| PPA '99 Highlights | |
The '99 PPA (Professional Photographers of America) Annual
Convention and Trade Show is now underway in Atlanta, and
we'll have some selected highlights from it appearing here over
the next few days. - There isn't a whole lot that's truly new,
so there won't be a huge amount of news from it, but there are
a few things worth reporting:
PPofA Day Two, 27 July, 1999
Well, actually it was day 3 or 4, depending on how you're counting,
but it was day two for us...
Kodak DCS330
We're indebted to Steve's
Digicams for breaking the news on Kodak's new DCS330 digicam,
in time for us to catch it at Kodak's booth at PPofA! (The press
release got to us a day or two late.) - We'd heard rumors of a
new product coming, but weren't expecting it for another month,
at the Seybold show in San Francisco. (That was the source of
our allusion to "fireworks" at Seybold, in the first
day's report on the D1.) We don't know if it was planned all along,
or if Kodak decided to move up their announcement date on the
DCS330 to counter all the noise Nikon's been making over their
D1. Certainly, Kodak's usual approach is to not announce products
until they're actually ready to ship, so this was a bit of an
advanced announcement, but OTOH, Kodak will apparently be ready
to ship their DCS330 well in advance of Nikon's D1. - Kodak is
saying "August", but not whether it will be early or
late August. Nikon is saying "September", but again
not whether early or late, and is hedging their comments somewhat
by essentially saying "but don't expect to find one at your
local camera dealer until January or later." (Actually, Nikon
didn't phrase it quite that baldly, but did clearly state that
they weren't expecting to be nearly able to meet demand for the
D1 for a long while after it officially begins shipping.)
Back to Kodak though, this is their section! - Part of the reason
the DCS330 will be shipping so soon, and apparently in quantity,
is that it's essentially an upgrade to their DCS315, a 1.5 megapixel
design, using the same Nikon Pronea 6i body, and apparently much
of the same internal circuitry. Enhancements include a 3 megapixel
sensor, boasting an image size of 2008 x 1504 pixels, a lithium
niobate low-pass filter to eliminate color artifacting, and a
PC connector for external flash connections. (The 315 would work
with external flash units, but only via the top-mounted hot shoe
connection.) Another significant enhancement is the use of Kodak's
new ITO (Indium/Tin-Oxide) gate electrode technology, which improves
overall quantum efficiency, and boost blue-channel responsivity
by an absolute factor of 3, and a factor of roughly 2 relative
to the also-increased responsivity of the green and red channels.
Say huh? Basically, Kodak's new full-frame CCD technology significantly
increases overall CCD sensitivity, with particular improvements
in the always-problematic blue channel. Bottom line, the 330 should
have much-improved image noise characteristics relative to the
315, particularly in the blue channel. Full-sized bitmap files
from the DCS330 are fuly 8.65 megabytes in size. The CCD on the
DCS330 uses 9 micron pixels, and has a rated sensitivity range
of ISO 125-400. It's physically larger than that in the DCS315,
which also has the same pixel size. The net is that effective
lens focal lengths are multiplied by 1.9x relative to a 35mm frame,
a considerable improvement over the DCS315's factor of 2.6x. (For
a review
of the earlier DCS315, visit John Cowley's Lone
Star Digital site.)
The DCS330 incorporates
the histogram exposure display that Kodak popularized on their
higher-end digicams, and that we applaud as the best way to communicate
exposure information. In addition to this, the camera has an LCD
mode that will blink any pixels that are overexposed, letting
you quickly see if you need to turn down the exposure settings.
(The histogram at left reveals that the scene was underexposed
overall, with no data values above roughly 75% brightness.) (Click
on the image to see a larger version.)
Argh! Bonk, bonk! (Writer hitting himself on the head) Unlike
Nikon, Kodak was letting people take actual pictures with their
demo units, but your brilliant author neglected to capture a sample
to take back with him for posting here! - OTOH, these were still
preproduction units, with labels saying "Engineering model
XYZ" on the bottom, so it's likely that results obtained
with them wouldn't have been representative of the final units.
We did impose on the booth personnel to let us look at an image
shot with the camera in Photoshop, and scrutinized the blue channel.
The result? Well, there was still some noise there, relative to
the other channels, but less than we're accustomed to seeing.
