GR V 12 Bit

[oliveoil]

I read some concerns re the GR V being 12 bit..The Nikon Coolpix A is 14 Bit..Please can anyone explain the relevance of this to me and could it be a possible deal breaker..esp if nikon lower their price.

Thanks for your opinions.

 

[BingoCharlie]

12 vs 14 bit raw should be the very least of anyone's concerns when deciding which compact fixed-lens camera to buy.

 

[abluesky]

BingoCharlie wrote:

12 vs 14 bit raw should be the very least of anyone's concerns when deciding which compact fixed-lens camera to buy.

Not necessarily.

14 bit provides increased tonal range within the same dynamic range as 12 bit.

While this means little regarding highlights, it could be significant for shadow recovery for severely underexposed photographs.

For street photography with an A or GR, low light or night photography would require high iso (3200-6400) to permit stopping down the aperture for increased DOF.

A 14 bit file would allow more latitude to underxepose and then recover the shadow detail in post processing.

Theoretically, the A would provide better IQ since it is 14 bit

 

[aleksanderpolo]

The difference between 12 and 14bit can be illustrated by K5 and K30:

Loading…

www.dxomark.com

Not much difference in color depth or high ISO, about less than a stop difference in dynamic range

 

[oliveoil]

Thankyou for explaining that.

 

[BingoCharlie]

abluesky wrote:

Theoretically, the A would provide better IQ since it is 14 bit

I get that in theory. More information is better than less information when processing raw. But I've read quite a bit about this subject (after worrying about it with a different camera) and have never seen anyone convincingly demonstrate a difference.

There are so many other factors to worry about than a razor thin, almost purely theoretical advantage from going from 12-bit to 14-bit raw. Differences in lens, sensor, processing, and autofocus performance will matter more in the real world than 12 vs 14 bit. The K5 and K30 use difference sensors and aren't really comparable at this level of micro-difference. Take the same camera. Shoot 12 bit. Then shoot 14 bit. It is impossible to see any difference.

I suppose if you were truly stuck at 50/50 between the two cameras then 12 vs 14 might tip the balance, but that seems unrealistic.

Do we in fact know that the GR shoots 12-bit? (I'm not skeptical, I just don't really know.)

 

[BingoCharlie]

For instance, give me the GR's sharp corners over the Coolpix A's mushy ones any day. This will make a much bigger real-world difference in my shooting than an extra two virtually worthless bits of raw data. (Frankly, Nikon's performance, for a fixed-lens camera, is appalling.)

Obviously, YMMV.

From DPR's test:

 

[AngryCorgi]

The 12-bit 16MP exmor is a different sensor than the 14-bit 16MP exmor (they have on-chip ADCs). The 14-bit variant has been shown to have ~0.8EV greater DR than the 12-bit variant at equal sensitivities, for one. Second benefit is the additional shadow tonality information. If you are simply a JPEG shooter only, then you will probably notice NOTHING different between these sensors. If you process raws, the difference, under extreme situations, is quite visible.

 

[BingoCharlie]

AngryCorgi wrote:

The 12-bit 16MP exmor is a different sensor than the 14-bit 16MP exmor (they have on-chip ADCs). The 14-bit variant has been shown to have ~0.8EV greater DR than the 12-bit variant at equal sensitivities, for one. Second benefit is the additional shadow tonality information. If you are simply a JPEG shooter only, then you will probably notice NOTHING different between these sensors. If you process raws, the difference, under extreme situations, is quite visible.

--
-AC-

I'm sure this will sound combative, but I don't mean it to be. Can you show me two pictures, side-by-side, from the same camera, where this effect is shown?

 

[Mousetrapper]

Though I would not mind 14-bit, I would not have it as a top priority when comparing cameras, it would actually be pretty far down on the list. So far down it would not even weigh in as a factor.

In "very extreme" situations I find that 14-bit would not make enough of a difference and I usually avoid shooting in such extreme light. HDR does not float my boat either.

I grew up shooting slides and exposing for highlights and with that approach I have yet to have a serious problem with modern digital cameras. There is still no camera that can take awful, extreme/harsh light and transform it into a really pleasant picture. A decent picture, yes.

Pretty much all the other aspects of the Coolpix A vs Ricoh GR comparison have me leaning towards the GR and that 12 vs 14 BIT difference will not affect my choice. Not one BIT.

