Guidelines for Submitters/Reviewers/Judges

rob77
21 Feb, 2018 11:41
In the last 6-7 days we went through some discussions about how to improve the current IOTD/TP system.
The as-is process (as far as I know) doesn't foresee, basically, the usage of some "official" guidelines for promoting images to the upper tiers.

People involved in the selection (again, as far as I know) are putting thumbs up to images based on their tastes, knowledge, feelings and surely on technical criteria.
IMHO, exactly this latter point could be object of improvements if we try to write down some guidelines - all together - for the staff.
Crystal clear, we are all humans after all. I am not expecting to take down to 0 the percentage of too denoisy images or out of focus or … whatever as TP or even IOTD.
But maybe in this way we can have some more objetive drivers to lean on!

Any suggestions?

Thanks and cheers
DavideCoverta
21 Feb, 2018 12:22
Roberto Colombari
People involved in the selection (again, as far as I know) are putting thumbs up to images based on their tastes
This is one of the first things that must be avoid, from my point of view…
Jooshs
21 Feb, 2018 14:10
If an image is to be worthy of an IOTD, I think there can be a set of minimum criteria that avoids hampering subjective tastes too much. I’d suggest the following.

- no stacking artifacts on the edges
- no double stars in the final stacks from misalignment
- minimum of filters and gear used and detailed capture times per filter
- no obvious mottling or rice stars from over use of noise reduction
- only one IOTD per data set per season

One other point… I wonder if it makes sense to limit the frequency of certain targets being selected or moved to TP. For example, I count roughly 5-6 M 106’s, Rosettes, M42’s etc in the last couple of weeks. This is probably a contentious issue and not everyone would agree, but in my opinion, these are very bright targets that are big and easy to make look impressive on an initial glance. I’d think there needs to be something unique about them to make them stand out a bit. Probably too subjective to really implement anything, but just a thought to consider. I’d much rather see an interesting target that’s not quite as striking garner some attention than a bunch more M 51’s.
rob77
21 Feb, 2018 14:26
For simplicity sake, I would just limit here suggestions for technical guidelines.
I think there is another thread in which we can discuss the frequency and targets.

Anyway, I agree with you.
So:
- no stacking artifacts on the edges
- no double stars in the final stacks from misalignment
- minimum of filters and gear used and detailed capture times per filter

I'd add:
- check the resolution at which the image is uploaded (not merely the size in pixels; the image could be cropped but at full resolution at the same time). If you see that the gear is a FSQ106 + SBIG8300, it's quite fast to have an idea of the full resolution the image should be, approx. 2.1"/px (Resolution = (CCD Pixel Size / Telescope Focal Length ) * 206). If the nice image you want to promote is presented at, let's say, up to 70% of the original resolution is ok, otherwise not.

- noise reduction: IMHO denoise must not be invasive. It's preferable a little bit noisy image than an image completely drown into denoise routines

Cheers
Edited 21 Feb, 2018 14:27
Jooshs
21 Feb, 2018 14:29
What if someone is doing a 9 panel mosaic and it is unrealistic to present anymore than 25% of the actual resolution because of limits on file upload sizes?

I think there has to be some faith in the judges to open the full resolution image and inspect the quality to make sure it doesn't just look like a good thumbnail.  Thoughts?
rob77
21 Feb, 2018 14:34
I am doing lots of mosaics. Very big ones.
I generally try to upload them, at least, at 50% of the original size. It is just a matter of waiting a little bit more for the upload smile

Up to a 3x3 (or even 4x4) mosaic I don't see so much issues in uploading a full resolution or a 30% reduced version. Isn't it ?

Cheers
JanD
21 Feb, 2018 14:36
I think you should be careful with the excessive application of rules and regulations.

Roberto Colombari
People involved in the selection (again, as far as I know) are putting thumbs up to images based on their tastes …

Would be sad if the taste would go missing. A picture can be technically perfect but sterile. I am not sure if the majority of the people would find it necessary to look with a magnifying glass at picture corners or pixels.