We also realized after we left the booth that the shot we were
looking at had been captured at the camera's top ISO rating of
400. (One of those delayed eyes-to-brain communications: "Oh,
duh - That's what the little number "400" on the CRT
meant.") The images did look very nice, with beautiful skin
tones and seemingly very good dynamic range, athough again, it
was difficult to make any detailed assessments in the trade-show
environment. The images we looked at close-up didn't seem as sharp
as we'd have expected from a 3 megapixel sensor, but (pay attention)
this may not have meant anything as at least one of the lenses
Kodak was shooting with wasn't working properly, either in autofocus,
or in auto-aperture mode. Likewise, all shots were handheld, via
available light, and so very prone to camera shake. (I felt I
needed to comment on the sharpness, to head off the inevitable
flood of emails asking about it, but don't want to be the source
of a rumor that the 330 is "soft." - Let's wait to look
at production units tested under well-controlled conditions before
making any judgements of that sort!)
Actually, the issue of the malfunctioning lens & other problems
was the source of some frustration for us: Granted, the booth
personnel hadn't even seen the devices until the day the show
opened, but there were a lot of rough edges around their presentation.
We heard one staff member explaining that the lithium niobate
filter (see below) was there in place of having to use a "hot
mirror" filter on the front of the lens (totally incorrect,
AFAIK); all of the AA batteries in the booth were dead
by the time we stopped by early Tuesday afternoon; and one of
the Nikkor lenses they were using to demo the cameras with just
plain didn't work. As a result of the dead-battery problem, interested
visitors were figuratively chained to the booth counter by the
AC-adapter power cord. Maybe I'm just tired and cranky after an
exhausting day at the show, but it doesn't seem that it should
be terribly difficult to make sure you have enough batteries and
chargers to keep everything ticking along for a new product's
debut at an important professional show! Ok, flame off...
Speaking of batteries,
we really liked the 330's (and 315's, apparently) implementation
of a AA-cell battery holder that makes 6 AA cells into a single
pack that slides into the side of the camera. (See picture at
right.) This struck us as a very handy arrangement, making it
easy to just pack along 2 or 3 sets of batteries on a shoot, and
to change between them very quickly. (For the record, Kodak specifies
battery life at 200 shots per charge, with high-capacity NiMH
cells.) The shot at right also shows the flash PC sync connector
on the right side of the body, just above the battery compartment,
as well as the lithium niobate filter, in an oblique view. (The
rectangular object with one corner cut out, inside the camera
body.) Dual TypeII PCMCIA slots are located immediately behind
the battery compartment. These are full-sized PCMCIA slots, so
they'll accept cards as large as the 500 megabyte Type III hard
disk cards, as well as the larger Type II solid state cards. With
a TypeII PCMCIA adapter, the unit should also accept the IBM Microdrive,
but our understanding is that this device hasn't yet been certified
for use in the 330. (No reason to believe it won't work, just
that Kodak hasn't tested and officially blessed it yet.) The 330
(and 315 before it) sports dual PC-card slots, with the promise
that you'll be able to load cards into both slots, for
greater recording capacity. Apparently, this capability isn't
yet implemented in any of Kodak's camera firmware, but is planned
for release in October, as a firmware upgrade. (Don't you love
cameras with upgradeable firmware?) Another firmware issue has
to do with file formats: There was some confusion over this: The
DCS315 produces "finished files" that can be read directly
into your computer. The DCS330s at the show apparently needed
to have their files processed by Kodak's software to turn them
into standard JPEGs and TIFFs. According to booth personnel, a
soon-to-arrive firmware upgrade will permit finished-file processing
in the camera.
Dual Slot Notes - GPS coming
While Kodak doesn't yet have the ability for their cameras
to address memory cards in the dual PCMCIA slots, they have grander
plans for the near future. It's turning out that GPS interfaces
are enormously popular in government applications, to pinpoint
just where the pictures were taken, important for municipal road
and facilities surveys. ("Say, Billy Joe, what should I put
on this here form for 'location'? - "I dunno, I think that
was the Jenkins' hog farm we passed a mile or so back, just say
'west of Jenkins'...") Apparently, there will soon be a GPS
interface available for the Kodak cameras that will let you plug
a GPS card into one slot, and the memory card into the second.
