[Eigengrau]

-> Falk Lumo: Camera equivalence

(I've read this statement so many times that I decided to make my own article I can refer to whenever this false information comes up)

Forgive me if I'm mistaken, but if this is the case, why design the Q with such a small sensor? One of the main justifications was that it enabled smaller lens design. And it seems to have worked, considering it is the smallest of the mirrorless group. I think you might be discounting aperture. In order to make cameras strictly equivalent (in terms of DoF) aperture would have to go through the roof - which just doesn't happen.

On the wide end of things, the APS-C equivalent of 31/1.8 would be something like 21/1.4. Not a lot of those lenses around. Or, consider the FA 20/2.8 and the DA 14/2.8 (which should be more like 2.0 to keep things equal). We can't really keep DoF constant because (judging by the lenses that have been made) that is impracticable on the wide end with APS-C. On the long end of things, APS-C seems to be a clear win, because you get the 200/2.8 instead of the 300/4.0. Smaller and faster for the same field of view and depth of field.

Since my background is in physics, I appreciate your desire to compare idealized cameras, but I think it discounts the realities of what lens designs are and aren't practical.

[RXrenesis8]

if this is the case, why design the Q with such a small sensor?

First and foremost, the Q was designed with a smaller mount. The sensor size is simply a byproduct of this arrangement. The shorter registration distance is the real winner in terms of smaller lenses.

Case in point: rangefinder lenses. They have to cover the same image circle as their SLR counterparts but they can be made smaller, wider, and faster. All thanks to being closer to the film.

[falconeye]

Since my background is in physics, I appreciate your desire to compare idealized cameras, but I think it discounts the realities of what lens designs are and aren't practical.

I assume you didn't read the paper.
The blog article is meant to drag people into reading it, not making preliminary comments. The blog article covers 10% of what the paper does.

If you however, did read the paper, please specify your concern. I can't see anything I didn't already discuss. Esp. wrt Q, please read the paragraphs about cost performance curves for given equivalence classes.

[drougge]

Forgive me if I'm mistaken, but if this is the case, why design the Q with such a small sensor? One of the main justifications was that it enabled smaller lens design. And it seems to have worked, considering it is the smallest of the mirrorless group. I think you might be discounting aperture. In order to make cameras strictly equivalent (in terms of DoF) aperture would have to go through the roof - which just doesn't happen.

As I see it, the small sensor is basically an excuse to make the lenses small by making them slow. f/1.9 doesn't sound that slow, but it's equivalent to an f/11 lens on FF. Of course shorter lenses are also smaller (unless they are retro focus and/or need to cover a larger image circle than their focal length), but much of the win is probably aperture related here.

[maxfield_photo]

Great paper Falk, I even learned a thing or two I didn't know, thanks. Ironically though I think the folks who stand to gain the most by reading it are the least likely to do so. I hope you take comfort in the fact that you are fighting the "good fight".

[falconeye]

Ironically though I think the folks who stand to gain the most by reading it are the least likely to do so.

I think so too. But now I can just dump the link and keep myself out of such discussions

Moreover, I need it as base to list arguments for full frame or to explain why APSC is expensive when surpssing a given image quality threshold. Of course, APSC can be good enough and in the paper I admit it.

[Eigengrau]

I assume you didn't read the paper.
The blog article is meant to drag people into reading it, not making preliminary comments. The blog article covers 10% of what the paper does.

If you however, did read the paper, please specify your concern. I can't see anything I didn't already discuss. Esp. wrt Q, please read the paragraphs about cost performance curves for given equivalence classes.

Having read the paper, I find that it is exceedingly well put together and raises some really interesting points. I find that it really illuminates a lot of facets of the situation very well, and brings in valuable new information to the argument. However, I'm not sure that I draw the same conclusions you do.

If I were to paraphrase what I feel like the outcome of the paper is, it would be:
We don't typically compare equivalent cameras. Here's what will happen if we do (we see big differences in ISO and aperture, among other things). Because of economic and design constraints, we still won't be able to compare equivalent systems most of the time.

