Well, just had my first go at voting (gardens) and here, for what they are
worth, are my thoughts about voting in general.

Technical - I didn't vote on technical, since I only really know POV. I just
gave everyone 10.

It would have been nice to be able to give a neutral vote (ie. "N" meaning
"I give this the average score for this particular image" - essentially
saying that you trust the judgement of other voters) so that I could have
given a technical score to the pov pictures. I don't know enough about stats
to know if this would work.

Alternatively, does anyone have any general info on the different packages
that could be assembled on a web site to help guide voters? I don't mind
doing the page if people send me anything useful.

Artistic - as well as letting my own aethetic judgment have an influence on
the score, I also tried to consider how well the artist had met their own
intentions.

Concept - low scores for off-topic, higher scores for an interesting use of
the topic, or for an ambitious scene.

General - I didn't read the text until after I'd voted, since IMHO it should
be irrelevant. Only once did reading the text change my vote, and in this
instance by bringing it down (I won't say which pic). I didn't penalise for
pics that looked 2D, etc, as long as this was appropriate to the subject or
tone of the pic - indeed a favourite entry couldn't look less like a
raytrace.

P.S. Looking forward to ruins already!

Post a reply to this message

>>Technical - I didn't vote on technical, since I only really know POV. I
just
>>gave everyone 10.

>>It would have been nice to be able to give a neutral vote (ie. "N" meaning
>>"I give this the average score for this particular image" - essentially
>>saying that you trust the judgement of other voters) so that I could have
>>given a technical score to the pov pictures. I don't know enough about
stats
>>to know if this would work.

Technical in my voting has to do with the level I perceive the person
rendering the picture has aspired to in the development and commission of
their work.  For example, a bryce picture with only a couple of objects and
a very realistic texture can be a pretty impressive picture.  However, on a
technical level, this ranks very low on my scale.

Making a technical call requires you to read the text file associated with a
picture.  It is so important for people to submit this with their pictures
from the aspect of scoring this item.  The description of HOW someone did a
particular portion of their work can determine how good a score they deserve
for this factor.

Sometimes, simply looking at a picture doesn't give you a good view on the
level of technical know-how that went into producing that picture.  I did a
large amount of work on my 1stabomb submission for the history round.  I
included hundreds of human figures that were composed of strictly POV
primatives but lost out because they weren't poser figures or human-like
enough for the viewers.  If you examine the zip file for that submission,
you will find figures for each of the officer components of each unit and
several different figures for each unit type.  While these figures looked
funny from close up, the farther they are away from the camera, the more
acceptable the primative detail is.   The other problem people had with this
submission was that since they DIDN'T read the text file, they DIDN'T get
the joke.

If you don't feel qualified to vote on a particular venue, giving a picture
a 10 doesn't aid in separating the good ones from the bad ones.  You instead
contribute to making the entire group more average.  For example, giving
everyone a 10, if there were 3 votes, 20,10,20, the picture ends up with an
average of 16.67, you've effectively dropped that picture by almost 4
points.  On the other hand, on a particularly bad picture, 10,1,1, you move
it from a 1 up to a 4.   Where both pictures are already getting average
scores, you will move the score less, but both will tend to be more average.
Better that you read the text file and decide if the person has made a
technical breakthrough or done nothing exceptional and give 1's and 20's
instead.  Your basis for this scoring is the level of expertise compared to
your own knowledge that the person has given you.

Unfortunately, in your case, what would be beneficial for everyone is to be
able to not score a picture for a particular area.  Then you don't change
the relative score for all the other votes.

Post a reply to this message

Tom,

Well, it has been about three years since I last voted, so it was almost
like the first time. ;)

My practice in voting is 1) a brief look at every image; 2)spend some time
on each image; 3) read that image's text file; 4) vote on that image, etc.
... let it mull about for a day or two and then review and see if my mind
has changed any (this time, my only changes were to up the ratings I had
given my *least* favorites -- felt like I had been too harsh on a couple).