(We're told that this will be an across-the-board capability for
all the high-end Kodak cameras with dual slots, and that it will
again, be a firmware upgrade.)
Low-Pass Filter
This shot shows the lithium niobate filter more plainly, just
inside the camera body, as well as a Calluna Type III hard drive
card poking out of the PCMCIA card slot. To our knowledge, Kodak
was the first company to implement a lithium niobate low-pass
filter in their portable cameras. It's function is to knock off
high spatial frequencies, without attenuating lower frequencies,
below some cutoff level. The net result should be a lower level
of color aliasing, with little overall loss of sharpness. (Nikon
is now also using a lithium niobate filter in their D1 model.)
A low-pass filter makes a lot of sense for digital cameras as
a way to eliminate color artifacts or "twinkles" around
high-contrast edges.
The diagram at right (from our
Buyer's Guide to Digital Cameras)
explains what happens to produce problems of this sort: Here,
we're looking at a sensor array that has a linear RGB color filter
array laid over the top of the CCD cells. Suppose we're shining
bright light onto the sensor, but a small dark object (with incredibly
sharp, well-focused edges) is obscuring the red-masked sensor
element. What will the camera/computer think it sees? - Basically,
it'll think it's looking at a very bright cyan-colored object
(blue and green at maximum, no red). That's an artifact! Ideally,
a low-pass filter would prevent there ever being an edge that
was so sharply defined as to produce a color artifact like this
(having high spatial frequencies), without affecting edges that
were only a little less sharp, which wouldn't have produced an
artifact on their own. Very tricky optics to make this happen,
which is why you haven't seen it on even high-end professional
cameras until Kodak's recent 520/620 models! So how well does
it actually work? Hard to tell: I'm pretty sure I saw some color
artifacts on some of the images Kodak shot at the show with the
330, but wasn't sure, not having shot the images themselves. (FWIW,
this is an issue that Nikon claims they've completely addressed,
and indeed their printed samples showed no sign of color artifacts.
BUT, we don't know what sort of processing those pictures had,
as Nikon still isn't letting anyone see raw image files.) About
all we can say is that Kodak's lithium niobate filter produces
dramatically better results than we're accustomed to seeing on
earlier Kodak cameras that didn't have it. More than that will
require actually working with the cameras in our studio. Oh -
the low-pass filter can also be removed by the user at any time,
in the event that it might interfere with whatever type of shot
the photographer is trying for.
Price: The $5,995 question.
Booth staff were quoting a list price of $5,995 for the DCS330
when it's released, and predicting ample quantities. For that
price, you get the camera body, a power adapter, and a battery
holder: No memory card or lens is included. (This is SOP for Kodak
professional cameras, and apparently the Nikon D1 will ship essentially
the same way.) Meanwhile, the 315 will apparently stay in the
lineup, at a lower price, although we don't know Kodak's pricing
plans for it. (See our speculation below though.)
Kodak vs Nikon
The really big question is how the DCS330 and Nikon D1 will fare
against each other in the market. There are a lot of obvious differences
in the cameras specs, design, and ergonomics, but the big unknown
will be image quality. The D1 wins on a number of spec-sheet features,
with a far more sophisticated camera body, much faster "burst"
shooting (4.5 fps, vs 1 fps for the DCS330), and a lens focal-length
multiplier of only 1.5x. Nikon also (rightly, I think) makes a
big deal of the camera's "feel", which is virtually
identical to their own pro-level F100. As noted, image quality
will be a large deciding factor for many people. It's also entirely
possible that many people will prefer the simpler function set
of the Pronea body that Kodak's using. Now wait, before you jump
on me for that, and think about who might be interested in these
cameras. Certainly, many professional photographers and photojournalists.
For photojournalists, the D1 wins hands down, but it *may* not
win relative to the higher-end Kodak models that already dominate
that market. (Price of the body may be a less-significant factor
in the photojournalism business, and having all of your
cameras be the same model could be very important for automating
the post-exposure processing for color management, etc.) We suspect
that Nikon-based pro commercial photographers would also gravitate
to the D1 as well. On the other hand, there's a large component
of Kodak's sales that go into government and corporate markets,
where the people using them aren't necessarily professional photographers,
but rather just people using them for documentation-type photography
in a workplace environment. People there may favor simpler devices
that demand less of the user. (This of course begs the question
of whose user interface is ultimately simpler to use: We've had
no time with the DCS330 to form an opinion there, and nobody
has had any time with the D1's user interface, because it doesn't
exist yet. Another factor in market success will be production
capability and distribution channels. Kodak claims to be able
to supply whatever quantity the market needs, while Nikon is frankly
stating that they expect shortages...
