Re: 3D SPEX/neotek flamed

[P3D <URBANIC@xxxxxxxxxxxxxx>]

Larry Berlin rants (caps his):
>WE THE PEOPLE HAVE TWO EYES. WE WANT TO USE BOTH OF THEM IN VIEWING STEREO
>IMAGES. WE HAVE BEEN PROMISED STEREO IMAGE ACCESS THROUGH TECHNOLOGY SINCE
>BEFORE WE WERE BORN, YET IT HAS TAKEN TILL NOW FOR IT TO COME ANYWHERE CLOSE
>TO THE PROMISES. WE ARE NOT WILLING THAT ANY COMPANY SHOULD FURTHER DEPRIVE
>OUR INALIANBLE RIGHTS TO MAKE AND USE STEREOSCOPIC IMAGES BY THE RESTRICTION
>OF AN ENTIRE MARKET THROUGH PROPRIETARY FORMATS AND PROCEDURES. YOU DON'T
>OWN MY TWO EYES, NOR DO YOU OWN MY ABILITY TO SEE WITH BOTH OF THEM.

Larry, let me put this in a way you can understand: GET A GRASP ON REALITY!
I got involved in a discussion of technical points between competing LCS
technologies, and next thing I know I have some kook ranting on.  I did not
make the above paragraph up; it is quoted ver batim, and (except as some
bizzare parody of the Declaration of Independance) I don't find it
appropriate in any context.  I do find it very hard to have a productive
dialog in such an atmosphere.

Nevertheless, lest some truly interested reader believes that the words that
he put into my mouth are indeed mine, I feel compelled to respond to this
post before I abandon this discussion.  Furthermore, given Larry's
vehemence, I don't really feel oblidged to pretend that he knows what he
is talking about.  Those of you that would prefer this forum to be ruled
by civilty and mutual respect, as would I, may want to skip my response.

In response to Andrew's (we'll leave out last names in this fiasco)
informative post I wrote:

>thanks for your thoughtful insight.  I am glad to see that people are actually
>thinking about this stuff instead of just reacting to what vendors promote.

to which Larry accuses me with:
>(John) then proceeds to further *vendor promote*...

Let me state this clearly for you Larry:  While I am an unabashed proponant
of Synch-Doubling, Neotek does not sell to the stereo enthusiast market,
nor are we interested.  Due somewhat to comsumer attitides such as yours, but
moreso to simple economics, we do not have any illusions about making money
from hobbyists.  To put it simply, the educational software market is very,
very big compared to the tiny, tiny, fragmented hobbyist crowd.  No company
has ever been succesful trying to please such as yourself, and more than a
few have gone very bankrupt.  The only two big companies that have been
around a while (Stereographics and 3DTV Corp) have only done so because they
sell expensive quality systems to groups that couldn't care less about your
Pollyanna stereo-for-everyman plans.

>While 1024 x 768 x 120Hz would be perhaps ideal, there is nothing
>wrong with 1024 x 768 x 85 Hz!

Yes there is.  IT FLICKERS.  You may, as you state, "disagree with your basic
premise that nothing slower than 120 Hz is acceptable" all you want, but it
doesn't change the basic physics.  You need at _least_ 60Hz to each eye to
get a normal flicker-free display. This is a simple fact confirmed by every
monitor design in the world.  You can not buy a computer display of less than
60Hz, nor will any modern video card allow you to select such a mode.  As a
matter of fact, almost all modern designs figure you really should start at
72Hz as 60Hz can subtly flicker under certain conditions.  Now _every_ LCS
system (including whatever you like, Larry) gives only every other frame to
each eye.  So, to maintain 60Hz to each eye you need 60Hz * 2 = 120Hz !
If you want to believe that 85Hz (equivalent to 42.5 Hz to each eye!) is
acceptable, fine, but I think it is very misleading to others that don't
have such systems and I hope they can see it must be otherwise.