Just my 2 cents.
Edited 21 Feb, 2018 14:39
rob77
21 Feb, 2018 14:38
Never said that. Taste must be part of the game jointly with some technical guidelines, IMHO
DavideCoverta
21 Feb, 2018 14:44
JanD
I think you should be careful with the excessive application of rules and regulations.
Roberto Colombari
People involved in the selection (again, as far as I know) are putting thumbs up to images based on their tastes …

Would be sad if the taste would go missing. A picture can be technically perfect but sterile. I am not sure if the majority of the people would find it necessary to look with a magnifying glass at picture corners or pixels.

Just my 2 cents.
Yes, you're right.
But we're just trying to avoid that an Image badly processed or imaged (out of focus, for example) is considered TP just because there is "something I like… for example the avatar of the processor"…  smile
JanD
21 Feb, 2018 14:50
Davide Coverta
Yes, you're right.But we're just trying to avoid that an Image badly processed or imaged (out of focus, for example) is considered TP just because there is "something I like… for example the avatar of the processor"…
I am just wondering, if you click through the last weeks of the IOTD and the Top pick gallery, which ones are the ones where you think "this one did not deserve it", or is it rather a picture which did not achieve to be a IOTD or Top pick which bothers you?
DavideCoverta
21 Feb, 2018 14:54
JanD
Davide Coverta
Yes, you're right.But we're just trying to avoid that an Image badly processed or imaged (out of focus, for example) is considered TP just because there is "something I like… for example the avatar of the processor"…
I am just wondering, if you click through the last weeks of the IOTD and the Top pick gallery, which ones are the ones where you think "this one did not deserve it", or is it rather a picture which did not achieve to be a IOTD or Top pick which bothers you?
Jan, I don't need to be unrespectfull versus somebody else, since we're here just to have fun…
If you search by yourself, Images not deserved to be TP are not a rarity! Look and after reply if you are sure that all TP are images that we should consider as a reference!
BR
rob77
21 Feb, 2018 14:55
Jan,
I think that here is not the right place to put the finger against an image or an astroimager.
Maybe you can clarify yourselves via PM smile .

Cheers
Edited 21 Feb, 2018 14:57
DavideCoverta
21 Feb, 2018 14:58
Roberto Colombari
Jan,I think that here is not the place to put the finger against an image or an astroimager.
Maybe you can clarify yourselves via PM.

Cheers
Is not needed… Jan can find easily the answer… He only have to search in the TP pages…
smile
JanD
21 Feb, 2018 15:10
Davide Coverta
Jan, I don't need to be unrespectfull versus somebody else, since we're here just to have fun…If you search by yourself, Images not deserved to be TP are not a rarity! Look and after reply if you are sure that all TP are images that we should consider as a reference!
BR
No offense meant, the question was of rather rhetorical nature. To become a TP, a picture has to be submitted by a person and reviewed by another. So there were at least two people who were convinced with a picture. Could it be disrespectful to try to change a, in my opinion, to the majority well working process just because someone is unhappy with a few decisions? Did someone speak up yet because he thought he deserved a TP but did not get one (I think you see were I want to lead to with this)?

But I see, I might see this differently than other people. I will not hijack this thread further… apologies.

Best,
Jan
DavideCoverta
21 Feb, 2018 15:14
JanD
(I think you see were I want to lead to with this)

And also you, can see were I want to lead with my previous words..
Search, the answer is so close  to you that i do not know why You're so sure on what you say…
Best regards
Edited 21 Feb, 2018 15:14
Jooshs
21 Feb, 2018 15:15
Roberto, the largest mosaic I’ve done is 2x2 drizzle stacked so far so I will defer to your experience there. That image was about 450 Mb so I had to reduce it to 20% to meet a 100 Mb size and that’s th extent of my experience. As you said, it still was published in near native resolution though.
rob77
21 Feb, 2018 15:16
JanD, let me clarify the topic.
Nothing has to or will be changed. The process will go on as it is right now.

With the due respect to reviewers and submitters, what this thread means is just to try to write (few) guidelines down that can help the selection of the images to be promoted to the upper tiers.