What to draw from this? Certainly not that I'm taking a position
of Kodak over Nikon: All I'm pointing out is that the Nikon/Kodak
race is only just beginning, and that Kodak won't willingly cede
their highly profitable professional camera business without a
fight. With their recent price drops on the high end of their
line (DCS520 and 620), combined with a very generous bundling
deal ($3,400 of free products, such as high-capacity memory cards),
Kodak will be applying heavy pressure on the upper end. (The DCS520
and 620 are currently selling for $10,000 list: We'd guess that
in a month or so, Kodak will drop the bundle, as well as the price,
bringing the cameras into the $6500 range(!) - This apparently
is why Nikon is only claiming a 20% price advantage for their
D1, and concentrating more on the "feel" of the camera
in their pitch.) On the low end, we expect to see the DCS315 price
decline, and wouldn't be surprised to see it at $4,000 before
the end of the year. Stay tuned, it's going to be a *very* interesting
year in high-end digicams!
Foveon Part II
Well,
actually, this is going to have to wait for a few days. I reconnected
with my old professor Carver Mead, and had an exceptionally interesting
discussion about some of the technology behind the Foveon camera.
There's clearly a significant body of new technology there that
will come out over time. I'll try to get an update posted here
before long, and we're hoping to do an extended story on the camera
in the near future. Stay tuned on this one, too...
(A quick snapshot though: The camera is tightly targeted at high-quality,
real-time "people" pictures - portraits, fashion, studio
wedding work, etc. Its controls seem to appeal to pros in that
market, although at a lease price of $1,600/month, it'll take
a fairly successful & busy pro to afford one. Foveon though,
claims that any pro grossing more than $150,000 through a portrait
studio per year has a very easy cost-justification. Much of the
technology is under wraps, while Foveon waits for more than 30
patents (!) to be processed, but we can say that at the results
level, the camera is very easy to use, and takes extraordinary
pictures that can be enlarged to enormous sizes without any hint
of pixellation or artifacts. Definitely a very interesting entry
for its narrow market, but we believe Foveon has the technology
and plan to revolutionize portrait studios!)
July 25, 1999:
Little new about Nikon D1, but spectacular samples!
We regret that there's little
new to report on the Nikon D1, not even any shots of the user
interface: The unit on display at the PPA show was the same previously
viewed and described by Rob
Galbraith. Most of its user interface code was simply not
present, and about all the prototype would do was take a picture
and display it on the back-panel LCD screen. Still, the excitement
level was high, as proud papa (well, proud uncle?) VP Richard
Lo Pinto of Nikon Professional brandished the alpha version of
Nikon's new baby, and masterfully spun the tale of the new unit's
wonders.
As expected, Richard was completely mum about details such as
the brand of CCD being used. Some sensor specs were available
though, including dimensions of the chip (23.7 x 15.6mm, or 46.5mm
across the diagonal) and the pixel dimensions (11.8 microns).
The nearly 12-micron pixels are indeed rather large when compared
to consumer-level devices, but slightly smaller than Kodak's latest
generation of enhanced-responsivity CCDs, which provide a 200-1600
ISO rating with 13 micron cells. Other details could be inferred
from the graphics shown on the computer slide show running alongside
the D1 demo station: If the grapic is truly representative, it
appears that the Color Filter Array (CFA) uses equal populations
of red, green, and blue pixels, which may be one way of dealing
with blue-channel noise. (Some high-end cameras use twice as many
green pixels as a way of increasing apparent resolution, since
the "luminance" channel of the human eye tends to track
green wavelengths most closely.) Another interesting tidbit was
that the sensor on the D1 uses a microlens array over the CCD's
surface: This implies that the sensor uses interline-transfer
technology, rather than the "full frame" approach of
Kodak's high-end sensors. What does this mean? Interline transfer
technology is generally less sensitive to light than the full-frame
approach, as it requires more cell area for electrodes and non-light-gathering
structures. The use of a microlens array can compensate for this
somewhat (by concentrating light on the light-sensitive sensor
area), but often at the cost of vignetting in the optics, due
to varying entry angles of light rays coming from the primary
optics to individual microlenses. This tendency could be corrected
by acquisition software, and we saw no sign of the artifact in
printed samples displayed by Nikon in the booth. On the other
hand, interline transfer CCDs permit electronic gating of the
exposure, which may contribute to the exceptional 1/16,000 second
minimum "shutter" time of the D1. Interline transfer
also facilitates real-time video viewfinder operation, as data
can be clocked out of the sensor while it's being exposed to light.