I don't think the text files are irrelevant at all when voting on tech -- so
many different ways to achieve one render, some of them hard and some of
them easy.  Also, as for the "artistic" I don't mind at all when I need the
text file to "understand" the picture (although, while I'm not particularly
concerned with "understanding", I have discovered that sometimes there is
enlightenment to be found).

As to tools for voting -- if you're a windows user, I strongly encourage you
to download http://www.halcyon.com/wordsltd/pov/winvote.htm -- my first time
using it, but it was a dream... no more toggling back and forth between text
editor, image viewer, other text editor with the text file.

All-in-all, I was impressed by the quality of work in this round... I'm
still torn between two images as to which is my "favorite" and there are
several right behind those that are really nice pieces of work.

Claire

Post a reply to this message

Robert J Becraft wrote in message <3873ace1@news.povray.org>...
>If you don't feel qualified to vote on a particular venue, giving a picture
>a 10 doesn't aid in separating the good ones from the bad ones.  You
instead
>contribute to making the entire group more average.  For example, giving
>everyone a 10, if there were 3 votes, 20,10,20, the picture ends up with an
>average of 16.67, you've effectively dropped that picture by almost 4
>points.  On the other hand, on a particularly bad picture, 10,1,1, you move
>it from a 1 up to a 4.   Where both pictures are already getting average
>scores, you will move the score less, but both will tend to be more
average.


However, giving everyone a 10 will not change the *relative* standings for a
category.  Additionally, most pictures already score around a 10 in each
catagory, and with an estimated 40 voters for this round, giving each image
a 10 will change the final score by no more than 0.25 points.

Mark

Post a reply to this message

jivvy <web### [at] freegraphicscom> wrote in message
news:38741dc8@news.povray.org...
> As to tools for voting -- if you're a windows user, I strongly encourage
you
> to download http://www.halcyon.com/wordsltd/pov/winvote.htm -- my first
time
> using it, but it was a dream... no more toggling back and forth between
text
> editor, image viewer, other text editor with the text file.
>
Found it - about 5 minutes after I'd written a whole bunch of perl to
display text and pic in html and written all my scores down - aarghh!

I agree re quality of round - very high and some lovely things.

Post a reply to this message

Robert J Becraft <cas### [at] aolcom> wrote in message
news:<3873ace1@news.povray.org>...
>> lots of stuff!

With regard to technical scores, I largely agree with what you say. The
problem from my point of view is two-fold.

In the first place, quite a high degree of intimacy with each prog is
required to give accurate technical scores. For example, it is not
necesserily enough to know what poser can do, it would also be nice to be
able to recognise "stock" figures from those specifically created by the
artist.

Secondly, I wonder if the whole principle of technical scores for such a
diverse range of applications is appropriate?

Certainly, to my mind, one solution would be to allow technical votes to be
skipped (adding on the average score for that user instead), although this
might complicate things with obscure progs, since one vote could be very
significant.

With regard to the text files (when not judging technical merit), I guess
this is matter of personal opinion when judging artistic merit and concept.
Certainly when I was unsure of the relevance of a particular image, I
checked the text, but it very rarely made any difference. Perhaps, in a
different topic, it could be more essential. For instance, the subject
"folly" illustrated with a man falling from the sky surrounded by feathers
would require that you know the story of icarus (spelling?). However, a man
sawing a tree branch while sitting on the wrong half would not.

To my mind, a garden, abstract or not, should in some way be a garden. I
will concede that I might have missed something that could have scored a few
points - for example if someone had illustrated "autumn, season of mists and
mellow fruitfullness" then I would have liked to have read the text and
scored appropriately. On the other hand, such a scene, if successful, would
also stand on it's own merit.

On the whole I find the text files are either irrelevant to my enjoyment of
a scene, or are excrutiating and convoluted justifications of off topic
subjects.

P.S. I took another look at your history entry - a perfect example of the
former. The subject is history, which your picture certainly is - all the
text file does is explain the rather confusing title. If it had just been
called "Romans" or something, it would still have been very on-topic and
wouldn't have needed the text explanation (or am I missing something?).