While looking at the physics of the situation helps us to speak more intelligently about these topics, in this case I prefer a more pragmatic approach focused on economics and especially end use. The reality is that I buy a camera because I want to shoot a certain type of picture in a certain type of environment, and I don't usually care how that happens, but I usually want to do it as cheaply as possible. My opinion of the situation is that people aren't as concerned with getting a specific amount of DoF as they are with getting bokeh when they want it, which would make strict equivalence less relevant.

On the original point, that FF lenses are larger/more expensive/etc: This is more an observation on my part of existing lenses than a reality of the universe. Sure, you can analyze why that shouldn't be the case, and why it perhaps won't be in the future, but what of the actual situation right now? If you want a lot of reach, it seems to me that my contention holds pretty convincingly (given what the buyer is trying to ultimately get), but I'll admit that at the wide end it doesn't hold up as well.

At any rate, those are my thoughts - like I said, I certainly find your paper enlightening, but I go a different place with it than others might, because I think people make decisions in a way that isn't always mathematically justifiable. Just my opinion.

[falconeye]

@Eigengrau,

I think you would be able to express all of your points in a more concise way when starting from a comparison of equivalent cameras first.

E.g., if you compare the equivalents for a DA*50-135/2.8, you'll find the really excellent Canon 70-200/4L in Canon land but no equivalent for Nikon. This is a much more concise statement than "FF lenses are larger" which - in this form - is false.

It shows too that the equivalent FF lens is actually cheaper. OTOH, the Nikon 70-200/2.8G is more expensive, heavier. But faster too which materializes the advantages of the FF system and at 24MP resolves much better with a larger pixel pitch than APSC with 16MP.

Add to it the fact that tele reach depends on pixel pitch rather than crop factor (type-2 equivalence in the paper) and you see that advantages at the tele end depend on options for pixel pitch and apertures. E.g., a D7000 hasn't more reach than a D800 so ATM, DX has no reach advantage over FX. Of course, this is constantly changing, e.g., with a 24MP D8000.

This kind of evaluation actually opens the eyes for a big opportunity for Pentax to go FF and still be different: Just fill the vendor-spanning gaps for excellent glass which is equivalent to the speed of APSC glass which in many cases is more than sufficient: DFA*24-70/4 DFA*70-200/4->DA*60-250/4 DFA*500/5.6 DFA*14-24/5.6 Such a system wouldn't be larger than FF and make it unique. Including lens cost, such a system would be half cost of D800.

[Eigengrau]

@Eigengrau,
This kind of evaluation actually opens the eyes for a big opportunity for Pentax to go FF and still be different: Just fill the vendor-spanning gaps for excellent glass which is equivalent to the speed of APSC glass which in many cases is more than sufficient: DFA*24-70/4 DFA*70-200/4->DA*60-250/4 DFA*500/5.6 DFA*14-24/5.6 Such a system wouldn't be larger than FF and make it unique. Including lens cost, such a system would be half cost of D800.

This is certainly an interesting idea, but the "non-equivalence" of current cameras seems to serve, at least partially, to correct the ISO problem. Because, (correct me if I'm wrong) if we went with the lens system that you suggest, the full frame camera would have to be perpetually operating at a higher ISO for the equivalent shot, which at least at face value will be a big turnoff to consumers.

[RonHendriks1966]

This is certainly an interesting idea, but the "non-equivalence" of current cameras seems to serve, at least partially, to correct the ISO problem. Because, (correct me if I'm wrong) if we went with the lens system that you suggest, the full frame camera would have to be perpetually operating at a higher ISO for the equivalent shot, which at least at face value will be a big turnoff to consumers.

Well most users don't use their lenses wide-open. That is for all those nice people on forums. So as long as those f4 lenses are sharp wide-open they won't even miss the f2.8. Problem is that no serious pro-users would buy in such a system.

[falconeye]

Problem is that no serious pro-users would buy in such a system.

That's true but isn't my suggestion.

My suggestion is to fil lthe gap with equivalent glass which is often missing from the current FF lineup.

I do not suggest that Pentax wouldn't (possibly over time) add the typical FF glass such as 70-200/2.8. Eventually, such heavier and possibly more expensive glass is necessary to fully exploit the possibilities of an FF system. And then pro-users are happy too.

My point is to overcome the current a priori choice that you either have a bulky FF system with great image quality or a portable APSC system with good image quality.

