The Nikon Negative Narrative - Rebuttal

WestEndFoto

DenverShooter said:

I am just not seeing a noise issue with my D850. It does't look any worse (and possibly better) than images I shot on my D800E at the same ISO.

Denver Shooter

I think that the high megapixel vs noise tradeoff is not what it used to be. If you downsample a D850 file to a D5's (and a D6's) resolution, the difference is very minor. However, there are some who consider that very minor difference a major difference. I am not one of them.

ggbutcher

Keep in mind, it's really "less light" on the sensor that's responsible for most of the noise. ISO simply puts gain into the electronic signal that allows you to use less light.

WestEndFoto

Not quite ggbutcher.

The noise is there regardless of the amount of light. This is photon shot noise, which is the only remaining “material” noise of concern for photographers. Photon shot noise is a consequence of quantum mechanics and cannot be reduced. However, clever algorithms can hide some of it. There are other sources of noise but little progress has been made in reducing it in the last decade.

When there is “less light”, the gain is turned up and if the noise is a significant ratio of the total signal, you will start to see it. If you are shooting at ISO 100,000, the noise is not overpowering the image because there is very little image to overpower.

WestEndFoto

Which leads to exposure. In school they teach you about the exposure triangle. There is no such thing, though it helps beginners understand how to take pictures - note that I did not say “set exposure”.

Exposure is number of photons filtered by aperture and time. All ISO does is turn the volume up on an underexposed image. If the camera is ISO invariant (to keep things simple you can assume that all are) you can shoot at base ISO, “underexpose” the image, and turn up the volume in post producing exactly the same result.

I go to great lengths to avoid “turning up the volume”.

spraynpray

DenverShooter said:

I am just not seeing a noise issue with my D850. It does't look any worse (and possibly better) than images I shot on my D800E at the same ISO.

Denver Shooter

I didn't have a D800/E, I jumped from the D750 to the D850.

ggbutcher

WestEndFoto said:

Not quite ggbutcher.



The noise is there regardless of the amount of light. This is photon shot noise, which is the only remaining “material” noise of concern for photographers. Photon shot noise is a consequence of quantum mechanics and cannot be reduced. However, clever algorithms can hide some of it. There are other sources of noise but little progress has been made in reducing it in the last decade.



When there is “less light”, the gain is turned up and if the noise is a significant ratio of the total signal, you will start to see it. If you are shooting at ISO 100,000, the noise is not overpowering the image because there is very little image to overpower.

Interesting, you got me thinking about it a little differently. If my best camera were still the D7000, I'd be retaining my misconception because that camera's sensor did not class-out as "ISO invariant". But my Z 6 is definitely such, and so would isolate the Poisson character of the photon arrival pattern. I've just spent a few minutes looking at some of my low-light Z 6 images, and there's very little variation in any of the darkest regions. I just love that camera...

Thing is, I think a lot of photographers are still in "ISO-variant land", owners of older cameras...

WestEndFoto

I think about it differently too. It caused me to shift from shooting in A mode primarily to shooting in M mode switching Auto ISO on and off.

[Deleted User]

Spray, at ISO 320 with my D800s for instance, and a wide DR landscape, I am not seeing any noise in a 16x24 print, even in the shadows. However, if I accidentally oversharpen in post with RAW files, the artifacts appear as noise. So with NX-D, you have the standard sharpening as well as unsharpen mask. I eliminate the former, except for some minor mid-sharpening allowance depending on the image, and set the latter at say 6 radius and 45-55 intensity. I rarely use clarity and if I mistakenly use both sharpening inputs, there is noticeable "noise" in the image. I will save the image as a 'tif for printing, and examine it at 200% before sending it to the printer.

If setting sharpening in camera for JPEG output, I usually don't push it past 5 or 6. But it is rare that I shoot in camera jpegs.

Wondering if this is a sharpening issue?

PB_PM

WestEndFoto said:

For me, if I am shooting above ISO 800 on my D850, it is because the presence of noise is a creative decision. Street, candid and event photography are genres where noise is usually part of the intent.

If I care about noise I am likely shooting at base ISO and certainly less than ISO 400 and will use a tripod if I can’t achieve that handheld.

I find that VR improves flexibility to the point where I am rarely fretting about noise.