Considerably later in your post (after you state there is "nothing wrong with
85Hz") you announce that 100Hz is "almost flicker free", and that 85Hz
"isn't too bad".  Perhaps you are disingenously arranging your statements
so that the facts don't interfere with your attempts to discredit my simple
statements. Perhaps not. But, in the event that you were trying to
make a case that some level of flicker is acceptable, I must state
strongly that for general purpose use (which you so passionatly advocate)
YOU ARE WRONG.  Besides that fact that even casual computer users of more
than a few years can recall when monitors went from 60Hz per eye to 72Hz
to reduce eyestrain, every serious, independant research paper to examine
the issues of eyestrain, eye fatigue and nausea as produced by stereoscopic
display systems comes to the same conclusion: any perceptible flicker, no
matter how subtle, will cause eye fatigue and potentially nasea.  This is
no surprise as it corresponds exactly to the effects seen with normal 2D
displays.  Bottom line: Neotek has  displays sitting in very prestiguous
medical schools where _all_ the students _must_ be able to view our
system for 2-3 hours at a time.  Let me know of any protracted viewing
being done with a commercial 85Hz system, Larry.

I made a statement that the only DVD-stereo system you can buy is a synch-
doubling one (ours, and I am proud of it).  Larry responded:

> I've heard better stories about stereo with DVD than you are telling so
> something isn't complete here. I see no problem with page flipping stereo
> with DVD, providing a page flipping stereo system is installed on a computer.

I am calling you here, Larry.  I won't try to refute ridiculous rumors
that don't make sense, but the sound of "something isn't complete here"
makes it sound like _I_ am the one making unfounded statements.
We have a working DVD product for sale; the only stereo-DVD product for
sale.  Tell me where is, or who has, a page-flipping DVD system.  You
"see no problem with page-flipping stereo with DVD" because you either are
trying hard not to see one, or because you have no idea how DVD works.
THERE IS NO SUCH SYSTEM.  It would be a technical nightmare to even attempt.
Now I suppose you'll write something about how someone should/could attempt
it.  Fine, but no one has, so don't make up facts, especially ones that
imply _I'm_ not being honest.

> I intend to use DVD with page flipping at some point. I can even do it with
> CD-ROMs so why not with DVD, which I understood to contain advanced
> provisions for stereo application.

I wished that you had asked this question (a good one) in a less contentious
context.  The answer is simply that DVD plays back entirely through hardware.
Most computer systems use a chroma-key overlay technique, just like your
local weatherman does to get his backdrop.  This leaves very little room
for any software hacks of the type that provide page-flipping.  Our system,
being hardware-based, simply slips right into the video stream, and doesn't
interfere with any of the sophisticated real-time hardware.  It works very
well on every DVD player released in this country thus far.  The proposed,
and even at one point specified, stereo standard was never implemented.
Neotek very much wishes it was.  It would allow us to sell our software
(a high profit item) without hardware (much lower profit).  However, once
again the world has ignored the stereo community like it doesn't exist.
That may sound harsh, but it is true and you better understand it if you
really want to do anything constructive, besides complain and wish.

>As to windows not supporting page flipping, so it is...
>development has been way off the ideal mark by the simplest of evaluations.
>It will be surpassed eventually and whether MS decides to wake up and do the
>job itself, or a competitor decides to end their ridiculous posturing by
>providing something far better, remains to be seen.

Great.  Another guy waiting for Microsoft to "get theirs".  This phenomona
is by no means comstrained to the stereo community.  Let me just say that
conventional wisdom recommends that you not hold your breath.

Randomly organized remarks, which I address in turn, follow.  Sorry for the
flow, but I didn't want to get accused of editing the original order.

>******  Your sync doubling on a screen that is 1024 x 768 is really 512 x
>768.

I think you mean 1024x386.  Regardless, if that doesn't suit you, simply
tell Windows to increase your resolution.  With synch-doubling, you can
pick whatever you wish.  We picked 1024 x 768 x 16bits at 120Hz as a minimum
because we felt that was where photorealistic quality starts.  If you
don't like that heavily reseached choice, blame us, not sych-doubling.
It allows you to use anything your graphics card can put out.

Larry mentiones several page-flipping resolutions he prefers:
>Each of the resolutions I mention here occupy the full monitor screen. The
>cost of equipment, assuming one already has a good computer and monitor, is
>about the same as what you want for the sync doubling.

Your definition of "good computer and monitor" is not clear to me.  We need
to actually put a minimum spec on the side of a box that guarantees a customer
a level of quality.  Basically, our minimum spec amounts to "2MB Video Card"
on a Windows machine.  Any customer using a 4-5 year old computer or monitor
will get the very same quality we see on our own systems, guaranteed.  BTW, as
a typical 17" monitor of the type that can support the modes you prefer costs
about $700 (and we'll forget about the computer), I think the cost to upgrade
to your proposed system is much greater than our system ($295 for the basics).

