Rozhovor s designery K10D (v angličtině)

[karak]

zdroj:http://ok1000.blogspot.com/2006/12/k10d ... entax.html

K10D Interview with development engineers of Pentax Japan

The major engineers of the K10D (Mr.'s I. Hirai, T. Hotta, and S. Morishita) have spoken. They're doubtlessly toiling away on future cameras, but with the present release of the K10D they're anxious to share their enthusiasm with an interview at ASCII 24. Just how different is the K10D from the K100D and previous cameras from Pentax? What part of the analog to digital conversion did the engineers focus on to help reduce banding in higher ISOs? How did the engineers design the weatherproofing? The veritical grip? And just what is their opinion on third party batteries? Again, with Ken Takeshita's help and special thanks to both the PDML list and ASCII 24, here is our translation of the Japanese ASCII 24 Interview where you can gleam some of the answers.

Interview with the K10D development engineers

What's the challenge involved in employing a 10MP CCD which is not particularly advantageous in sensitivity control? The first hurdle is to actually produce an image.

ASCII 24: The K10D finally adopted a 10MP sensor, but wasn’t it difficult to follow-up the previous successful image processing using a new CCD?

Pentax Team (Mr. I. Hirai, Imaging System Division, Development Department): The pixel pitch is smaller and the sensitivity has a one stop disadvantage (ISO 100 to 1600). This means a higher degree of noise, of which we were fully aware, so we strived for “film-like quality,” which was always our goal. It’s true that noise increases as sensitivity increases, but we wanted to make the noise look more natural. A special concern was not to create a “digital-like” image due to unnatural suppression of noise by image processing. In order to achieve more natural looking images, we conducted various fine tunings.

ASCII 24: The K100D employed the same 6MP CCD as the *ist D, DS2 and DL2. If over-emphasis on spec figures can be set aside, there seems to be more advantage in continuing to use the same CCD to obtain consistent and stable images. A brand new CCD was adopted this time around, but was this approach really necessary?

Pentax Team: Trial and error continued until the last minute, to the point people began to worry if the schedule could be maintained. But this is the same for our competition.

With a new CCD, the first difficulty is to produce any images. It doesn’t take long to produce an acceptable level of quality images at the base sensitivity (ISO 100), however, as the sensitivity increases step by step, at some point, terrible images will be produced. The real effort starts from this point, trying to pin point possible reasons, mainly by trial and error, often having to make new Cbs (circuit boards). At the same time, firmware is developing along and each group has to closely coordinate in order to slowly improve image quality. Fine tuning can only commence after this point.

The frequency of designing new circuit boards might be hardest hurdle in the whole K10D project.

ASCII 24: What was the most difficult task in properly tuning images?

Pentax Team: It must be measured against noise. Homogeneous noise across the whole image could often be permitted, but countermeasures against localized noise is difficult to achieve.

ASCII 24: What would be the main reason for such localized noise?

Pentax Team: If we knew that, life would be so much easier. Of course, each case is closely analyzed and dealt with, but there are a number of cases of localized noise and it is really time consuming to solve these. Changing Cb’s is one measure, but that has to be repeated many times.

We were fully aware that we were going to face very critical eyes on the noise issue because of increased resolution. In addition, the base sensitivity of the sensor was lower, which made tuning to the acceptable level for production very difficult.

ASCII 24: Does this mean that measures at the analogue signal stage might be important, since noise tends to come into a path between the CCD and A/D converter?

Pentax Team (by Mr. T. Hotta who was in charge of firmware design): Indeed. Recent Cbs might look at one piece, but there are actually layers to this and a slight change in wiring alone significantly changes how noise appears. In a digicam, DDR2 memory that functions at a very high frequency and many other parts all produce radio waves. A TV tuner portion of a videocard for PC, for example, is tightly shielded in order to avoid the interference between the digital circuit and analog circuit. If noise enters before signal’s digitization, the digital image output would almost be unusable. When increasing the gain for sensitivity, at ISO1600 for example (4 stops over base sensitivity), just a slight voltage variation in the order of a few milli-volts will impose a huge impact on image quality.

ASCII 24: Images produced by a digital camera are naturally recorded in digital form, but it sounds like effort to achieve the required quality is accumulation of “analog” work.

Pentax Team: Yes, work on the analog portion is critical, without which the later stages cannot be accomplished. It is usually a schedule-critical task and affects overall performance.

ASCII 24: The issue of interference between analog and digital circuits was raised, but I can see the importance of skillfully separating the two.

Pentax Team: Separating grounding circuits is one way, but it is not simply a matter of separation. Sometimes, good results could be obtained by a larger grounding area. This is another area requiring the trial and error method.