Totally depends on what shutter speeds you need to get sharp shots. Even on a bright sunny day if you move into the shade and need a shutter speed of 1/2000s at F8 to get the subject sharp, the ISO is going to go up. For street shooting an ugly grungy look seems to be the thing. I'm not into that, just personal taste. I happen to use the entire native ISO range, I just fix it later if I don't like the character of the noise in a shot. With the software out today there is no reason to stick to low ISO for moving subjects. I wouldn't shoot landscapes or still life that way, I stick to ISO64 for that, but for things that move I just turn on auto ISO and forget about it.

I've shot a good number of the advanced amateur FX DSLRS, D700, D750, D800, D810, and the D850. The D700, D750 have the best noise character. The high resolution bodies are like cheap film stock at high ISO, not something I like.

DenverShooter

A D750 is 24 MP and the D850 is 45 MP. The '850 is almost twice the resolution (1.875X). I think that would have to somehow be taken into account in terms of "noise performance"?

PB_PM

Thank you captain obvious.

DenverShooter

PB_PM said:

Thank you captain obvious.

So how exactly are you compensating for the vast difference in resolution when you make those claims about noise performance?

PB_PM

I’m just stating facts, is that a problem for you? It’s not a commentary on whether it makes the camera good or bad, just a fact. I happen to use noise reduction software to deal with it, also a fact. What’s the big deal?

WestEndFoto

PB_PM said:

I’m just stating facts, is that a problem for you? It’s not a commentary on whether it makes the camera good or bad, just a fact. I happen to use noise reduction software to deal with it, also a fact. What’s the big deal?

I think we are just asking for support for those facts. Without support, facts are at best opinions.

And opinions are OK.

For example, I stated an opinion a while back that BSI sensors are more expensive than FSI sensors. Despite a rebuttal that got personal, I still stood behind my opinion. It was an opinion because absent an invoice or price list from a producer that produces both FSI and BSI sensors, it is not a "fact". However, I still stood by the opinion because given how sensors are fabricated, what I have read, the opinion of a well known and informed commentator that I won't go around quoting which will jeopardize trust my opinion was strong enough that I would bet a significant sum of money on it to support. Opinions are often held in high regard. About a year ago I was an expert witness. My testimony was my opinion, but it held enough weight to be pivotal to the case. The judge basically bet $1.1 million that my opinion was correct. And yes, judgements are often bets based on opinions.

In the end, we usually don't have the luxury of having "facts". That is why a healthy debate based on opinions and facts, not personal attributes, is important.

[Deleted User]

Opinions should not be substituted for facts. It does nothing but create a "false narrative", which is why I started this rebuttal.

Unfounded opinions waste time of readers, and one has to wonder what is the purpose of those "assertions" other than to bring attention to a needy personality.

If you like something, ok fine, and opinion is welcome from personal experience - but we don't need arguers trying to elbow others using worthless "fake news" as basis.

If one doesn't have direct knowledge regarding the actual cost of a product (or its components), one should abandon the argument so that the stream of discussions can focus on better outcomes.

WestEndFoto

flip said:

Opinions should not be substituted for facts. It does nothing but create a "false narrative", which is why I started this rebuttal.

Unfounded opinions waste time of readers, and one has to wonder what is the purpose of those "assertions" other than to bring attention to a needy personality.

If you like something, ok fine, and opinion is welcome from personal experience - but we don't need arguers trying to elbow others using worthless "fake news" as basis.

If one doesn't have direct knowledge regarding the actual cost of a product (or its components), one should abandon the argument so that the stream of discussions can focus on better outcomes.

That is a very high standard Flip. There is very little that can be said that is supported by a "fact" and this forum would be dead if that was the standard. The legal system would also grind to a halt.

Consider the difference in price between a Z5 and Z6 at launch. How to explain that? It would seem that the difference in price between an FSI and BSI sensor would go along way to explaining that. I would think that this is something that is of interest to people. But who are you going to find that has direct experience of Sony Semi's costing? Are we not allowed have this discussion because of that? Are we not allowed to draw on the collective opinions of varying credibility available to us given the internet? Is it fake news because there is no invoice to support the assertion - just the opinions and views of well regarded experts in the field whos job it is to be so.

The issue here is not fake news etc. It is the ability for a diverse group of people with different opinions sometimes supported by facts to have a civil discussion. It should be up to the reader to make their own judgement about the value of such opinions.

[Deleted User]

Sensor cost is a rather simplistic approach at best to substantiate your narrative. Cost accounting is quite complicated and with the number of components and manufacturing costs, we have no idea what components represent what percentage of costs per unit. A lower sensor cost does not necessarily equate to a lower product price.