>Now, if you want to talk about *inate limitations* lets talk about half
>height images.

Where did this term "half-height" come from?  Electronic images do not have
a "height", they have a resolution.  And if you mean something like "half-
vertical resolution", then I think your term also encompases the horrible
interlaced systems as well.  If you wish to communicate with professionals
in this field, I suggest the standard term "sync-doubling".

> Have you written your software to be able to read a full height image into
> your sync doubling method?

Of course.  Almost all high-end and professional work is done as stereo-
pairs.  From 100 year-old cards to 50's 3D movies to IMAX to digital
cameras, the stero pair is the lingua franca of stereoscopy; at least as
source material.  It often gets converted to some horrid format to accomodate
a cheap display system, but as long as the original pairs are available, the
work can be shown at its highest quality.  Our products assume that that is
where you are starting.

Me:
>>> DON'T USE LOSSY COMPRESSION FOR 3D.

Larry:
>For starters, JPG images do induce a degree of loss. However in using normal
>amounts of compression (25% to 65%) the loss is VIRTUALLY NEGLIGABLE

Define "virtually neglibable".  Coming from a guy that believes 85Hz is a
fine way to view stereo, I am sure your definition would differ from mine.
Actually, I wouldn't even like to try to define such a thing, it sounds too
much like a marketing blurb.  I prefer that our customers simply encounter
_no_ loss, and so do they.

>Or switch to using PNS which is the stereo form of PNG. PNG
>offers a no-loss compression that is royalty free, but don't expect a lot of
>compression. The image files are still rather large, so this is for some
>amount of compression, without losing any image quality. Ideal for stereo,
>but still a bit large for general internet usage. It beats NEO in that it's
>non-proprietary and commonly available to anyone and defeats the argument
>about loss compression!!!

What is your problem with me, Larry?  I make a simple statement: "Don't
use lossy compression", then you go on a tirade and then finally say to use
PNG because it offers no-loss compression, which is exactly what I suggested.
Are you just trying to pick a fight?  If so, don't worry about "beating"
NEO, I have explicity stated 2 times thus far that I am not proposing, nor
is it available, as a public stereo standard.

>Now as to vertical resolution being less important... Hogwash. Yes, it is
>horizontal detail that is critical to stereo viewing, but the eye does
>resolve vertical resolution in essentially the same degree as horizontal
>resolution.

Did you try my experiment?  Many did and apparently found it interesting.
Do you disagree with my results?  I think that it clearly demonstrates that
the eye resolves reolution better in the horizontal direction.  Furthermore,
any book on visual physiology concurs.  I guess it is all "Hogwash".  When
and if you do agree, I hope you can see where it modifies your view on
stereo formats considerably.  Call your favorite vendor and see if they are
aware of this simple fact.

> What you are trying to say is that since stereo uses mostly horizontal
> detail, it's somehow OK to throw out 50% of the vertical resolution.

No, I am saying that you can pick any vertical resolution that you think is
sufficient for your purposes merely by setting your video card to higher
resolution.  Sync-doubling allows you to do that.  If you want a subtler
appreciation of the benefits of synch-doubing, and you are one of the many
that accept that the eye has better horizontal resolution, then you may
want to note that for a given file-size, a sync-doubled over-under image
makes more efficient use of the pixels.  As Larry does not believe that the
eye has superior horizontal resolution, I am sure he will disagree.

>The line-pair cards is not an accurate representation of this factor.

If you are referring to my experiment, I think it is clear to most that it
proved its point: the eye has better horizontal resolution than vertical.
I've just noted that sync-doubling fortuitously exploits that.  You
disagree with the results apparently, so further argument along that line
is pointless.

>CARDBOARDING is far and away more caused by a narrow stereo base than by JPG
>compression. It is not even closely related to JPG compression.

I can do some simple, back-of-the-envolope, math that demonstrates that
foveal resolution far exceeds (almost 2 orders of magnitude) JPEG resolutions
for anything 50% compressing a 1 pixel vertical edge of a 640x480 pixel
image displayed on a 17" monitor at 30" viewing distance.  It's basic
trig plus the knowledge that a 50% compression on a vertical 1:5 pixel lined
image results in a 3 pixel (2.6 mean extension rounded to 2) line.  A good
test case if one wants to actually quantify things.  Given this awful loss
off parallax information, you get considerable cardboarding.  Stereo-base
width can also certainly effect things, but that is basically competant
photograpahy.  I take that for granted, although I do appreciate the
tremendous skill that it takes.  If anyone doubts my logic, a simple
experiment would be to display an image that you believe has no cardboarding.
Increase the lossy compression and see if such a thing appears.  Let us know
your results.