I think the Coolpix A looks pretty nice though I am not sure I would be happy with the lens performance though stopped down it looks awesome. I applaud Nikons effort with the camera and when it takes a nose dive with the pricing I think we might see quite a few being sold.

Both sensors are great sensors and I would be happy with either one. The NEX5n I briefly toyed with produced very nice files so I would be happy with that sensor in any camera.

 

[AngryCorgi]

BingoCharlie wrote:

For instance, give me the GR's sharp corners over the Coolpix A's mushy ones any day. This will make a much bigger real-world difference in my shooting than an extra two virtually worthless bits of raw data. (Frankly, Nikon's performance, for a fixed-lens camera, is appalling.)

Obviously, YMMV.

From DPR's test:

To be fair to BOTH cameras, the samples show flaws in both lenses (different areas). The upper-right mid-zone where the cards are is an example:

GR top, A bottom - upper-right mid-zone...

So, neither lens is all roses.

--
-AC-

 

[oliveoil]

I'm A RAW shooter..Used to Leica M9 files and rather sensitive about the way a sensor renders.

 

[AngryCorgi]

BingoCharlie wrote:

AngryCorgi wrote:

The 12-bit 16MP exmor is a different sensor than the 14-bit 16MP exmor (they have on-chip ADCs). The 14-bit variant has been shown to have ~0.8EV greater DR than the 12-bit variant at equal sensitivities, for one. Second benefit is the additional shadow tonality information. If you are simply a JPEG shooter only, then you will probably notice NOTHING different between these sensors. If you process raws, the difference, under extreme situations, is quite visible.

--
-AC-

I'm sure this will sound combative, but I don't mean it to be. Can you show me two pictures, side-by-side, from the same camera, where this effect is shown?

Its a fair question, and one I get a lot when explaining this to people. No combative tone in that, IMO.



I can show you the difference in 12-bit vs 14-bit in terms of shadow tonality if you like:

D7000, 14-bit RAW, ISO100, Pushed 5EV

D7000, 12-bit RAW, ISO100, Pushed 5EV

You see how the 14-bit file retains some mid-tones in the shadows when pushed, while the 12-bit file more harshly clips to colors (especially visible in the reds here).



In regard to a direct comparison between a 14-bit and 12-bit exmor sensor, I do not have one of each to directly compare. DXOmark tests this and I have found their #s to be very accurate in the past when comparing other cameras, though. That is the source of the "0.8EV" variance in DR referenced above. A few conspiracy theorists think DXOmark cooks their data and is paid to do so, but all the evidence I have suggests this is NOT the case.

--
-AC-

 

[aleksanderpolo]

No doubt that there is going to be some difference in DR, 0.8EV seems to be a reasonable number. But when you need to push 5EV to see the difference, I think we are pushing deep into the pixel peeping level.

Would a 14bit raw cooler? Sure. How much is one willing to pay for the ability to push 5EV a little better? $300? That's left for the buyer to decide.

 

[AngryCorgi]

aleksanderpolo wrote:

No doubt that there is going to be some difference in DR, 0.8EV seems to be a reasonable number. But when you need to push 5EV to see the difference, I think we are pushing deep into the pixel peeping level.

Would a 14bit raw cooler? Sure. How much is one willing to pay for the ability to push 5EV a little better? $300? That's left for the buyer to decide.

The demonstration is intended to pronounce the effect, and is not necessarily a real-world application. The missing shadow tonality is there in less strenuously manipulated examples too. Again, if you shoot JPEG, it isn't going to matter much (except that the GR JPEG engine appears to be not-too-great) in terms of 12-bit vs 14-bit.

 

[Andrewteee]

oliveoil wrote:

I read some concerns re the GR V being 12 bit..The Nikon Coolpix A is 14 Bit..Please can anyone explain the relevance of this to me and could it be a possible deal breaker..esp if nikon lower their price.

Thanks for your opinions.

More is always better. Spinaltap told us that!

 

[abluesky]

abluesky wrote:

Theoretically, the A would provide better IQ since it is 14 bit

I get that in theory. More information is better than less information when processing raw. But I've read quite a bit about this subject (after worrying about it with a different camera) and have never seen anyone convincingly demonstrate a difference.