Cheers
Edited 21 Feb, 2018 15:25
carlocolombo
21 Feb, 2018 15:16
sono carlo colombo. Bene vorrei entrare in punta di piedi nei vostri discorsi selettivi,   io personalmente se una foto non pi piace non alzo il pollice, ma allo stesso tempo non vorrei scoraggiare chi magari ce la messa tutta  per fare una bella foto,  io stesso potrei essere in quelle condizioni, ma poi noto che difficilmente per quel che mi riguarda , le foto premiate sono notevoli, e molto raro che io non sia d'accordo sulla scelta delle premiate,  a questo punto però ho un suggerimento da vecchio  astrofilo, mi piacerebbe vedere sempre di più oggetti poco fotografati perché poco noti se non sconosciuti, ora pesate che nel fare una ricerca, proprio dentro astrobin ne già rintracciati  17  potrebbe essere interessante  vedere  qualcosa di nuovo? ciao a tutti voi e cieli sereni  c.colombo
Edited 21 Feb, 2018 15:19
rob77
21 Feb, 2018 15:17
Josh Smith
That image was about 450 Mb

JPG at 80% of quality?? How big was it ?

Cheers
Jooshs
21 Feb, 2018 15:18
No idea! 😊 I’ve never saved it as a jpeg. Probably not very big 😊
rob77
21 Feb, 2018 15:21
Ahhhh ok, it was TIF 450Mb. Understood.
If you save it at JPG 80% of the quality (you really don't lose much), it would become something below 100Mb (?, or even smaller).

I usually do this when I upload huge files.

Cheers
whwang
21 Feb, 2018 16:44
My browser refuses to load any JPG that's larger about 30 MB, and 20MB will be painfully slow to load and to display.

So even though I am perfectly capable of uploading a 100MP JPG, if I can't see the result on my own browser, I won't do it.
Edited 21 Feb, 2018 16:44
rob77
21 Feb, 2018 16:54
Ok but how many huge mosaics we have around? I haven't seen so many as far as I can remember. We can manage them as exceptions as far as it is written in the image data what it is.

It's clear, it's anyone choice to reduce or not but you'll understand that it's unfeasible to judge an image posted at 1/3 of it's original size.
Beeing a good image even at nearly the full res., IMHO, is really a plus and should be taken into account.

Cheers
Edited 21 Feb, 2018 16:56
whwang
21 Feb, 2018 17:13
The universe is big, so is astrophotography the hobby. There are numerous possibilities.  Some images are meant to be viewed at high magnification, otherwise its real beauty won't be appreciated.  But this is not necessarily true for all images. A beautiful astro-landscape doesn't need to be large to touch my heart, for example.

I think we are all human. We made mistakes. Sometimes an image with significant flaws could have been picked as a TP or IOTD. But those are rare.  I still trust that the submitters, reviewers, and judges have the wisdom to judge it in a case by case basis.
tolgagumus
21 Feb, 2018 17:17
[quote="Josh Smith" no stacking artifacts on the edges- no double stars in the final stacks from misalignment
- minimum of filters and gear used and detailed capture times per filter
- no obvious mottling or rice stars from over use of noise reduction
- only one IOTD per data set per season

One other point… I wonder if it makes sense to limit the frequency of certain targets being selected or moved to TP. For example, I count roughly 5-6 M 106’s, Rosettes, M42’s etc in the last couple of weeks. .
I agree with
1. no stacking artifacts on the edges -
2. no double stars in the final stacks from misalignment -
3. minimum of filters and gear used and detailed capture times per filter - (some minimum of details)
4. no obvious mottling or rice stars from over use of noise reduction

I do not agree with one IOTD per data set because one who is going to keep track of which data set is which, two how is it possible to determine if same set was used, three assuming the same set was used if the second persons processing is clearly better than the first who already was awarded, how can we tell him you are disqualified.

Same as the target, I have seen 1000s of M42s and M31s. But sometimes I see one that really jumps up.

Once the picture is through a initial technical requirements stage the only criteria should be "I like it"
 
Register or login to create to post a reply.