> make your images in a no-loss format (TIF, BMP, TGA, PNG, PCX  plus many
> others. Or for stereocopic formats use BMS, or PNS) hould it ever be needed.
> OF COURSE it's stupid to expect to re-edit a JPG image and get reasonable
> quality. Who even suggested such a thing?

Well, it seems to me that if you promote JPS/JPG as a "standard", then you
are promoting it as useful for trading or archiving.  If you have to always
send along a TIF, BMP or some other non-lossy image for any future editing
or re-registration, then what is the purpose, casual Web browsing only?
I prefer a format that allows me to re-register, edit, window, archive,
trade, post, etc. without needing some hidden "original" that actually
retains quality.

Me:
>We, however, made a very conscious decision that if setting
>some mimimum standards meant we can't provide "browsing" quality images,
>then that was what we would do.  Two years ago that meant we didn't even
>intend to offer a Netscape "Helper" application. 

Larry:
>That's not a decision for image quality, that's a dumb marketing
>mistake. You ingored 80% of the full marketing picture and made a bad
>decision. That's all that such a thing represents. It's kind of like MS
>missing out on the internet idea till later when someone proves it was more
>than what they had thought ...

First, I think that by definition it _is_ a decision for image quality.
Second, our "dumb" marketing mistake has made us a few dollars in a very
stable educational market, with a future that we have reason to be optimistic
about.  I don't want to name names, but I suspect a few of the companies
you admire for their marketing prowress and paying attention to the hobbyist
market are either dead, or had recent large layoffs (for which Neotek is truly
sad as they are not competing in our market, but do develop enthusiasm for
stereoscopy to the benefit of us all).  Please provide an example of a
stereoscopy company whose marketing plan you admire.  And I am sure that
Bill Gates would like to hear if you are available to help save his
company from all of their horrible marketing mistakes as well, especially
their Internet problem that you mention.

>*****  No one said anything about dumbing down image quality either. The
>intent of JPS was for internet exchange of stereo images since that's likely
>to be the number one mode of access in the forseeable future.

The Internet is likely to be the the number one mode for the exchange of 
images?  That will be news to all the people that actually consider CD
to be the only viable commercial way to distribute high-quality electronic
imagery.  For example, the education market.  Tell Kodak you'd like to
retrieve your PhotoCD over the net next time you drop off a roll of film.
The net will eventually catch up to the bandwidths necessary to support the
sizes of data you see on CD.  Of coures by then, everyone will be creating
high-res video on DVD (we already are), so the Net will be hopelessly slow
for that.  The Net is bound to be one step behind whatever storage media
is on a desk top machine for some time to come.  We choose to aim at that
higher level.  You are satisfied viewing images that are Web'able in size.
To each his own, just don't confuse the capability of the two.

> I have dozens, maybe hundreds of stereo images that are over 1 MB each.

We have, as do many of our customers, many thousands of such images.  Most
much greater than 1MB.  Don't confuse your needs with institutional users.

>you will have to release the format to everyone's use. (of course you could
>do that tomorrow if you decided to do so. Why wait another one or two
>years??!!!)

We consider our format a valuable asset that took considerable time and money
to develop.  We put a lot of programmer hours, customer feedback and painful
rewriting into developing our format so that our paying customers can get
better performance, capability and convenience than what is available
publically.  So, why exactly should we release it so that people like you,
who believe that JPS is superior (or at least a great idea), can particpate
in future development and extension?  If you were one of our customers who
had in essence paid for the development of such a system, would you feel
that is equitable?  And in general, do you believe that anyone that invests
in devloping a superior algorithm or technology should release it to the
public for free?

>Then the question is, *why would I bother with yet another format, since
>long before NEO becomes usable, there are so many others that
>aren't proprietary, yet serve the purpose very well? 

I think that there are many users that believe NEO is currently useful.  But,
to answer your question, as you are happy with JPS, there is no reason why you
would find NEO useful.

>The decision to use a proprietary format for any stereoscopic format is one
>that is totally self serving with no thought at all to the general stereo
>market place.

