MicrostockGroup Sponsors


Author Topic: Isolation contains stray areas that are either too feathered or rough rejections  (Read 7604 times)

0 Members and 1 Guest are viewing this topic.

123XXX

« on: March 19, 2011, 03:30 »
0
When you get a rejection saying "The execution of isolation contains stray areas that are either too feathered or rough.", I am wondering if the word "rough" in this case means rough in the sense that the edge is too hard and distinct or that the image is too roughly isolated with edges that are too soft? There is a difference between the two possible meanings and I am not sure what the intended meaning is of this cookie cutter rejection which I see quite often?

Also, I have had a lot of trouble with getting certain types of people images accepted in general and the rejection notice is always the same as above. The hair is almost always the issue and I am often not sure what to do with stray hairs and the edges of the head. Remove them or just trim them?

I am always worried if I remove too much loose hair and make the edge too fine and hard looking that it will look fake. Doing that does make the image look like a cutout for sure, but based on the rejections I am getting, and what I am seeing getting accepted much of the time, I have a feeling the hard edge around the head with no feathering is usually most acceptable?

Sometimes I try and even blur the edge of the head by a pixel width after I have removed all the stray hairs in order to make it look like a more natural tapering off of sharpness from natural lens optics as the front of the head area in the photo transitions into the back of the head and you have this natural softening curved edge. Again, this doesn't seem to pass though, but to me it looks more natural than a hard edge all the way around the head where it meets with the background.

Can anyone possibly shine a bit of light on this gray area (no pun intended) so I don't waste to much time on isolations and submitting over and over again?

Cheers..
« Last Edit: March 19, 2011, 03:46 by 123XXX »


123XXX

« Reply #1 on: March 19, 2011, 03:43 »
0
 :) - Ooops sorry. Hit post again by accident.
« Last Edit: March 19, 2011, 03:46 by 123XXX »

« Reply #2 on: March 19, 2011, 08:24 »
0
Submit a polite ticket to Scout asking for more clarification on the rejection. They're pretty good at narrowing down the problem.

« Reply #3 on: March 19, 2011, 09:06 »
0
It's been a long time since I had this rejection - so my response may no longer be relevent....

I think iStock likes a hard edge, no feathering. It is also possible they don't like very jagged edges such as hair that is cut with the magic wand.  For hair, cut off any stray hairs and tend towards the harder look.  The buyer can feather the edges into whatever background they are adding.

I've only done a couple of hair and soft fabric cutouts so don't have much experience.  I have done a lot of car cutouts and for these the pen tool is the best choice to make smooth curves along car body edges.  Other tools and approaches tend to leave notches and jagged edges where one would expect smooth edges.

If you are doing cutouts, be sure to leave the clipping paths in the file.  Having paths included is worth more money to many buyers - even for very simple objects.  Your path will match the cutout object better than a path a buyer might create after the fact so your path will save them time.

The iStock error/rejection messages tend to cover a multitude of sins.  Often the possibilities are conflicting so it is difficult for a new reader to understand the real issue from among the inclusive list.  To me, "too feathered" and "too rough" could well mean opposite problems.  In my case, I was more successful with no feathering and using hard edges that were not highly detailed/jagged.

Good Luck with your next try.

123XXX

« Reply #4 on: March 19, 2011, 11:38 »
0
Many thanks Stan. That is a useful opinion. It makes sense from the standpoint of giving a hard edge so it is easy for the buyer to select the person and die cut them out from then background and then they can soften the edges later if they want.

The only conflicting experience I have had is on one image where I composited a person against a locational background and it has been rejected 3 times on the same basis already. Yet the compositing work is print ad quality I assure you and I can't see how it can keep getting rejected.

I also can't imagine having a hard edge on the person on this kind of image would be good since it would make the person look stuck on top of the background rather than actually standing in that location and there would be no other reason for a buyer to be wanting to select the person out from the image again since it is a finished image already (versus something shot on white).

Seems there may be no cut and dry answer I guess and it comes down to the image, inspector, background type, and various other factors.

« Reply #5 on: March 19, 2011, 15:58 »
0
I used to get this rejection reason a lot and found it was mainly because of a: jaggy artifacts that are the result of jpeg compression (which are unavoidable) and b: reviewers who are inexperienced/uninformed and don't understand that it is unavoidable.

Scout usually sorts it out but it is annoying.

« Reply #6 on: March 19, 2011, 19:09 »
0
Some of the inspectors don't like shallow depth of field if a light object meets a white background in the OOF area.

123XXX

« Reply #7 on: March 19, 2011, 23:08 »
0
I used to get this rejection reason a lot and found it was mainly because of a: jaggy artifacts that are the result of jpeg compression (which are unavoidable) and b: reviewers who are inexperienced/uninformed and don't understand that it is unavoidable.

