14-bit vs 16-bit Raw Bit Depth

[David Redfearn]

My GFX100S II is arriving tomorrow. One question about settings: I know the RAW files are very large, and will be larger if I select RAW 16-bit depth. I assume there is a (theoretical) color rendering advantage to the16-bit mode. Do users just leave it at 16-bits or is this something selected for a specific use case with most shooting done with 14-bit?

David

 

[Greg7579]

This has been discussed every week for 6 years. Here is the answer. Shoot 14 bit raw lossless compressed.

We all do, and there is no loss of quality from 16 to 14 bit or with lossless compressed vs uncompressed, and the file size saving is really big.

It's the first thing Jim taught me when I started with GFX.

 

[NAwlins Contrarian]

My GFX100S II is arriving tomorrow. One question about settings: I know the RAW files are very large, and will be larger if I select RAW 16-bit depth. I assume there is a (theoretical) color rendering advantage to the16-bit mode. Do users just leave it at 16-bits or is this something selected for a specific use case with most shooting done with 14-bit?

Our Intrepid Lead Scientist in Residence has made several tests and blogged about them. One small quote that sums up my sense of the situation:

My earlier read noise tests indicated that overall read noise of the GFX100 was virtually unaffected by the choice of 14 versus 16 bit raw precision .... There are some small differences in the raw noise pattern, but they fade into utter insignificance when compared to the horizontal banding th[at] comes from the OSPDAF system. If you eliminate the banding in post-production [and I think since then, a firmware update has eliminated it], the differences might conceivably be worth worrying about, but they are tiny.

(Emphasis added.) Source: https://blog.kasson.com/gfx-100/visual-comparisons-of-fuji-gfx-100-14-and-16-bit-raw-precision/

 

[JimKasson]

My GFX100S II is arriving tomorrow. One question about settings: I know the RAW files are very large, and will be larger if I select RAW 16-bit depth. I assume there is a (theoretical) color rendering advantage to the16-bit mode. Do users just leave it at 16-bits or is this something selected for a specific use case with most shooting done with 14-bit?

16-bit precision offers virtually no advantage over 14-bit precision with the GFX 100x cameras. The black point accuracy is a bit better for nosebleed ISOs. ES is slower. My advice is to forget 16-bit precision unless you know you want it for a good reason.

 

[SrMi]

My GFX100S II is arriving tomorrow. One question about settings: I know the RAW files are very large, and will be larger if I select RAW 16-bit depth. I assume there is a (theoretical) color rendering advantage to the16-bit mode.

The theoretical advantage is not to in color rendering but in noise.

Do users just leave it at 16-bits or is this something selected for a specific use case with most shooting done with 14-bit?

David

 

[David Redfearn]

OK, thanks! 14-bit it is!

David

 

[weihaowang]

My advice is to forget 16-bit precision unless you know you want it for a good reason.

Hi Jim,

Are there any possible reasons that you can imagine when 16 bit from GFX100x can really make a difference?

 

[fcracer]

My GFX100S II is arriving tomorrow. One question about settings: I know the RAW files are very large, and will be larger if I select RAW 16-bit depth. I assume there is a (theoretical) color rendering advantage to the16-bit mode. Do users just leave it at 16-bits or is this something selected for a specific use case with most shooting done with 14-bit?

16-bit precision offers virtually no advantage over 14-bit precision with the GFX 100x cameras. The black point accuracy is a bit better for nosebleed ISOs. ES is slower. My advice is to forget 16-bit precision unless you know you want it for a good reason.

 

[SrMi]

Note that with X2D, the raw file is the same regardless of bit size. The benefit of using 14 bits is only faster readout, which matters mostly for electronic shutter use.

 

[Greg7579]

My GFX100S II is arriving tomorrow. One question about settings: I know the RAW files are very large, and will be larger if I select RAW 16-bit depth. I assume there is a (theoretical) color rendering advantage to the16-bit mode. Do users just leave it at 16-bits or is this something selected for a specific use case with most shooting done with 14-bit?

16-bit precision offers virtually no advantage over 14-bit precision with the GFX 100x cameras. The black point accuracy is a bit better for nosebleed ISOs. ES is slower. My advice is to forget 16-bit precision unless you know you want it for a good reason.

Jim, I keep seeing YouTubers mention how amazing the 16-bit colour is on the GFX cameras. Giving them the benefit of the doubt, I think they’re just using 16-bit as a descriptor without any substance because they’re not sure where the extra visual improvement is coming from. I don’t see any real life benefit of 16 vs 14 but maybe I’m missing something. Am I missing anything with a reasonably exposed image?