Merits of a 2 channel readout

ASCII 24: Readout from the CCD is in 2 channels, but isn’t that a bottleneck for burst rate?

Pentax Team: It’s sufficient to obtain the nominal performance of 3fps. With regard to the back end, bus control has sufficient room and can accommodate faster sensors. The maximum speed works best for post-processing, and this performance can be maintained with faster sensors.

ASCII 24: So it may have been better to use a 4 channel readout for faster signal acquisition, but isn’t 3 fps the limit with 2 channel readout?

Pentax Team: The maximum frame rate via 2 channel readout is indeed limited, and if a faster frame rate is required the sensor itself becomes a bottleneck. However, 2 ch R/O has merits over 4 ch R/O, namely lower power consumption and the ease to obtain higher resolution.

ASCII 24: More specifically?

Pentax Team: As you know, a single layer type CCD has Bayer type filter arrangement (R & G, G & B filters are arranged alternately). In a 4 channel readout, R, Gr, Gb and B are each read into separate channels. There is concern about the errors between Gr (G filter in red row) and Gb (G filter in blue row). In the case of 4ch R/O, each is read by a separate path, creating errors by different paths. These errors should be corrected in later processing, but if it’s not done properly, the resolution would have to be lowered, or banding shows up. On the other hand, in 2ch R/O, since one channel reads R & B, and another reads Gr & Gb, Gr and Gb signals can be processed equally in hardware. If the speed is of utmost importance, a 4ch might be appropriate, but we believe that, from the viewpoint of overall balanced performance, 2ch would be best.

ASCII 24: The image processing engine is named “PRIME.”

Pentax Team: Overall imaging performance has been much improved. The key is the “image processing engine” that has been renewed and produces very high performance. We thought we should make this appeal to users and put an name on it for the first time. We are, of course, trying to please with other portions such as the A/D converter, DDR2 and other upgrades as well.

ASCII 24: Reason for using DDR2 memory?

Pentax Team: Because of its overwhelming breadth at the bus. When considering fast framing of 10MP images, ordinary memory bandwidth cannot respond to the speed without some kind of stress. When we calculated the required memory bandwidth required to clear this problem, DDR2 was found to be most optimum under the current technology. Price is coming down and quality has been improving.

ASCII 24: I understand that the bus speed is faster now. What about the clock speed of main Cb?

Pentax Team: DDR2 was used in order to increase frame rate, but the data bus bandwidth is also broadened for faster media writing.

ASCII 24: By how much?

Pentax Team: More than twice than before. But the writing speed depends on the performance of the media too. You should consider that sufficient bandwidth is required there as well.

ASCII 24: Wasn’t it difficult to develop firmware because of the number of new hardware pieces?

Pentax Team: With regard to firmware, it had to be developed from scratch partly because of the newly developed PRIME engine. It is a complete different one from those used in other models. In order to improve the “quality’ of firmware, UML (Unified Modeling Language) was utilized. So, we have used a new approach and methods not found in this industry beforehand.

ASCII 24: What is UML?

Pentax Team: It’s an object oriented analysis method. Demand analysis (i.e., evaluation of required functions in this camera) was conducted before we started development of the firmware. There only very few manufacturers who introduced UML into assembly.

ASCII 24: We understand that Pentax determines target users very clearly before starting development of products. What sort of user demographic have you targeted for the K10D?

Pentax Team: The K10D is a step up from the K100D and it was Pentax’s first attempt to separate lower and upper models, forming a lineup. The K100D mainly targeted entry level but with the K10D, the target is those who have a good understanding of photography, and love photography. It could be what is called “advanced amateur,” in the 40’s to 50’s age bracket.

ASCII 24: What have you emphasized in targeting advanced amateurs?

Pentax Team: We discussed the addition and deletion of each function, suitable for the target users. For example, Auto Picture mode was eliminated, and two wheels would be necessary for a quick change of shutter speed and aperture, etc. To be an higher end model of K100D, a 10MP sensor would be suitable, and as resolution increases, the SR had to be beefed up. Weather/dust proofing is also demanded by this target user group.

ASCII 24: Pentax and Samsung are supposed to be cooperating each other. Is any Samsung technology incorporated in the K10D?

Pentax Team: We announced the collaboration with SamsungTech last year, but the development of the K10D commenced before that. Basic developments regarding the K10D are all supplied by Pentax. The GX-10 as sold in Korea has their own original GUI, which is different from K10D.

ASCII 24: Any shared parts between the K100D and K10D?

Pentax Team (by Mr. S. Morishita in charge of mechanism/body design): A portion of the mechanics is shared, but the K10D is mostly a brand new design.

ASCII 24: The K10D body became larger than previous models.

