this new STF works fine with the hybrid phase-detect / contrast-detect system in the A99 and recent E-mount cameras. Did you have to do anything special in the lens design to allow the autofocus, or was that just firmware to be able to handle apodization?
YN: Ah, both.
DE: Both. So there are things -- you probably can't tell me...
<laughter>
DE: ...but are there things in the lens specifically to help the autofocus?
YN: Yes, contrast AF and phase-detection AF is totally different mechanically, in the lens side.
DE: Yeah, for contrast-detection the lens has to be able to move and stop very quickly.
YN: Yes, you are right.
DE: From that standpoint, wouldn't that be true of any phase-detect lens, not just an apodization lens? So is the 100mm f/2.8 STF different in that way, that it has to be fast? Is it different from any other lens you make that is also compatible with phase-detect? I'm wondering is there anything different with the apodization lens that helps the phase-detect part of autofocus?
YN: I see, yes. We cared about what the light path is. If we move the lens like this, wobbling, maybe the ND filter's effect might be like this, like a shaking. Any STF should be low, lower than that effect.
DE: It should, there should be less like change in the apparent aperture size when it wobbles focus?
YN: I think so.
DE: Yes. And I would imagine too that having the very large front element so you don't get cat's eye, that also means that the phase-detect elements at the edge can see more of the light paths?
YN: Yes, yes.
DE: How did you decide to do the STF versus a fast portrait prime? Because you mentioned that portrait photographers a big market for the A7 series; I would think that they would like a very wide-aperture, short tele, so what made you decide the STF versus very wide aperture?
YN: Yes, that just depends on the customer's voice, so yeah, at this moment the STF and standard middle-range prime lenses is definitely needed for professionals.
DE: That was just a bigger requirement?
YN: I don't know why, but maybe Alpha 7's customers [are portrait photographers], maybe the landscape photographer is increasing. Maybe next they will request more wider [lenses].
DE: Yes. And this may be the last question. Video is a very big feature for your cameras, especially the E-mount. As you're coming out with new lenses, how much consideration or how much effort is involved in considering video performance when you're designing them? What are the challenges associated with making a lens perform and handle well for still images and also for videographers?
YN: Basically, we consider the video function at the initial stage.
DE: Mmm! At the very beginning?
YN: All functions.
TK: Yes, especially E-mount.
DE: Especially for E-mount, because [video is a] very big part of your market. We were actually very surprised that Fujifilm announced Cine lenses, but they're coming first for E-mount, not even for their own cameras!
YN: <laughs> Maybe there are a lot of customers with our [cameras].
DE: I think you have a lot of customers, yeah, yeah.
YN: That's based on customer request.
TK: <laughs>
DE: And yeah, their customer request was to please make something for Sony's cameras. The other part of the question you answered that you considered video from the very beginning. Are there any challenges in making a lens be both a good still lens and also good for video?
YN: Yeah, sometimes it's challenging. Especially the focusing unit, the size and weight of the focusing unit is key for the videos. If the focusing unit is small, maybe for video shooting we can easily make it possible to...
DE: It's easier to move quickly and...
YN: But maybe in case of the prime lens or the high more quality lenses, the focusing block should be bigger then, so that's why we have a lot of actuators. But sometimes it is really too big to control...
DE: Quickly enough for video, yeah. Ah, that's very interesting.
YN: Sony wasn't really expecting to attach both video and still functions...
DE: They expected that both are good, yeah. That's a primary trade-off that you're having to make is to keep the focusing elements light enough but still have good quality.
YN: Yeah, for all the market voice. There are a lot of portrait shooters that they listened to, they have to shoot video together with one camera.
DE: Yeah. So even the portrait market, a lot of that now involves video. Ah, that's very interesting. Yeah, that's a good point, I guess I had remembered that from a discussion with another lens designer, they were saying in this particular case they had wanted to develop a lens for ultimate sharpness, but that also meant it was going to be slow focusing, it was not a sports lens because it had a very large moving element in it. Very interesting.
TK: We just got the report from WPPI. [At WPPI], the video seminar used to be only [about] the still photographers; portrait and wedding. But now one-third of the lecture is related to video.
DE: Really?
TK: There's a lot of demand for video for the wedding and portrait photographers as well.
DE: Still photographers, they're fighting to keep control of their customers because people will hire a still photographer and a videographer, and they want to do both. Well, I think that's all the questions I have. Thanks very much indeed for your time!
TK, YN: Thank you!