P.P.S. How should pictures be scored for technical where renderer, tools and
source are not given? 0?

Post a reply to this message

Tom Melly wrote:

> In the first place, quite a high degree of intimacy with each prog is
> required to give accurate technical scores. For example, it is not
> necesserily enough to know what poser can do, it would also be nice to be
> able to recognise "stock" figures from those specifically created by the
> artist.

Looking at the POV and 3DS renders (75% of the entries in this round), it looks
like in both cases the artists covered the whole range of the prog's
possibilities, from the newbie to the seasoned artist, so that one can have a
quick idea of the relative difficulties involved in doing the picture. For the
lesser used programmes, I'll probably have to look at user galleries.

Generally speaking, I tend to think, however, that one should do the best with
the tools he/she has and simply not use them if they don't live up their
expectations. For instance, if you want to do a human figure without Poser, this
is all right if it is as good as a Poser output, or if it different in a way
that is artistically significant (like some of the blobman pictures posted in
p.b.i) : if the artist does a bad Poser wannabe with primitives then I think
he/she should not have tried it in the first place (of course you can make a
very impressive human figure with primitives). In the case of Poser, if the
Poser model looks bad then I think he/she shouldn't have used Poser that way :
for instance, I never used close-ups of Poser models until version 4 because I
found them ugly. In that way, the technical judgement may be disconnected from
your knowledge of a tool's abilities.

> P.P.S. How should pictures be scored for technical where renderer, tools and
> source are not given? 0?

Perhaps write to the author and post the author's answers here... I guess that
some participants don't mention the tools used because they must be obvious
context-wise, particularly for professional tools like 3DS. It is also difficult
sometimes to keep track of all the external help (I begin to have trouble
crediting the authors of the poser morphs targets I've used, for instance).
Posting sources may be a problem for many acceptable reasons. Not mentioning the
renderer is a problem though, but it happened only in 3 cases.

G.

Post a reply to this message

Gilles Tran <tra### [at] inapginrafr> wrote in message
news:3874A205.42DEC387@inapg.inra.fr...

>....
> Looking at the POV and 3DS renders (75% of the entries in this round), it
looks
> like in both cases the artists covered the whole range of the prog's
> possibilities, from the newbie to the seasoned artist, so that one can
have a
> quick idea of the relative difficulties involved in doing the picture. For
the
> lesser used programmes, I'll probably have to look at user galleries.
>

A good idea/point - next time I'll take the plunge.

Post a reply to this message

I think that when an artist's scores for other entries are recorded, scores for
that artist's own entry of

            mean - 0.1 std dev

should be recorded for their own score!


Tom Melly wrote:

> Well, just had my first go at voting (gardens) and here, for what they are
> worth, are my thoughts about voting in general.
>
> Technical - I didn't vote on technical, since I only really know POV. I just
> gave everyone 10.
>
> It would have been nice to be able to give a neutral vote (ie. "N" meaning
> "I give this the average score for this particular image" - essentially
> saying that you trust the judgement of other voters) so that I could have
> given a technical score to the pov pictures. I don't know enough about stats
> to know if this would work.
>
> Alternatively, does anyone have any general info on the different packages
> that could be assembled on a web site to help guide voters? I don't mind
> doing the page if people send me anything useful.
>
> Artistic - as well as letting my own aethetic judgment have an influence on
> the score, I also tried to consider how well the artist had met their own
> intentions.
>
> Concept - low scores for off-topic, higher scores for an interesting use of
> the topic, or for an ambitious scene.
>
> General - I didn't read the text until after I'd voted, since IMHO it should
> be irrelevant. Only once did reading the text change my vote, and in this
> instance by bringing it down (I won't say which pic). I didn't penalise for
> pics that looked 2D, etc, as long as this was appropriate to the subject or
> tone of the pic - indeed a favourite entry couldn't look less like a
> raytrace.
>
> P.S. Looking forward to ruins already!

Post a reply to this message