Photokina 2012: Nikon talks about the D800's outer AF issue

"I AM here4U" at Photokina 2012

Note: A copy of this blog article has been published in German language too. Please refer to the corresponding version (Deutsch) before posting comments. Thank You.

Prior to Photokina, I had scheduled a meeting with Nikon Germany, partly in order to discuss the known issue regarding the outer AF focus fields of the Nikon D800 camera.

So, on Tuesday I had the pleasure to discuss the topic extensively with both Michael Wollburg (Manager Customer Support Nikon GmbH) and a Nikon technician in charge. The meeting was very friendly and constructive.

Below, I am going to disclose my impressions which may differ from the original wording of what has been said (I have been there as member of accredited press; but I did not opt for an interview; I rather wanted to learn the bare facts). Therefore, any mistakes in the remainder are only my fault. Now for the infos:

• Nikon has acknowledged, found and understood the root cause of the issue. It has been eliminated in the current production (however, I guess we'll never know when and from which serial numbers on).
• The root cause is a misalignment of the AF module when mounted, outside of Nikon's own production tolerances. But be asserted we are still talking micro meters here ...
• The issue for affected D800 can be solved in selected Nikon service centers; such as Düsseldorf, Germany.
• The procedure is currently rolled out to more Nikon service centers.
• The fixing procedure for Nikon is a tedious one. It includes writing individual calibration values into the firmware. For larger deviations, the AF module will first mechanically be re-aligned. This may actually include the AF auxiliary mirror in some cases.
• This method is believed to deliver an autofocus precision which is at least as good as of cameras from a fresh batch. I could not clarify if there is a chance for both methods being non-equivalent in some way. However, Nikon Germany does not think so. They rather wholeheartedly believe that the in-service calibration procedure resolves the issue as good as current production does, if not better.
• There is some dispute about the eventual success of the service measures from Nikon so far. Not only are there differences between countries and early attempts have faced difficulties and a lack of a clear procedure. But there is additional uncertainty about what to reasonably expect, esp. with the large sensor resolution of the D800. Therefore, the eventual success of the service measures from Nikon must be gauged relative to what is feasible technically. Every AF module from every vendor has finite AF field deviations. Lumolabs determined (depending on the lens too) deviations of up to 40µm between AF fields (cf. www.falklumo.com/lumolabs/articles/D800Focus/SensorArray.html) as being usual in the market and acceptable for good image quality where 20µm would be very good. Such are accuracy values as found in other cameras (e.g., D3). According to Nikon, we shall expect the same (or better) for freshly produced batches of D800 from Nikon, or for D800 now being calibrated in service. LumoLabs intends to verify this claim. Until then, we cannot really know if isolated continued reports about inacceptable AF accuracy after service are due to a failure during calibration or an unreasonable expectation. It is this blog which actually proposed a method to distinguidh between the two. Nikon is supporting our contribution to bring the discussion back to facts.
• Nikon does actually not know how many D800 of the early batches have been affected. Despite all oddities, the so-called service-rate of the D800 is unremarkable and only "sligthly increased" (compared to other camera models).
• I take it that Nikon will not publish an official statement about this issue. And I take it too that meanwhile this has been internally decided in Japan. So, don't ask for it ... ;)
• Independently of corporate information policy, Nikon wants to help all customers with the problem without asking much questions. This includes returned goods from dealers and service calls after expiration of the warranty period. (Note however, that Nikon Germany cannot speak for the rest of the world.)
• I do now actually recommend affected D800 photographers, to contact Nikon or their dealer i order to schedule an AF calibration job done on it.
• Personally, I do think it is a very good idea to additionally create an A2C2 chart (before and after). This should help your camera's resale value too.

By the way, Nikon Germany has internally forwarded both our LumoLabs white papers and even had a closer look at FoCal. The AFMA unit step values we determined (approx. 4µm at f/2.8) have been received as plausible and it seems to be an accepted fact within Nikon that AFMA unit step values do indeed depend on a lens' nominal aperture.

It will be possible to check Nikon's claims (resolution of the AF issue both in service (normally) and current production) by using our testing methodology. If claims turn out to be true (as to be expected) then I think the entire AF field problem issue can settle. Slowly. But nevertheless make me make a clear call towards Nikon Japan: With your current information policy (or lack thereof) you are heading towards difficult times. I AM HERE FOR YOU ...

Falk

Further reading:
LumoLabs: Nikon D800/E outer AF sensor accuracy (First blog article)
LumoLabs: Using FoCal for testing of an AF sensor array (Second blog article)

Photokina 2012: Nikon spricht über das D800 AF Problem

"I AM here4U" at Photokina 2012

Note: A copy of this blog article will be published in English language too. Please refer to the corresponding version before posting comments. Thank You.