Pentax Team: Pentax cameras have been traditionally compact. But the body ended up becoming larger because of larger SR unit and the incorporation of weather proofing. We did not put priority into making the body compact, nevertheless, there was a size limit which we tried not to exceed.

It is true that design staff was debating if this size was acceptable when laying out the body design, obviously because Pentax has always stuck to the concept of compact size and light weight. However, the marketing department responded that performance should not be compromised in order to aim at a higher-model in the lineup. They also said that the size should still be attractive with weather proofing incorporated. The K10D is indeed bigger compared with the usual size of our traditional models, but it is not big compared with competition in same price and performance range.

ASCII 24: Why did SR unit become bigger?

Pentax Team: The magnet and coils were beefed up in order to secure stronger actuation of sensor. As resolution increases from 6MP to 10MP, pixel pitch decreases. This required beefing up of SR unit with some margin.

ASCII 24: What would be the benefits of a beefed up actuator?

Pentax Team: Response and braking accuracy have been improved. The mechanism has been further refined from that used for the K100D. Consequently, the correction improved by another 0.5 stop (2 to 3.5 stops in the K100D and 2.5 to 4 stops in the K10D).

ASCII 24: What’s the degree of weather proofing?

Pentax Team: In an SLR, it is rather difficult to apply the JIS (Japan Industrial Standard) classification because lenses have to be changed. We set our own in-house standard for weather/dust proofing, anticipating actual environments users might encounter, and exceeded those standards.

In the case of an SLR, the body is made of so many parts and cannot avoid complicated seams. All seams had to be sealed, then the weather proofing of moving control parts has to be evaluated beginning from the selection of material. For waterproofing, because just one missing or loose sealing will allow water or dust entering into body, we had be extremely careful. It is ultimately a choice of either “doing” it or “not doing” it altogether, and there is no in-between. In order to obtain perfect weather/dust proofing, every detail had to be plugged, taking a lot of time and cost. However, we believe we plugged all seams.

ASCII 24: When opening and closing the battery cover, I could feel a tight closure, but not so much on the memory card flap.

Pentax Team: For the memory card cover, a sealing material is pasted on body side and water/dust tightness is obtained by the cover being pressed on it. Perhaps that’s why you might have felt that way. In order to secure the seal tightness, the form of cap was so designed that it won’t be loosened by just one action.

ASCII 24: Hyper Program was resurrected. It was incorporated in the *ist D but not for the *ist DS and thereafter.

Pentax Team: Hyper Program was employed for the first time in the Z-1 that was released 15 years ago. It is the same thing in K10D.

ASCII 24: It’s hard to describe but it’s a very useful function in actual use. This time, exposure modes Sv and Tav have been added, too, which feels very refreshing.

Pentax Team: In retrospect, the installation of Hyper Program the on Z-1 was too ahead of its time.

But we believe that this exposure mode, which is very intuitive, has now become suitable and is finally appreciated in digital era. This function is hard to describe in words and just getting used to it is the best way to use this mode.

ASCII 24: While many makers are now employing a pentamirror finder, Pentax are using costly glass prism.

Pentax Team: The prism itself is the same as that used for the *ist D, but a brighter screen is used.

ASCII 24: The default setting for the K100D was “bright” but it’s “natural” for the K10D.

Pentax Team: We decided on “natural” because of target users being advanced amateurs who we thought would rather prefer “natural” setting. However, by seeing it “bright”, you can obtain the same taste of colour as the K100D.

ASCII 24: AF response seems improved. I felt a stronger torque on focusing ring during AF.

Pentax Team: It is probably due to the change of power source. It was changed from AA batteries to a dedicated Li-ion battery. As it can apply 7.2V consistently, the AF motor can take advantage of the higher voltage.

ASCII 24: Any other advantages in changing to Li-Ion battery?

Pentax Team: In using AA batteries, we found varying degrees of quality in offshore made ones, which required quite a bit of built-in margin in designing the camera, and could have made a negative impact on response time etc. We believe it was advantageous that we eventually adopted Li-Ion batteries which allowed us designing more tight and precise power supply.

With regard to the change of power source, this has been contentious since the beginning of product planning. On the one hand, we have been singing the praise on convenience of using AA batteries, and if we abandoned it, we were concerned if users wouldn’t accept it. On the other hand, from the development side, opinions were presented that more advantages could be enjoyed by switching to Li-Ion batteries. In the end, we believe switching to Li-Ion batteries was the right decision.

ASCII 24: Don’t you have any concerns over third party copies of batteries?

Pentax Team: Pentax have been using proprietary batteries for P&S digicams. We have been constantly watching low quality third party copies, but it is essentially a cat and mouse game. At present, there is no effective measure against poor quality copies. We do not believe it is beneficial to users by raising our cost incurred from this measure.