They are probably looking at their GFX raw files for the first time on a 4K, 5K or 6K 32-inch pro IPS Black or Mini-LED high color gamut HDR new-tech screen for the first time.

They do like everyone else and go.... Wow!

Then they brainlessly attribute it to 16 bit. LOL.

 

[Greg7579]

My advice is to forget 16-bit precision unless you know you want it for a good reason.

Hi Jim,

Are there any possible reasons that you can imagine when 16 bit from GFX100x can really make a difference?

In fairness, there are a few really good photographers on this Board who shoot 16 bit (Manzur especially) because they run big arrays and don't car aboiut200 MB files sizes, so they shoot uncompressed raw and 16 bit.

I think they do it for some kind of future proofing. They don't see the difference now but think something may breakout later on future raw editing software.

In other words, they think there is some hidden future advantage, but they aren't sure what it is. They don't care about file size, so they shoot 16 bit and uncompressed raw.

 

[JimKasson]

Note that with X2D, the raw file is the same regardless of bit size. The benefit of using 14 bits is only faster readout, which matters mostly for electronic shutter use.

Right. The OP asked about the GFX, so I didn’t talk about other cameras.

 

[SrMi]

Note that with X2D, the raw file is the same regardless of bit size. The benefit of using 14 bits is only faster readout, which matters mostly for electronic shutter use.

Right. The OP asked about the GFX, so I didn’t talk about other cameras.

The title was not camera specific, therefore I thought it may be helpful to share information about other MF cameras as well.

What do you use with X2D, 14 or 16 bits?

--
https://blog.kasson.com

 

[JimKasson]

Note that with X2D, the raw file is the same regardless of bit size. The benefit of using 14 bits is only faster readout, which matters mostly for electronic shutter use.

Right. The OP asked about the GFX, so I didn’t talk about other cameras.

The title was not camera specific, therefore I thought it may be helpful to share information about other MF cameras as well.
What do you use with X2D, 14 or 16 bits?

14, so I don't have to remember to go to 16 when I'm using ES.

My H2D-39 had 16 bit readout, too, but at least 5 bits of that was noise.

 

[JimKasson]

Note that with X2D, the raw file is the same regardless of bit size. The benefit of using 14 bits is only faster readout, which matters mostly for electronic shutter use.

Right. The OP asked about the GFX, so I didn’t talk about other cameras.

The title was not camera specific, therefore I thought it may be helpful to share information about other MF cameras as well.

In general the decision rests on the relationship of the read noise to the smallest quantization level. If there is more than 1.5 or 2 least significant bits of read noise, finer quantization buys virtually nothing.

 

[JimKasson]

My GFX100S II is arriving tomorrow. One question about settings: I know the RAW files are very large, and will be larger if I select RAW 16-bit depth. I assume there is a (theoretical) color rendering advantage to the16-bit mode. Do users just leave it at 16-bits or is this something selected for a specific use case with most shooting done with 14-bit?

16-bit precision offers virtually no advantage over 14-bit precision with the GFX 100x cameras. The black point accuracy is a bit better for nosebleed ISOs. ES is slower. My advice is to forget 16-bit precision unless you know you want it for a good reason.

Jim, I keep seeing YouTubers mention how amazing the 16-bit colour is on the GFX cameras. Giving them the benefit of the doubt, I think they’re just using 16-bit as a descriptor without any substance because they’re not sure where the extra visual improvement is coming from. I don’t see any real life benefit of 16 vs 14 but maybe I’m missing something. Am I missing anything with a reasonably exposed image?

I think not.

 

[JimKasson]

--
https://blog.kasson.com

 

[David Redfearn]

I understand why you might want to shoot 16-bit - in the hopes that future raw decoders will improve and pull more out of these images. But why Uncompressed when lossless compressed raw is available (and returns the original image perfectly, bit-for-bit).

There was a long controversy about Sony's "lossy" compressed raw (all that was available for many years). You could definitely see a difference, but only in specific situations and it was very hard to see. Now Sony has lossless compressed raw (which is what I use for my A7RV) so that issue is gone now.

Saving space with these huge raw files is a plus - saves me some money as I deal with data volumes by buying more SSD drives. I wonder how fast I will amass a huge volume of the raw MF files?

David

 

[JimKasson]

I understand why you might want to shoot 16-bit - in the hopes that future raw decoders will improve and pull more out of these images. But why Uncompressed when lossless compressed raw is available (and returns the original image perfectly, bit-for-bit).

Two reasons, one real and one bogus.

The real one: uncompressed files are less affected by bit rot.

The bogus one: Lossless isn't really lossless.

 

[Stan Disbrow]

Hi,

Ha! Yeah.... That's the ticket!

Stan