Ich hatte im Vorfeld der Photokina mit Nikon Deutschland vereinbart, ein Gespräch u.a. zum bekannten Problem der äusseren AF Sensoren bei der Nikon D800 Kamera zu suchen.

Am Dienstag hatte ich nun Gelegenheit, die Angelegenheit ausführlich mit Michael Wollburg (Manager Customer Support Nikon GmbH) sowie einem anwesenden Techniker zu besprechen. Das Gespräch war sehr konstruktiv.

Im Folgenden gebe ich meinen abschliessenden Eindruck des Gesagten wieder, nicht unbedingt den Originalton (ich war als Vertreter der Fachpresse akkreditiert, wollte aber kein Interview, sondern Informationen). Für Fehler in den Aussagen bin also allein ich verantwortlich. Hier nun die Infos:

• Nikon hat den Fehler acknowledged, gefunden, verstanden und in der aktuellen Produktion abgestellt (wann genau oder ab welchen Seriennummern werden wir wohl nie erfahren).
• Der Fehler ist ein Justagefehler des verbauten AF-Moduls, der ausserhalb Nikon's eigener Toleranz liegt. Allerdings reden wir hier immer noch von Mikrometern ...
• Der Fehler bei hiervon betroffenen D800 kann z.Zt. z.B. im Servicepoint Düsseldorf behoben werden.
• Weitere Servicepoints werden z.Zt. zur gleichen Leistung befähigt.
• Die Fehlerbehebung ist für Nikon aufwändig, hierbei werden Kalibrierungswerte in der Firmware hinterlegt und bei grösseren Abweichungen zunächst das AF-Modul mechanisch feinjustiert, was ggfs. den AF-Hilfsspiegel mit einschliesst.
• Die nach dieser Prozedur abschliessende Genauigkeit sollte der ab Fabrik mindestens entsprechen. Etwaige grundsätzliche Unterschiede der Prozedur in der Fabrik und im Service konnte ich nicht abschliessend klären. Nikon Deutschland vermutet solche Unterschiede nicht, sondern glaubt vielmehr aufrichtig, dass die Servicemaßnahme das Problem ebenso beheben wird, wie eine Neulieferung dies könnte, wenn nicht gar besser.
• Der letztendliche Erfolg der Fehlerbehebung muss auch unter dem Blickwinkel des technisch Machbaren gesehen werden, jedes AF Modul jedes Herstellers hat unvermeidbare Abweichungen. LumoLabs hat (auch abhängig vom Objektiv) Abweichungen von bis zu ca. 40µm zwischen den AF Feldern (cf. www.falklumo.com/lumolabs/articles/D800Focus/SensorArray.html) als branchenüblich und für gute Bildschärfe akzeptabel eingestuft, von 20µm bereits als sehr gut. Dies sind die Werte, die von anderen Kameras (z.B. D3) vorgelegt werden. Laut Nikon wird eine neu produzierte D800 oder jetzt im Service justierte D800 diese Werte erreichen oder übertreffen. LumoLabs strebt an, dies demnächst im Labor zu prüfen. Bis dahin ist unklar, ob vereinzelte Berichte über nach Service weiterhin unbefriedigende AF Genauigkeit überkritischen Kunden oder fehlgeschlagener Justage zuzuschreiben sind. Hier im Blog haben wir ja eine Prozedur vorgeschlagen, dies zu unterscheiden. Nikon unterstützt unseren Beitrag zur Versachlichung der Diskussion.
• Nikon weiss tatsächlich nicht wirklich, wieviele der D800 der frühen Chargen betroffen sind. Die Servicequote der D800 sei trotz allem "unauffällig" und nur "leicht erhöht".
• Ich gehe davon aus, dass Nikon keine offizielle Stellungnahme zu diesem Thema mehr abgeben wird und dies intern in Japan auch längst so entschieden hat. Man braucht also nicht länger danach zu fragen ...
• Unabhängig von der offiziellen Informationspolitik will und wird Nikon betroffenen D800-Kunden großzügig helfen. Dies betrifft auch Reklamationen nach Ablauf der Garantiezeit sowie Rückläufer aus dem Handel.
• Ich empfehle betroffenen Kunden nun, sich an Nikon oder ihren Händler zu wenden, um ihre Kamera zum nächsten passenden Termin justieren zu lassen.
• Persönlich empfehle ich noch einen A2C2 Chart davor und danach, auch zwecks Dokumentation im Falle des Wiederverkaufs.