Finally, a clear statement of utter truth.  We did indeed develop a
propriatary format in order to serve paying customers that were not
part of the stereo community (many of whom don't know that they now are!).
And while we took a good look at the existing stereo market place, we
decided that compromising to accomodate the trendy companies of the time
(most of whom are now defunct) would be unwise.  So we only concerned
ourselves with the very high-end techniques and products, and developed the
best format for _our_ purposes (not Larry's or anyone else's) that we could.
So, you are right on the money for once.

>General formats for everyone to use work exactly opposite. They encourage
>stereoscopic usage overall without regards to the viewing system. This
>increases the overall marketplace and improves the selling environment for
>any viewing system rather than just one brand.

And has resulted in the large number of high-quality stereo displays and
successful suppliers that have so captured the average consumer's interest.
(Extreme sarcasm).

>*****  Yet you admit that double sync maintains the same bandwidth of image,
>making your full screen images equivalent to a 640 x 480 image in the end. 

Your math is way wrong, but let's say that it isn't.  You would still have a
FLICKER-FREE image that works on any old computer.

Me:
>You can go directly to the .neo feature page at: http://neotek.com/neo.htm
>However, we are not promoting this as any kind of standard.  Indeed it is
>proprietary, and we spent a lot of time and money developing something from
>scratch targeted directly at stereoscopy. 

Larry:
>And ignoring the marketplace as if it didnt' exist... NEO as proprietary
>does no good.... It drastically narrows your own marketing reach. 

OK, I'll bite.  What marketplace do you keep referring to?  The large and
rich one that has supported all those other defunct stereo companies.

>As long as you are comfortable doing stereo just for yourself and your
>customers directly, and no interest in providing your customers with wider
>access to other stereo images, fine stay with your closed door policy.

You got that exactly backwards.  Our customers have access to almost any
other format that is out there (JPG too).  They can turn any image source
into a NEO file.  We just aren't interested in promoting our format amongst
others that aren't our customers.

> This is hogwash!!!  JPS will do more for stereoscopy than you have
>ever dreamed was possible.

You're absolutely right.  All the ailing LCS stereo market needs is a lossy
format that gets used primarily for low-resolution Web pages.  That will
sure impress all the non-believers that believe stereoscopy is just a
novelty gimmick.  I sincerely hope you are right, I just don't think there
is any chance you are.

>Are you really going to limit your customers by not providing such an
>access for their supposedly high quality systems?

Once again, our customers can import any type of image they want.  They
just get to use NEO to trade, store, manipulate and author (a whole subject
I've never raised) images.

>If I owned your system I would not want my viewing experience limited by
>your aversion to compression schemes.

First of all, it is only LOSSY schemes that we abhor, and second, you would
probably be a very disatisfied Neotek customer for a number of reasons.

>I want to see and work with all the stereo images. AND I don't
>want to have to buy every viewing system on the market in order to do so.

Good for you.  I want a car that is good off road, very fast around a race
track and can pull a Winabago.  The differance is, I don't get mad and go
on a tirade because GM won't make one and then give it away for free.

>You would prefer that NO INDIVIDUAL PERSON has easy access to any
>kind of stereo unless you sold them the system for doing so.

Where did you get that conclusion from?  We would prefer that every teacher
in the world buy our system for teaching.  We hope that an entertainment
market of general consumers finds our DVD system worth the price when we
start to release some seriously mainstream material that is in the conversion
pipe.  Outside of that, we don't have much of an immediate interest, and I
personally would like to see as many people involved in stereo in its many
fascinating forms as possible.

>If I consider my options, I'll opt for JPS over NEO any day of the year,

Good, so why are you accusing me of taking away your "inalienable" rights.

>Share your ideas if you will, but I for one won't let you get away with
>misleading sales double talk that belongs on the car sales lot,

Let me state for the umpteenth time that I am not selling anything.  Our
system is not intended for, nor marketed at, the hobbyist market.  A quick
scan of our Web pages would show that.  The one very obscure link that we
have to any hobbyist material is because we have a distributor that we
are very happy to accomodate who believes that there is some kind of
market around the hobbyist crowd.  I think history proves otherwise, but
it's his investment.  And, while we have many hobbyist that are partners
and collaborators in material development and publishing, we work with them
to make profitable products for the non-stereo world.

Well, I can see that the break after the late-summer academic sales rush
has given me way too much time to respond to this post.  Also, that crazy
first paragraph from got me on a bender. At any rate, thanks for your
patience, please don't hesitate to e-mail, and, to your undoubted relief,
I probably won't be making any more posts for a while.

John


------------------------------