Of course, one may run two systems side by side, Pentax for APSC and e.g., Nikon for FF because of this current gap. But sooner or later, the current FF lineups will expand to close these gaps.

[RonHendriks1966]

Of course, one may run two systems side by side, Pentax for APSC and e.g., Nikon for FF because of this current gap.

Well I see a lot of sportsphotographers with the combo D3s/D300S or 1D Mk IV/7D since they run the same lenses. So there is more to win with a FF camera for Pentax because running two systems isn't that great in practice for most (pro) users. It does happen but not that often.

[jsherman999]

.


Interesting commentary and speculation on FF, the upgrade path... and the beginning of the squeeze from above on aps-c, perhaps?

Feb 25 (news and commentary)--Dealers on Friday got some new pricing from Nikon, specifically on the D700. The new suggested price is soon going to be US$2199 (currently US$2699). But here's an interesting kicker: there apparently won't be a minimum advertised price (MAP) associated with that, which would mean we'll likely see someone drop under the US$2000 mark.

Some people have questioned my slight shift on predicting what Nikon will introduce next. Actually, it hasn't been a slight shift. If you go back and read what I wrote in 2010 versus where we are today, I think you'll find that Nikon went a bit different direction than I originally expected. The post-quake thinking at Nikon seems to be a bit different than the pre-quake thinking, too. Nikon seems a bit more emboldened in its decision making since the last management change. Looking back on my conversations with Nikon executives over the past couple of years as well as anonymous tips I receive, I can see that I didn't pick up on all the clues that were dropped. Mea culpa.

But let me explain one thing that still seems to be hanging a bunch of you up: entry FX. First, it should be clear that a US$2000 D700 is very much an "entry FX" model ;~). And a danged good one, at that. Many of you seem perplexed by why an entry FX model makes sense, and why a US$1000 difference in price between it and a D800 works.

First the rationale: the market for new DSLR sales boils down to upgraders. The notion of "new camera users" coming into the market is mostly wrong. Young adults aren't opting for DSLRs, and that would be only a small percentage of the purchasers now, anyway. The side-grade from film SLR to DSLR is now mostly complete.

So today Nikon is actively soliciting Coolpix users to upgrade to CX (Nikon 1). CX users will be solicited to upgrade to DX. And DX users, well, it's only natural to upgrade them to FX. But if the entry FX body is 3x the price of the top DX body, that's a pretty big money leap. Entry FX can't be more than 2x the top DX price if it is to encourage upgrading. Indeed, it probably should be 1.5x (which would be about US$1800). That puts us right at the likely D400 pricing, which is one reason why I think the D400 could go either way (DX or FX).

Yes, a DX D400 at US$1900 and an FX D800 at US$3000 are almost 1.5x apart, too. So what's the advantage to making a D400 FX? Lenses. Indeed, the "where are the DX wide angles" question continues to be an interesting one. One might leap to say that this is more evidence that the DX line might stop at the D7000 point: someone who pays US$1600-2000 for a DX body is going to want lenses that don't exist. But those lenses do exist in FX.

I still think a D400 could go either way and is more likely to be DX, but given Nikon's recent aggressive push, I can't rule out an FX D400, thus what I wrote in the next article. The new US$2200 pricing on the D700 just throws another wrinkle into the mix.

[falconeye]

Interesting commentary and speculation on FF, the upgrade path... and the beginning of the squeeze from above on aps-c, perhaps?

The beginning was two years ago.

[Pål Jensen]

If I were to paraphrase what I feel like the outcome of the paper is, it would be:
We don't typically compare equivalent cameras. Here's what will happen if we do (we see big differences in ISO and aperture, among other things). Because of economic and design constraints, we still won't be able to compare equivalent systems most of the time..

The whole thing is folly because formats aren't equivalent. Trying to making them so is nonsense. Eg. I could say that the DA 560/5.6 is the equivalent of a 840/5.6 on FF. Then someone could say that it isn't really equivalent cause they don't have exactly the same DOF at the same numerical apertrure! As if anyone cares! :ugh:

No one has ever bought into a lens or camera with the demand that DOF is exactly 5.6 cm at 1.5 metres at a certain magnification and aperture value!