Nikon hat übrigens unsere beiden LumoLabs Whitepapers zum Thema intern kommuziert und sogar auch FoCal evaluiert. Auch scheinen die AFMA-Schrittweiten (ca. 4µm bei f/2.8) in etwa zu stimmen und sie ist i.d.T. von der nominellen Blende des Objektivs abhängig.

Mit unserer Methode werden sich die von Nikon gemachten Aussagen (Fehler in der Produktion behoben, im Service i.d.R. behebbar) überprüfen lassen. Sollten sie sich (wie zu vermuten) als zutreffend erweisen, so sollte an dieser Front langsam Ruhe einkehren können. Dennoch trotzdem mein klarer Ruf nach Nikon Japan: Mit eurer Informationspolitik stehen euch schwere Zeiten bevor. I AM HERE FOR YOU ...

Falk

Weitere Lektüre:
www.panobilder.de/artikel/nikon-d800-af-sensor-genauigkeit/ (deutscher Hintergrundartikel)
LumoLabs: Using FoCal for testing of an AF sensor array (Blogartikel, englisch)

Photokina 2012: The Leica event

Leica Hall of Fame award ceremony, with Nick Ut, Phan Thi Kim Phuc, Karin Rehn-Kaufmann and Markus Lanz all on stage.

On September 17 at the Photokina 2012 Leica event, Nick Ut was awarded with the Leica Hall of Fame Award 2012 by Karin Rehn-Kaufmann. The price was for his famous "Napalm Girl" photo depicting Phan Thi Kim Phuc who became victim of the Vietnam war 1972.

In other words, Leica created enormous suspense to all awaiting Leica's new products. I almost thought the prize award would be what the event is all about. After all, it is quite an iconic photography and having both Nick and Kim Phuc on stage made it an impressive event. Even for people waiting for the whatever Leica would reveal. Like me ;)

So, it wasn't until 9:00 PM that Leica would launch their new products. And then it was pretty fast.


Four new Leica cameras

Andreas Kaufmann during product presentation of the new M

Leica revealed four new Leica cameras within just 30 minutes ... quite like a fireworks ;) :

• The Leica X2 in full color
• The Leica S3, called "new Leica S"
• The Leica M9mkII, called Leica M-E
• The Leica M10, called "new Leica M"
• Additional lenses (e.g., 3 new S lenses)
• An R mount to M adapter

The new S triggering interest on stage

The new M behind glass

First thoughts

Well, my very first thought has been, nice, but it isn't the Leica mirrorless many were expecting.

But this thing is bigger than one may think at first sight. Both the S and M are now CMOS, supporting live view and video. And Leica is making use of this.

The presentation has been rather short on details. But I grasped some additional material and a press kit on CD. So, there are some additional observations I'd like to share. As there are:

1. The M has no autofocus. But it has focus peeking and magnification, aiding in manual focus thru the lens, a novelty for rangefinder users. And yes, the M remains a true rangefinder camera!
2. The electronic manual focus is further assisted by the known EVF which can be attached.
3. Adaptability of R lenses means macro and tele applications are now quite feasible.
4. The broschure for the new M includes a text about a special microlens design avoiding color casts towards the edges with the new 24 MP sensor. This is a known problem with the 24MP NEX-7. Leica calls it the "Leica Max 24-MP sensor". I don't know if it is the same technique already known from the M9 or an improved one. But at least, it may mean that most (other) mirrorless cameras will be no competitor for M lens users.
5. The supplier for the new M CMOS sensor is CMOSIS in Belgium.
6. The new S has the same CCD sensor as the S2, no changes except for the sensor board. Both infos given to me at the Photokina Leica booth.

So, the new M is pretty close to a true mirrorless. It almost is the only full frame compact mirrorless camera. However, I consider autofocus to be a mandatory requirement for a mirrorless, without it it remains an almost pure rangefinder.

But Leica is now close. They deliver autofocus (S) and contrast AF (the X1 and X2) already. They even improved AF in the new S (added predictive AF). Therefore I think it is a matter of one more iteration of the M series, introducing AF-capable M series lenses. Leica probably fears to devaluate existing M lenses too much. So, it is better to deploy EVF (with peeking and manual focusing) and a true AF in two installments. At least, that's my interpretation of Leica's strategy here. It implies that there won't be a separate mirrorless line. Only yet another new M ;)

Of course, with the Sony RX-1 and anticipated NEX-9 which now most probably will become the first true full frame mirrorless,  Leica has only a given window of opportunity to populate this niche.


The main act

And then came the main act. At least, I heard organizers calling it this way ... ;)

Seal

Thanks for stopping by,

Falk

Updated September 18 (the original version contained some doubts that the new S sensor still is a CCD sensor. Meanwhile, I clarified with Leica and it is the same sensor as used as in the S2.)