Banding is fixed in Firmware 2.20

Started Dec 17, 2019 | Discussions
Ulrich73 New Member • Posts: 10
Re: Almsot perfectly repeatable under these conditions

This is an example: raw file processed with C1, NR off, +1.5EV exposure, +15 shadows.

If you look at the right side of the image, striping is there in green and brown areas; also clearly visible in the upper part of the image, in the darkest brown areas. If you download the file and open it in photoshop, the striping is more evident.

In the original raw file as viewed in C1 after the aplications of the settings listed above, and in the converted 16bit TIFF file, the effect is more evident; it looks like that the reduction to a 8bit jpeg tends to reduce the problem.

This is the kind of situation in which I can see the "banding" issue; in all other situations that I tested, the sensor is very reliable and robust, I can pull shadows as much as I like without seeing any sign of it, very impressive.

As others already said, it looks like the sensor behaves differently in these situations, something different happens that messes things up. The good thing is that, with little effort in photoshop, it goes away more or less easily; still, I find it a bit frustrating.

(Not sure if exif is present, if not: can you please tell me how to iclude it? thanks)

Maybe more evident in B&W, same settings:

 Ulrich73's gear list:Ulrich73's gear list
Fujifilm X-E1 Fujifilm XF 35mm F1.4 R Rokinon 12mm F2.0 NCS CS
Virtual Photon Senior Member • Posts: 1,416
Re: Almsot perfectly repeatable under these conditions

Ulrich73 wrote:

This is an example: raw file processed with C1, NR off, +1.5EV exposure, +15 shadows.

If you look at the right side of the image, striping is there in green and brown areas; also clearly visible in the upper part of the image, in the darkest brown areas. If you download the file and open it in photoshop, the striping is more evident.

In the original raw file as viewed in C1 after the aplications of the settings listed above, and in the converted 16bit TIFF file, the effect is more evident; it looks like that the reduction to a 8bit jpeg tends to reduce the problem.

This is the kind of situation in which I can see the "banding" issue; in all other situations that I tested, the sensor is very reliable and robust, I can pull shadows as much as I like without seeing any sign of it, very impressive.

As others already said, it looks like the sensor behaves differently in these situations, something different happens that messes things up. The good thing is that, with little effort in photoshop, it goes away more or less easily; still, I find it a bit frustrating.

(Not sure if exif is present, if not: can you please tell me how to iclude it? thanks)

Maybe more evident in B&W, same settings:

Thanks for the examples.  It took me a while but I can see it in the colored version when I looked at it at 300 and 400%.  It's pretty faint.

 Virtual Photon's gear list:Virtual Photon's gear list
Nikon D500 Nikon Z7 Nikon Z6 Nikon Z50 Olympus PEN-F +12 more
El Duende
El Duende Contributing Member • Posts: 721
Re: Almsot perfectly repeatable under these conditions
1

I don't see any banding in *any* of the photos posted in this thread (and I looked hard). Not sure what people are talking about, blowing up 300 - 400% is always going to reveal some flaws in an image but banding is not one that I notice and have never noticed on any of my Z6 images.

-- hide signature --

paul v
-----------------------------------------------------------------------------------------
Photo Gallery: [ http://photos.transmitmedia.com ]

 El Duende's gear list:El Duende's gear list
Nikon D500 Nikon Z6 Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II Tokina AT-X 17-35mm f/4 Pro FX Nikon AF Nikkor 50mm f/1.4D +4 more
Ulrich73 New Member • Posts: 10
Re: Almsot perfectly repeatable under these conditions

Virtual Photon wrote:

Ulrich73 wrote:

This is an example: raw file processed with C1, NR off, +1.5EV exposure, +15 shadows.

If you look at the right side of the image, striping is there in green and brown areas; also clearly visible in the upper part of the image, in the darkest brown areas. If you download the file and open it in photoshop, the striping is more evident.

In the original raw file as viewed in C1 after the aplications of the settings listed above, and in the converted 16bit TIFF file, the effect is more evident; it looks like that the reduction to a 8bit jpeg tends to reduce the problem.

This is the kind of situation in which I can see the "banding" issue; in all other situations that I tested, the sensor is very reliable and robust, I can pull shadows as much as I like without seeing any sign of it, very impressive.

As others already said, it looks like the sensor behaves differently in these situations, something different happens that messes things up. The good thing is that, with little effort in photoshop, it goes away more or less easily; still, I find it a bit frustrating.

(Not sure if exif is present, if not: can you please tell me how to iclude it? thanks)

Maybe more evident in B&W, same settings:

Thanks for the examples. It took me a while but I can see it in the colored version when I looked at it at 300 and 400%. It's pretty faint.

Yes, after converting to jpeg it gets much less visible; in 16bit tiff and during raw processing in CO is much more evident.