Scout usually sorts it out but it is annoying.

Interesting. What types of subjects are you having this problem with mainly? Again, my problem is pretty much limited to hair on the heads of people.

123XXX

« Reply #8 on: March 19, 2011, 23:54 »
0
Some of the inspectors don't like shallow depth of field if a light object meets a white background in the OOF area.

That would make sense as it makes it more difficult to isolate a subject from a white background if you have a soft edge.

« Reply #9 on: March 20, 2011, 02:05 »
0
I used to get this rejection reason a lot and found it was mainly because of a: jaggy artifacts that are the result of jpeg compression (which are unavoidable) and b: reviewers who are inexperienced/uninformed and don't understand that it is unavoidable.

Scout usually sorts it out but it is annoying.

Interesting. What types of subjects are you having this problem with mainly? Again, my problem is pretty much limited to hair on the heads of people.

I found that any object that is light coloured against white background (ie: yellow - I had trouble with isolated Lemons for example) when jpeg-ed resulted in gradient errors where the algorithm tries to fill in the "missing" colours. I think that's how to describe what's going on. You would have to understand exactly how the compression works and my knowledge is limited. All I know is that I would isolate an object perfectly with pen tool and then when I saved as a j-peg and re-opened the isolation would not be perfect anymore. It took me a while to figure out why I was getting the rejects as I was sure the isolations were flawless before uploading, but that's what was happening in my case.

Is the hair your having trouble with blonde / fair ?

Cheers

« Reply #10 on: March 20, 2011, 05:23 »
0
Some of the inspectors don't like shallow depth of field if a light object meets a white background in the OOF area.

That would make sense as it makes it more difficult to isolate a subject from a white background if you have a soft edge.

They don't like a natural transition that isn't "isolated" by messing about in photoshop, either. Maybe they just can't tell the difference.

123XXX

« Reply #11 on: March 21, 2011, 10:18 »
0
I used to get this rejection reason a lot and found it was mainly because of a: jaggy artifacts that are the result of jpeg compression (which are unavoidable) and b: reviewers who are inexperienced/uninformed and don't understand that it is unavoidable.

Scout usually sorts it out but it is annoying.

Interesting. What types of subjects are you having this problem with mainly? Again, my problem is pretty much limited to hair on the heads of people.

I found that any object that is light coloured against white background (ie: yellow - I had trouble with isolated Lemons for example) when jpeg-ed resulted in gradient errors where the algorithm tries to fill in the "missing" colours. I think that's how to describe what's going on. You would have to understand exactly how the compression works and my knowledge is limited. All I know is that I would isolate an object perfectly with pen tool and then when I saved as a j-peg and re-opened the isolation would not be perfect anymore. It took me a while to figure out why I was getting the rejects as I was sure the isolations were flawless before uploading, but that's what was happening in my case.

Is the hair your having trouble with blonde / fair ?

Cheers

That sounds strange to me. Paths and selections never change from opening and closing a file. I generally only work in layered PSD files though. On a JPG you can't save selection's, only paths. I am still a bit fuzzy on what your problem is and I don't mean to sound judgmental, but it sounds like an error somewhere in your workflow that is causing the issue.

As for backgrounds, you can get posterization, banding and artifacts occurring on solid color backgrounds when using a spot strobe on the background, but I don't see how that can happen on white. In addition, you can just fill the white area with white to make is continuous. The softness on the edge of the lemons might also be occurring from light bounce back from your white background. The more reflective the background and the closer your object is to the background the greater the potential problem. The softness could also be coming from a bit of lens flare from bounce back of light into the lens as well. There are various possible issues I am afraid if your light source is not flagged properly.  

In regards to my hair issue, I am only shooting black hair, but I just realized when looking at some of my rejections closely that there were some loose hair remnants around the head. What was happening was on the model's thin strands of hair that parts of the hair would get blown out in the middle sections of the hair strands. This gave the appearance of loose hair separated from the head. I have removed that now and resubmitted, but still awaiting reinspection to see if that was the issue with these photos.

In regards to the composited image on the dark background being rejected, that one is still a mystery.

Anyway, thanks for your thoughts.
« Last Edit: March 21, 2011, 10:27 by 123XXX »

123XXX

« Reply #12 on: March 21, 2011, 10:24 »
0
Some of the inspectors don't like shallow depth of field if a light object meets a white background in the OOF area.

That would make sense as it makes it more difficult to isolate a subject from a white background if you have a soft edge.

They don't like a natural transition that isn't "isolated" by messing about in photoshop, either. Maybe they just can't tell the difference.

Interesting point. You would think they would appreciate a gradual hardness to softness from close to the edge of the subject to the definitive edge, but if they consider those to be over feathered pixels at the edge then that could explain the problem. It means they don't look for a natural looking aesthetic, but rather a technical correctness on pixel edges. Seems strange though because on images shot naturally which haven't been isolated or composited you would never have a hard edge like that either. ???

