|
|
Groups: Member
Joined: 8/1/2007
Posts: 127
Location: Surrey, UK
|
Having just forked out £1200 for a Nikkor 24-70mm f2.8 lens, I was rather disappointed to get my first submission with it failed. But not for SoLD. It was for interpolation artifacts.
Having inspected the images at 100%, I can see none. I wouldn't expect any as they were taken on a Nikon D300 and the files were only interpolated from 34MB to 50MB using Bicubic Smoother in PS. There is possibly a little graininess in the shadowed foreground, but I would not expect a fail for that, and it's certainly not Int. Art.
I know there are many folks on here who swear that QC never get it wrong, but can anyone else see it? Once QC get on your case it's like shovelling s**t uphill.
Hope this link works.
http://s853.photobucket.com/albums/ab91/crivers_2009/?action=view¤t=newlidlstore06.jpg&newest=1
Clive (now the demotivated one). 
|
|
Groups: Member
Joined: 4/3/2008
Posts: 185
Location: Leafy, Posh South Bucks UK
|
when you have masses of blue sky interpolation artefacts usually show up as blockiness in the sky? Maybe ......?
pp
|
|
Groups: Member
Joined: 9/18/2007
Posts: 155
Location: No idea - too drunk
|
Evening Clive.
I suspect they are looking at the main raised yellow arm of the JCB - top edge where the yellow meets and runs alongside / beneath the dark piston thing on top. Looks like a problem in that area.
Gary.
|
|
Groups: Member
Joined: 7/2/2009
Posts: 165
Location: California
|
That picture is too small. I'd have to see it at 100%. Gary, how can you see anything with it so small? Am I missing a button somewhere?
|
|
Groups: Member
Joined: 7/14/2007
Posts: 223
Location: Scotland
|
I bet your 24-70mm and D300 is too sharp - they are seeing jaggies on those diagonal lines (the file is not big enough to check). I would be amazed to see any interpolation artefacts on this. But they just kicked out a set of mine for a dust-dirt problem, and while it was not dust-dirt (actually bits of out of focus foliage disconnected from their twigs by extreme 500mm mirror lens effect) I have duly removed the shading. They picked up 'camera shake' on another but again, it was not, it was actually a flash shot but pushing the limits of an 18-250mm lens wide open - the aberrations at the edges look a bit like shake. In that case I just deleted the image. My check procedure with Impression Media 2 zooms to the centre of each shot, which was OK when the edges were not.
So - 'interpolation artefacts' might just be the closest they can get to describing a problem. We didn't have dirt, or camera shake, but it was pretty clear why the two images had been kicked back.
David
|
|
Groups: Member
Joined: 4/20/2009
Posts: 70
Location: South Africa
|
Would need to see the pic at 100% ... you'll seldom see interpolation artifacts in smaller images.
Pretty Pictures wrote:when you have masses of blue sky interpolation artefacts usually show up as blockiness in the sky? Maybe ......?
That was the reason for my last rejection, and it's difficult sometimes to spot unless you specifically hunt for it. I can understand that it may happen if you need to adjust brightness/exposure a bit too much but don't understand why it occurs in a well exposed, bright daylight shot at low iso.
Is there a way to remove that blockiness?
iSpy photography
|
|
Groups: Member
Joined: 7/2/2009
Posts: 165
Location: California
|
ispy wrote:Would need to see the pic at 100% ... you'll seldom see interpolation artifacts in smaller images.
[quote=Pretty Pictures]
Is there a way to remove that blockiness?
I have painstakingly painted out artifacts before, but only to a couple of images I really wanted to use...not for the weak of heart, I'm afraid...
|
|
Groups: Member
Joined: 4/3/2008
Posts: 185
Location: Leafy, Posh South Bucks UK
|
if you shoot RAW you can remove blockiness by sliding the Luminance slider to the right in ACR; if you magnify a portion of blue sky whilst doing so you can see the blockiness disappearing; the danger with the luminance slider is that you can render the whole image slightly less sharp; if the image was sharp in the first place the very slight loss in sharpness is acceptabe. If there is any doubt I would import the image twice once without a luminance adjustment and once with and then you can reveal or hide parts of the image in adjustment layers.
pp
|
|
Groups: Member
Joined: 6/13/2007
Posts: 1,347
Location: Ireland
|
Clive Rivers wrote:Having just forked out £1200 for a Nikkor 24-70mm f2.8 lens, I was rather disappointed to get my first submission with it failed. But not for SoLD. It was for interpolation artifacts.
I think you've started off on a "false premise" by assuming your new lens will fix a problem that is not actually a lens-related one - but which has more to do with the sensor.
Interpolation (sometimes called re-sampling) means "to increase the number of pixels in a digital image." If you're using the same sensor as before, I can't really see why you'd have more success in interpolating an image from one size to another, than previously. The image might be sharper, or have more resolution or less flare than with a lower quality lens - but the pixels are still being pushed up as far as ever before.
And, you might have fallen into a false sense of security with the new lens and cropped even tighter on the original image - giving the poor old pixels even more work to do.
That said, QC may have confused the blue-sky noise for IA's. I often "spot" blue sky for excessive noise in Photoshop - and I have a sensor that produces 47.5MB files that only need 1% interpolation.