1. We have no idea the projected contribution margin per unit of the Z5 and Z6. For instance, Nikon may have decided that C Margins would be less for the former as they would be selling many more units. Nikon may have purchased many more FSI sensors so the per unit cost is less. Yet FSIs may be the same or similar in cost to BSI assuming the same number of units purchased. There is also the issue of supplier costs due to changes in manufacturing, demand, and ability to supply timely, and contract options to change the price due to these variables. On the other hand Nikon may have locked in prices prior to purchase. We have no knowledge of any of this and negotiations and unit cost may change on a dime, even if supplying the same sensor.
2. There is competitors pricing and Nikon may have lowered the price to where it meets market. Again corporate may have decided that C Margins for the Z5 would be less (and therefore pricing would be lower).
3. As I understand it, the Z5 has fewer components and technical capabilities, so Nikon likely has found ways to reduce its costs by giving the consumer less.
4. Nikon may have found a way to reduce direct labor costs and/or variable costs for the Z5 vs Z6 which again equates to a potential for lower price point.
5. A good CFO will have worked out the fine details of each component cost per unit, and based upon a number of scenarios, executives will make a decision as to price point for each product. In any event, the margins have to cover corporate allocation costs in theory, and if they plot a lower barrier (meaning an operating loss), they are doing it to meet the market pricing, again, nothing to do necessarily with sensor cost.

At least with Thom he mostly talks business product strategy based on his earlier life experience. Whether his comments are insightful is hard to say, but at least it is not based upon conjecture.

WestEndFoto

Maybe. Good points. I am not trying to explain why FSI sensors are cheaper, merely that they are.

Capt_Spaulding

To muddy the opinion/fact/conjecture water a bit. I spent a 40 year career formulating hypotheses, subjecting those hypotheses to "critical" testing and (attempting to) publishing the results. As many to most of you know, very few practitioners of the sciences will present their findings as "fact." Rather, most will simply say the data failed to disprove the hypothesis thereby conditionally confirming it. The result is still a "best guess" with a known probability that it is incorrect.

In the world of the law, we've seen time and again that eyewitness testimony is quite often unreliable. Facts can be very slippery creatures.

spraynpray

Excellent conversation WEF/flip/Capt_Spaulding. I think we can all agree that facts outweigh guesses but in the absence of facts, best guesses are all we've got but are dependant on the diligence of the guesser.

My comments on noise and my use of the D850 need clarifying: Any landscape where movement isn't an issue is shot at lowest ISO, on a tripod and VR off. I do not over sharpen (or use clarity much if at all as it is an image wrecker), I am happy to supply any of my images at full res for close inspection if anybody asks.

My main problem is with my astro shots and I find that I have given up shooting them for two reasons: Noise and difficulty of achieving absolutely critical focus. I admit I am OCD about those areas, some people shoot the 500 rule and are happy with their images but I find that the 500 rule only gives tolerable results if shooting the slowest moving parts of the sky which are also the parts with the least interesting features. I want lowest noise and round stars - that's not too much to ask for is it? My initial problems with the D750 were critical focus and bad aberrations but I stop my 24-35 f2 Art down to 2.8 and get OK results with the abs, but critical focus is a swine. I use a loop on the screen at 1:1 zoom but still have problems sometimes.

I am intrigued by @ggbutchers post but also think surely the D850 is new enough to be in the same bracket as the Z* generation sensors?

My Canon shooting buddy has no problems with focus - his live view seems better.

WestEndFoto

Thank you Spraynpray. I think that a diversity of opinions on this forum is something to aspire to. That means that we will often disagree. However, I believe that strengthens us, not weakens us. I think the core competency of a team member is to present their honest views, strive to understand other views, reward and encourage the same core competency in others and ensure that the arguments is focussed on the issue and not the participants.

Regarding your OCD, many very successful people have this trait. I would encourage it. I have it in some areas myself, both in my work and personal life.

Pistnbroke

Live view sharpness is affected by picture control setting for sharpness ..try +9
If its still at factory setting of +3 it will be soft soft soft even 5 or6 does not have bite of 9 RAW not affected

spraynpray

Interesting. I'll try various settings next time.

Capt_Spaulding

Thanks PB. I'll definitely give that a try.

Pistnbroke

Glad to be of help. As you know I only shoot JPEG .When I got my D850 I set it to Rockwells recomended 7. Rubbish. Not until I set to 9 did the pictures sing ( no not ring ha ha ) Why i took Rockwell's advice I dont know every camea from the D70 on I have had at +9 and as spray says the clarity control is a picture killer.