ASCII 24: So, there is no effective measure for this?

Pentax Team: In Japan, any copies are regarded something “dark”. Of course, we recommend using our genuine batteries. However, in the U.S., for example, third party copies are much more popular than in Japan. Among copies, of course there are poor quality ones, or even ones using questionable brand/logo design. However, there seems to be little perception that copies are necessarily of poor quality. There is this subtle difference between Japan and there.

ASCII 24: The battery grip option is available and this looks pretty good.

Pentax Team: The battery grip is often called a vertical grip and there has always been a strong demand for it by users who frequently shoot portraits. Also, some of the overseas users complained that the grip of K100D was still a bit too small and we enlarged the K10D grip.

ASCII 24: Batteries are inserted in both the body and grip. Which batteries get consumed first, from the body or the grip?

Pentax Team: It can be set by menu, but, basically, the larger remaining capacity takes priority. If only for more secure holding, you do not have to insert batteries into the grip. But it would be a waste of space, so we created a space for a remote and a card within the grip, which are protected from water/dust as the grip is also weather proof.

ASCII 24: Finally, please give us some last words from each of you regarding the attractions of the K10D from the viewpoint of design/development.

Pentax Team: We adopted a brighter screen which enables easier focusing. Pleae enjoy the Hyper Program and Sv/TAv modes. In terms of image quality, the priority is placed on depiction of detail. It is indeed different from 6MP images. Enjoy it.

We believe the basic specifications are excellent but please enjoy the user-friendly handling too. The “image magnification” with one-handed operation, and fine WB adjustments are unique to the K10D and not found in the competition.

As a designer for mechanical components, I would like to emphasize the beefed up SR and DR, etc. that are “supporting” features for shooting. Of course, complete weather-proofing, not found in the comparable competition, is also an important feature.

We can finally offer a product which appeals to those who “love photography”. We will continue to offer a lens line up with an emphasis on image quality.

Original interview copyright 2006, ASCII 24. Translated by K. Takashita and M. Gaudet.
Link: Japanese ASCII 24 Interview. http://ascii24.com/news/i/keyp/article/ ... 5-000.html

[monika]

co řikali?

[karak]

Je mi jasny, ze davat sem veci v anglictine neni uplne idealni, ale tenhle material mi prisel celkem zajimavy a pro nekoho to muze byt treba i inspirace pro zdokonaleni v anglictine

[asdf]

No, může to být zajímavé. Asi si teda budu zdokonalovávat angličtinu.

[vkramar]

co řikali?

Např. že Z-1 předběhl svoji dobu a že K10D líp ostří díky vyššímu a stabilnějšímu napětí Li-Ion akumulátoru oproti AA...

Zítra bych měl mít trochu času, tak se pokusím pár bodíků vypíchnout v jazyce českém.

[vkramar]

Takže tady je pro angličtinou nevládnoucí Pentaxisty VELMI stručný výtah pár věcí, které mě zaujaly:

Add 10 Mpx senzor

V boji s vyšší šumem způsobeným mj. menší roztečí snímačů se snažili spíše o přirozenější vzhled než o jeho maximální potlačení.

Nižší rozteč si také vyžádala posílení mechanizmu stabilizace, za cenu větších rozměrů. Ale zase by i díky tomu měla stabilizace zvládnout o 1/2 EV delší časy než u K100D (pozná to vůbec někdo? ).

Zpracování signálu

Dali přednost 2kanálovému zpracování (1. R + B, 2. Gr + Gb) před 4kanálovým (R, B, Gr a Gb samostatně) i za cenu nižší propustnosti, aby se vyhnuli chybám během zpacování při oddělení Gr a Gb. Přesto však díky DDR2 pamětem a dalším vylepšením údajně více než zdvojnásobili frame rate.

Rozměry a konstrukce

Přednost dostala robustnost a odolnost před miniaturizací. Mj. i proto je K10D oproti K100D o poznání větší a těší. Dále si mírné zvětšení vyžádal i např. rozměrnější mechanizmus stabilizace.

Pro ty, kterým se tělo zdá příliš malé, nebo pro ty kteří fotí převážně na výšku, je k dispozici battery-grip (alias vertical-grip). Dá se použít i bez baterie, pokud jsou osazené obě baterie (v těle i v gripu) používá se přednostně ta s vyšší zbývající kapacitou, ale je i možnost zásahu v menu.

Napájení

Jedním z důvodů pro přechod od AA k Li-Ion akumulátorům byl požadavek na zajištění vyššího a hlavně stabilnějšího napětí. Příjemným vedlejším efektem je např. silnější AF.

[monika]

vkramaráku dik za překlad