I think that the interesting aspect of it is that, other than under these conditions, it never appears, even pulling shadows as crazy

 Ulrich73's gear list:Ulrich73's gear list
Fujifilm X-E1 Fujifilm XF 35mm F1.4 R Rokinon 12mm F2.0 NCS CS
jthomas39 Contributing Member • Posts: 563
Re: Almsot perfectly repeatable under these conditions

Yes, some faint banding, in the brown area. This is a screen crop of the 100% zoomed image, from the top left side.

It doesn't show as well in the reduced image compression. Click the "original size" link.

Ulrich73 New Member • Posts: 10
Re: Almsot perfectly repeatable under these conditions

jthomas39 wrote:

Yes, some faint banding, in the brown area. This is a screen crop of the 100% zoomed image, from the top left side.

It doesn't show as well in the reduced image compression. Click the "original size" link.

I tried to convert the same image from RAW with DXO photolab. With the default settings, striping does not come up, even pushing shadows. The difference is that CO applies a strong level of microcontrast to the image by default, DXO does not. Anyway, if you activate the "contrast" setting in DXO and increase microcontrast, it gets immediately visible.

 Ulrich73's gear list:Ulrich73's gear list
Fujifilm X-E1 Fujifilm XF 35mm F1.4 R Rokinon 12mm F2.0 NCS CS
frankjoke Regular Member • Posts: 349
Re: NEFs

I checked both NEF's about the differences in mEXIF with exiftools and found about the fact that one is created 6 minutes later and has 39mminstead of 40mm on the other it has 'Contrast Detect AF' "On" instead of "Unknown (2)" to the other!

So my feeling is confirmed that it has to do with some conditions on the shot defines the appearance of banding is somehow confirmed.

The reason why I may not see banding in my pictures is that I do not have this circumstance happening (noticeable often) in my way to shoot.

-- hide signature --

Greetings
Franz with Z7 + Irix, Tamron, Nikon F and Z lenses.

 frankjoke's gear list:frankjoke's gear list
Nikon Z7 Nikon Z 50mm F1.8 Tamron SP 150-600mm F5-6.3 Di VC USD G2 Nikon Z 24-70mm F2.8 Nikon Z 85mm F1.8 +3 more
Horshack Veteran Member • Posts: 8,680
Re: NEFs
2

frankjoke wrote:

I checked both NEF's about the differences in mEXIF with exiftools and found about the fact that one is created 6 minutes later and has 39mminstead of 40mm on the other it has 'Contrast Detect AF' "On" instead of "Unknown (2)" to the other!

So my feeling is confirmed that it has to do with some conditions on the shot defines the appearance of banding is somehow confirmed.

https://www.dpreview.com/forums/post/61841033

https://www.dpreview.com/forums/post/61841053

https://www.dpreview.com/forums/post/61856758

BJN
BJN Veteran Member • Posts: 5,110
Re: Banding is fixed in Firmware 2.20

I'd say it's reduced considerably...not completely gone. Can't say I've run into it yet in my photos.

-- hide signature --

BJ Nicholls
SLC, UT

 BJN's gear list:BJN's gear list
Panasonic Lumix DMC-GM5
sirhawkeye64 Veteran Member • Posts: 6,035
Re: ***RETRACTION, BANDING STILL REMAINS***

Timo Vahala wrote:

Nature of this phenomenon is erratic, depending on circumstances you may get pics without it, and then boom it`s back.

Timo

I think that's what Jim Kasson found in his tests.  For some instances, you can shoot up to ISO 6400 and not really see it, but in others it was visible at lower ISOs.

 sirhawkeye64's gear list:sirhawkeye64's gear list
Nikon Z7 Nikon Z7 II Fujifilm X-T20 Nikon Z5 GoPro Hero8 Black +14 more
SrMi
SrMi Senior Member • Posts: 2,081
Re: ***RETRACTION, BANDING STILL REMAINS***

sirhawkeye64 wrote:

Timo Vahala wrote:

Nature of this phenomenon is erratic, depending on circumstances you may get pics without it, and then boom it`s back.

Timo

I think that's what Jim Kasson found in his tests. For some instances, you can shoot up to ISO 6400 and not really see it, but in others it was visible at lower ISOs.

The PDAF banding discussed by Jim Kasson et al. is seen only at lower ISOs (less than ISO 640, I believe). At higher ISOs, the noise is hiding any possible PDAF banding.

See Jim Kasson's "Z7 PDAF banding FAQ" for more information.

 SrMi's gear list:SrMi's gear list
Leica SL2 Fujifilm GFX 100S Sony a1 Hasselblad X1D II 50C Olympus E-M1 III +1 more
Keyboard shortcuts:
FForum MMy threads