(The prototype did not have a "live" LCD view, however.)
We did manage to clear up one point of confusion that's been raised
over the camera though: People (or at least Rob Galbraith) have
noticed a "vibration reduction" mode on the list of
features on the data sheet. This led some to wonder if there were
some sort of active anti-shake system implemented, or if there
were a feature similar to the "Best Shot Select" of
the CoolPix 950. We learned that the reality is much more mundane,
and is actually a purely mechanical solution to shooting with
long teles on tripods: In "vibration reduction" mode,
the camera simply delays tripping the shutter for a short while
*after* the mirror flips up, to give vibration from the mirror
actuation time to die out. The reasoning for not including a "best
shot select" was that pros want to know exactly when they're
taking a picture, and don't want to have to worry about the camera
automatically choosing the wrong version of an image. (The 4.5
frame per second "motor-drive" speed would accomplish
much the same thing, assuming the photographer had enough image
storage on-board to preserve all of the images thus captured.)
We were particularly impressed with another spec-sheet parameter
that didn't really hit home until we were talking with Mr. Lo
Pinto: The shutter-release lag time of the D1 is a scant 58 milliseconds,
or only 0.058 seconds! - Quite a contrast to consumer-level digicams,
where 0.5 seconds is considered fairly fast!
One interesting note we picked up on was Nikon's comment to a
prospective customer that the D1 was priced "20% below the
competition." - We found this interesting, because many people
are comparing the D1 to much more-expensive pro cameras from Kodak,
selling in the $10-16,000 price range. It may be that Nikon is
prefiguring the impact their unit may have on the competition,
but based on our speculation about the CCD above, it's also likely
that Nikon themselves aren't putting the D1 in the same ballpark
with Kodak's DCS520 and 620, from an image-quality standpoint.
On the topic of image quality though,
Nikon showed one set of printed samples that absolutely blew our
socks off! (Still no electronic files, hardcopy prints only.)
The images in question showed a night scene of a suspension bridge,
shot at a 1/2 second exposure time. (No other exposure information
was given, such as ISO rating or aperture.) What Nikon did was
to first output the image as a 9x12 full-frame print, and then
proceed to crop into the image successively, magnifying each cropped
version back to the full 9x12 size. We were absolutely amazed
by how well the image held up to totally unreasonable levels of
enlargement: The highest magnification shown corresponded to a
full-frame print of roughly 5 x 8 feet! At that level, we could
clearly see some individual pixels, and noise was evident in the
3/4-tone areas. What amazed us was the extent to which tiny details
(such as the support wires of the suspension bridge) were faithfully
rendered in the final enlargement. Even more surprising was the
*complete* absence of any color artifacts around tiny white lights
on the bridge: Past experience would have led us to expect at
least some level of artifacts on these objects. Nikon claims a
new and proprietary color-interpolation algorithm used on the
D1, and we'd be willing to believe it, given the images shown.
OTOH, Kodak has recently begun using an optical low-pass filter
on their high-end cameras, which should also contribute to greatly
reduced artifacting. Bottom line, we're going to have to wait
for a trusted third party to test the D1 and publish the images
before we can arrive at any firm conclusions about image quality.
(Hint, hint...) Nikon showed an image sample that compared the
color-artifact performance of a brand "X" camera to
that of the D1 (see the blurry picture of a video display of a
PowerPoint slide show above right): While the difference was obvious,
the comparison was essentially meaningless, since the "Brand
X"camera in question wasn't identified. (A DCS 410?)