There are so many other factors to worry about than a razor thin, almost purely theoretical advantage from going from 12-bit to 14-bit raw. Differences in lens, sensor, processing, and autofocus performance will matter more in the real world than 12 vs 14 bit. The K5 and K30 use difference sensors and aren't really comparable at this level of micro-difference. Take the same camera. Shoot 12 bit. Then shoot 14 bit. It is impossible to see any difference.

I suppose if you were truly stuck at 50/50 between the two cameras then 12 vs 14 might tip the balance, but that seems unrealistic.

Do we in fact know that the GR shoots 12-bit? (I'm not skeptical, I just don't really know.)

That's kind of been my experience as well: practical results fail to prove the theory.

 

[Tom Caldwell]

oliveoil wrote:

I read some concerns re the GR V being 12 bit..The Nikon Coolpix A is 14 Bit..Please can anyone explain the relevance of this to me and could it be a possible deal breaker..esp if nikon lower their price.

Thanks for your opinions.

Firstly we must make the assumption that what was seen was actually true (I am not disbelieving, just noting that an argument about a fallicy is not an argument at all).

Meanwhile to try and make a statement without being elitist I have to argue that those who base an argument about whether of not to buy a GR (or previously a GRD) simply based on sensors and IQ of sample images are missing quite a lot on just what the Ricoh camera might be about. Which is "a very satisfying camera to use". If interface and advanced firmware count for nothing then cameras can very easily be compared on what images they produce and everyone can get used to the eccentricities of the camera that eventually does their business.

So every car has seats and a steering wheel, therefore we can compare each scientifically on how quick from zero to 100 or how econonical a V8 engine can be with automatic in econo-mode? Maybe how great the sound system is? No doubt everyone has their own needs and so it must be but in the end all cars are "just transport" as much as all cameras "just take images". Sometimes the nth degree ends up unaffordable or impractible. As much as we admire a Ferrari there are better cars to drive to the shops.

Therefore looking at a Nikon A (if the price reduces) because of 14 bit raw is missing quite a few other details of comparison upon what many others might make a decision. Like whether you are actually going to enjoy the ride.

 

[Tom Caldwell]

Ricoh reserves their "GR" designation for what they regard as their best lenses. Just as much as other manufacturers put coloured rings around the barrels what they think are their best quality lenses. This is probably because those that buy cameras are unable to normally tell the difference and it is behoven of the manufacturers to explain in some way that "hey guys you are paying extra money for the quality". I am not sure what "standard" the Nikon A lens is by Nikon's own reckoning scale. Nikon does make great lenses, some have a reputation for being superb, however there is no doubt that they might make "cooking" versions as well. But I am not into Nikon lenses enough to know what their special hand signals to Nikon purchasers are a to lens quality.

Ricoh for a good number of years has specialised in small lenses, many of the collapsible variety. Collapsing lenses generally are not premium lenses. But no doubt Ricoh has put a lot of work into lens quality on small cameras and this might be a spin off from very substantial involvement in copier lens design and manufacture. Good copiers don't distort their images.

The highest quality lenses are not always the fastest and notably many of the opposition do make faster lenses by type than Ricoh. Maybe they are also high quality lenses and make good images, I am not in a position to judge but the "GR" - emphasised in red lettering - on the "GR" camera (that is also an evocative label) means that Ricoh are proud of the lens that it is installed. For users seeing will be believing.

Meanwhile I would be very interested to know if Nikon regard their lens in the A as "something special" from the Nikon lens manufactury - or just another "high quality" Nikon lens, which would be no real slouch anyway.

 

[Tom Caldwell]

aleksanderpolo wrote:

No doubt that there is going to be some difference in DR, 0.8EV seems to be a reasonable number. But when you need to push 5EV to see the difference, I think we are pushing deep into the pixel peeping level.

Would a 14bit raw cooler? Sure. How much is one willing to pay for the ability to push 5EV a little better? $300? That's left for the buyer to decide.

I can add to that - would 90% of users take images where they would be able to see the difference? Make that 99%.

Therefore a lot of this debate is based on the fanatical need to have the very best of something and for most they might not really care whether it takes 4.5 seconds to get to 100 or 5.0 seconds, but 4.5 seconds sells better.