Stephen
|
|
Groups: Member
Joined: 7/2/2009
Posts: 165
Location: California
|
Stephen Power wrote:[quote=Clive Rivers]That said, QC may have confused the blue-sky noise for IA's. I often "spot" blue sky for excessive noise in Photoshop
Yeah, if you think the problem is in the sky (is it at all splotchy?), then run a filter like Dfine or Noise Ninja selectively on the sky only.
|
|
Groups: Member
Joined: 2/23/2009
Posts: 247
Location: Ormond Beach Florida
|
All the comments a great except that unless we can see the pic at 100% then due to the size restriction we really can't judge accurately. A full size image reduced down will almost always look blocky.
|
|
Groups: Member
Joined: 8/1/2007
Posts: 127
Location: Surrey, UK
|
Hi guys,
Many thanks for all your comments, although I don't feel I have got to the bottom of the problem yet.
Apologies, I didn't realise that the file on Photobucket would be so small, so I have added 4 crops which I hope will roughly equate to 100% view. I am still learning about posting pics on Photobucket et al.
http://s853.photobucket.com/albums/ab91/crivers_2009/?action=view¤t=lidlcrop04.jpg&newest=1
http://s853.photobucket.com/albums/ab91/crivers_2009/?action=view¤t=lidlcrop01.jpg&newest=1
http://s853.photobucket.com/albums/ab91/crivers_2009/?action=view¤t=lidlcrop03.jpg&newest=1
http://s853.photobucket.com/albums/ab91/crivers_2009/?action=view¤t=lidlcrop02.jpg&newest=1
If you mouse-over the top left corner of the image, you can zoom to 100%
I would emphasise that the image in question was not cropped at all and was only interpolated from 34MB to 51MB. There are no IAs and no jagged diagonals visible at 100%, as you will see from the crops. The sky is as smooth as a baby's bum!!
My best guess is that QC have seen slight noise in the shaded foreground block paving and have taken it to be IAs. I would be interested to hear whether anyone feels this is worth a fail. Seems a bit harsh to me.
Stephen,
My reason for buying the 24-70 lens was that my previous one was a 24-120 and was not quite sharp enough for me to feel confident in it. The reason was not IAs as I have never been failed for this in the past. As David said, it is almost impossible to get IAs with this small amount of interpolation using Bicubic Smoother in PS.
Thanks again for all your help and patience.
Clive
|
|
Groups: Member
Joined: 7/14/2007
Posts: 223
Location: Scotland
|
I don't see why that failed. The noise is not excessive (by Alamy standards, iStockphoto would not let it past, but they ask contributors to almost destroy the original quality of images to get them past). Certainly nothing there is what I would call 'interpolation artefacts'. Sure you have the right shot? I got very upset a couple of weeks ago about a QC rejection which took down a whole big upload, and it turned out I had two similarly named files by mistake and was checking the wrong one.
David
|
|
Groups: Member
Joined: 4/24/2007
Posts: 979
Location: In a house.
|
Hi Clive
There is alot of noise in the shadows - have you pushed the shadow areas with fill light? That may account for the edges of the block paving looking rather jagged.
Chris Potter
|
|
Groups: Member
Joined: 8/12/2009
Posts: 61
Location: Crete, Greece
|
Look at the black piston on top of the JCB, it’s covered in blocky artefacts.
|
|
Groups: Member
Joined: 5/31/2007
Posts: 201
Location: UK
|
Hi Clive, I wonder if using iso 800 has contributed to IA's? Although to be honest I cannot see any problem areas and it's an image that I would be happy to submit, from what I can see from the photobucket samples. I think you have tried to establish what/where the problem is and it would be acceptable, in this instance, to ask MS for further clarification 
|
|
Groups: Member
Joined: 7/2/2009
Posts: 165
Location: California
|
Yeah, there are artifacts in the shadows, most noticeably for me in the yellow pavement bricks. I wouldn't have chosen ISO 800 for a shot taken at 24mm and 1/750s. You could have spared some shutter speed there and dropped the ISO a stop (or two).
|
|
Groups: Member
Joined: 7/14/2007
Posts: 223
Location: Scotland
|
I'd still say that it just looks like noise, maybe with a touch too much sharpening. Interpolation artefacts should be something specific to interpolation. This noise would still be visible at native size. Even so, and despite the ISO 800 being a bad choice for the conditions, it points to a tighter QC now than in the past and I think Clive is right to be surprised.
One other factor - check all the other pictures in the same submission. If ISO 800 was used for a lot, and some reduced quality was seen as a result, QC may have simply have picked this as one example and not necessarily the worst (though usually they do pick the worst).
David
|
|
Groups: Member
Joined: 7/2/2009
Posts: 165
Location: California
|
David Kilpatrick wrote:Even so, and despite the ISO 800 being a bad choice for the conditions, it points to a tighter QC now than in the past and I think Clive is right to be surprised.
I have to agree that it's a bit surprising. I can find some artifacts, which could just be sharpened noise, but it's not that bad.
|
|
Groups: Member
Joined: 8/1/2007
Posts: 127
Location: Surrey, UK
|
Hi guys,
Thank you for all taking the time and trouble to help, much appreciated.
I agree that the shots should have been taken at a lower ISO, (note to self - check ISO before shooting!! ).
The images were sharpened slightly and this has emphasised the noise in the shadows.
I have reprocessed some of the images from the original RAW files unsharpened and have resubmitted them, except for 2 images which are most borderline, one of them being the image I have posted. I will submit those 2 separately when the bulk have been QCed.
Yes, David, they did pick the worst one!!
Regards and mucho sales to you all.
Clive
|
|
|
Guest |
RSS feeds