Overall, the D1 shows much promise, but any worthwhile opinion
will have to await release of actual working models. - By the
way, the unit to be displayed in Florida at the Visual Edge '99
conference at the end of August will still be Unit 43, the same
one as at PPA, and that Rob Galbraith reported on earlier. There
should be an almost fully-functional unit appearing at the Seybold
conference August 30. (BTW, our grapevine predicts there may be
some fireworks at that conference in the high-end arena, so stay
tuned...)
New Nikon Lenses & Flash
At the same time as they announced the D1, Nikon also announced
a couple of new lenses and a new flash. The lenses can be used
with any Nikon camera ,but were coupled with the D1 launch because
they fit the camera's needs rather well. The first new lens is
the 17-35mm zoom. This very wide-angle zoom addresses the need
for wider angular coverage in the D1, due to the sensor size being
smaller than the 35mm frame (Making about a 1.5x increase in "equivalent
focal length", although Nikon took pains to point out that
it isn't a change in effective focal length, but just a cropping
of the film frame.) The second lens is more interesting in our
view, being a true tilt/shift lens, with greater range than the
equivalent lenses from Canon. It's an 85mm unit, with "micro-Nikkor"
technology, making it ideally suited for tabletop product photography.
We were surprised at how much movement you could get from this:
Really pretty significant tilt and shift. (Sorry, no specs, we'll
try to remember to pick up a data sheet when we're back down to
the show on Tuesday.) The reason we feel this lens is so key is
that it finally allows small-format users adequate perspective
control for product photography: Couple it with the D1, and you'd
have an incredibly productive catalog-shooting machine! The SB-28DX
flash is a modification of the standard SB-28: All the same functions,
only modified to work properly TTL with the D1: The SB-28DX reads
flash exposure based on light reflected off the shutter of the
D1, which is painted (anodized?) a neutral gray. This gives flash
metering in almost "real time", after the mirror flips
up, immediately before the shutter itself fires. (The TTL metering
of the SB-28DX and the SB-28's multi-unit TTL capability could
combine with the 85mm tilt/shift lens for a very effective studio
system for small-product photography. - We predict this will be
a huge market for the new system!)
Mini-Softbox
Taking a break from the high end of the market, we saw a cute
flash accessory in the Photoflex booth: It's an *inflatable* soft-head
to attach to any conventional flash unit! These have been around
for a while I think, but this was the first chance we've had to
try one. The shots below show the soft box itself, and the good-natured
Photoflex employee Ben Clay, who cheerfully posed for an A/B comparison:
The effect of the soft box is a bit subtle in this on-the-spot
test, but the image on the right is clearly softer and more evenly
illuminated. The mini-softbox comes in two sizes, to fit a variety
of flashes, but generally will only work with off-camera flash
heads, due to its size and shape: It worked OK on our venerable
D-600L, due to the 600's pop-up flash head geometry...
 |
 |
 |
HIGH end - The Foveon Camera
We'll have more to report on this
later in the week, after we have the chance to meet with Carver
Mead, company founder and inventor of the camera: There's been
a lot of buzz in recent months about a new-concept camera from
California startup Foveon. Images from the camera displayed at
enormous sizes in the booth showed some of the reason for the
excitement: Single-shot captures that scale all the way to ~4
feet across with excellent quality: The images below show the
Foveon camera (which is integrated with a laptop computer), and
a shot of head engineer Richard Lyon that was enlarged to something
like ~3-4 feet across. Next to the shot of Lyon is an ultra-closeup
clip of a tiny portion of the shot near his hand: Scale that back
to the full frame, and you'll have some idea of the level of detail
we're talking about! Foveon has been keeping technical details
very close to its vest, but we're hopeful to have more information
after we interview Carver Mead Wednesday. (Article probably to
be posted Thursday.) (A personal note by Dave This will be an
interesting reunion of sorts for me -- Carver taught me digital
electronics at Caltech over 20 years ago, and Richard Lyon was
in the class two years ahead of mine. - Carver is truly a great
original thinker, and one of the prime movers of the semiconductor
industry, and Dick is possibly the brightest engineer I've known
in a long and varied career! Whatever they've come up with in
the Foveon promises to raise the technology of the whole area
a solid notch or two!)
 |
 |
|
|
|
Stay tuned for a second PPA update, probably on Thursday!