« Reply #13 on: March 21, 2011, 15:16 »
0
I used to get this rejection reason a lot and found it was mainly because of a: jaggy artifacts that are the result of jpeg compression (which are unavoidable) and b: reviewers who are inexperienced/uninformed and don't understand that it is unavoidable.

Scout usually sorts it out but it is annoying.


Interesting. What types of subjects are you having this problem with mainly? Again, my problem is pretty much limited to hair on the heads of people.


I found that any object that is light coloured against white background (ie: yellow - I had trouble with isolated Lemons for example) when jpeg-ed resulted in gradient errors where the algorithm tries to fill in the "missing" colours. I think that's how to describe what's going on. You would have to understand exactly how the compression works and my knowledge is limited. All I know is that I would isolate an object perfectly with pen tool and then when I saved as a j-peg and re-opened the isolation would not be perfect anymore. It took me a while to figure out why I was getting the rejects as I was sure the isolations were flawless before uploading, but that's what was happening in my case.

Is the hair your having trouble with blonde / fair ?

Cheers


That sounds strange to me. Paths and selections never change from opening and closing a file. I generally only work in layered PSD files though. On a JPG you can't save selection's, only paths. I am still a bit fuzzy on what your problem is and I don't mean to sound judgmental, but it sounds like an error somewhere in your workflow that is causing the issue.

As for backgrounds, you can get posterization, banding and artifacts occurring on solid color backgrounds when using a spot strobe on the background, but I don't see how that can happen on white. In addition, you can just fill the white area with white to make is continuous. The softness on the edge of the lemons might also be occurring from light bounce back from your white background. The more reflective the background and the closer your object is to the background the greater the potential problem. The softness could also be coming from a bit of lens flare from bounce back of light into the lens as well. There are various possible issues I am afraid if your light source is not flagged properly.  

In regards to my hair issue, I am only shooting black hair, but I just realized when looking at some of my rejections closely that there were some loose hair remnants around the head. What was happening was on the model's thin strands of hair that parts of the hair would get blown out in the middle sections of the hair strands. This gave the appearance of loose hair separated from the head. I have removed that now and resubmitted, but still awaiting reinspection to see if that was the issue with these photos.

In regards to the composited image on the dark background being rejected, that one is still a mystery.

Anyway, thanks for your thoughts.



I don't think you understand what I'm saying. I probably didn't explain well. Many people have had this problem. See here: http://www.istockphoto.com/forum_messages.php?threadid=186131

123XXX

« Reply #14 on: March 23, 2011, 00:03 »
0
Sorry I misunderstood and thanks, I will have a look. Thankfully all those pictures in question on the white background were accepted the second time around. I think it was those stray looking hairs as I mentioned which were separated by the lighting.

In terms of the composited image though that hasn't been approved because of the hair on the person I layered onto a locational background, I think I am still left with my original question as to what the word "Rough" actually means in their standard rejection on isolation issues. Does too rough mean soft edged pixels which look over feathered or an edge that is too hard and definitive and just drops right off?

« Reply #15 on: March 23, 2011, 02:57 »
0
I think I am still left with my original question as to what the word "Rough" actually means in their standard rejection on isolation issues. Does too rough mean soft edged pixels which look over feathered or an edge that is too hard and definitive and just drops right off?

Who knows  ::)

RacePhoto

« Reply #16 on: March 27, 2011, 02:04 »
0
I think I am still left with my original question as to what the word "Rough" actually means in their standard rejection on isolation issues. Does too rough mean soft edged pixels which look over feathered or an edge that is too hard and definitive and just drops right off?

Who knows  ::)

Just got this one for an image of a glass of beer against a white backdrop. Somewhat similar. Point is, it was shot isolated, not cropped or fixed. All I did was levels and some pure white on a small shadow, that wasn't touching the glass. Yes I did levels so the white was white and that's about it. I did not do noise reduction. Could be that the problem is the camera, I used a S-90 as I was just playing around at the bar. Accepted SS and BigStock.

++Posterisation, noise reduction


 

Related Topics

  Subject / Started by Replies Last post
21 Replies
7722 Views
Last post May 26, 2008, 21:12
by Tomboy2290
11 Replies
3309 Views
Last post December 27, 2011, 14:01
by BaldricksTrousers
5 Replies
2054 Views
Last post June 02, 2013, 13:06
by cathyslife
13 Replies
3567 Views
Last post July 10, 2013, 14:38
by old crow
4 Replies
1996 Views
Last post February 21, 2017, 13:55
by Pauws99

Sponsors

Mega Bundle of 5,900+ Professional Lightroom Presets

Microstock Poll Results

Sponsors

3100 Posing Cards Bundle