Thoughts About Full Frame

[Paulo Bizarro]

A never-ending discussion topic. Personally, it is only cost that prevents me from getting a full-frame DSLR from the EOS line. With a smaller sensor camera, my 24 TS lens would become useless, and my 135 f/2 would become useless.

I wonder how much a Digi Rebel fitted with a full-frame sensor would cost? I don't need full-frame sensors in shooting machines like the 1 series, but I would welcome full-frame sensors in lower tier cameras. Even then, perhaps they would remain too expensive...

In the meantime, I am quite happy with Velvia 100F.

[mikepak]

Here is my opinion on this.

With a given manufacturing technology, it always cost more to produce 35mm size sensor versus APS size sensor.  It always was in the past, it is now, and it always will be.  Even if the semiconductor process yield reaches 100%, the 35mm size sensor will cost 2X more (just from area alone).  This feat (100% yield) is rarely reached by even best manufacturers.  Typically, 35mm size sensor will cost any where between 3X to 9X more than APS size sensor (this estimation can be derived from a given probability of defect density).

So, given the higher cost of manufacturing the 35mm size sensor, which Camera makers would ever put the 35mm size sensor into their entry line?  It simply will not happen.  So, I see that consumers willl always get the APS size sensor on their Camera, while the 35mm size sensor will strictly be reserved to Pro-line camera ($3000 and up).

[didger]

Whew!!! This has gotten way over my head.  I've noted in the countless past instances when there's been pages of very technical/mathematical discussions to try to prove some point or another, no consensus was ever reached.  Everybody keeps believing whatever they believed to start with and they're sure that their math, etc. proves it.  I'd rather just have camera/lens combinations to test by shooting my test grid array and decide what's what by looking at the results.  "Probieren, probieren geht ueber studieren" is what my German engineer daddy often said.  It means (more or less) "experimentation rules over theorizing".

If a compact and light weight (small or large sensor) camera as good as my 1ds comes along at a tempting price I'll buy it, but not based on any theoretical discussions related to the technology involved.  I'm a dummy, ya gotta show me.

[didger]

This important COST factor will always drawn line between professional market and consumer market.

I can't imagine anyone seriously contesting this.  The issue initially was about quality issues, and reducing costs is not the same thing as increasing quality.  
As for the idea that small sensor cameras will be rivaling large sensor, this is surely a bit speculative.  Tomorrow's small sensor cameras may rival today's large sensor, but not tomorrow's large sensor.  Canon is obviously committed to sustaining the growth of large sensor cameras.  Of course the highest quality will always be most expensive and mainly for the pro market.  That's always been true for film and can't help but always be true for digital.

[EAD]

I mean by this that full frame and MF backs will remain a Pros-wealthy amateurs thing and I dont see the prices dropping considerably in the near future. It still amazes me  the amount of pros that worked with 35 mm film cameras and produced very good results( probably moved by their adventorous back-packing/helicopter-hanging/mountain climbing needs). Look Jean Arthus-Bertrand´s world from the air shot nearly exclusively in canon 35mm. APS sized sensors are capable of an amazing quality right now and will probably improve considerably in the next couple years,  allowing a lot of proffesionals to work with them proficiently. I agree that larger sensors will ALLWAYS give better image quality, but I have the feeling it is going to be much more interesting/profitable for the big companies to concentrate in getting the best out of smaller sensors. APS size will (does)allow enough quality for making it questionable the extra cost for FF for many pros allready. I think the landscape of the Photography market will change considerably in the years to come and APS size will be strengthened and maybe even favoured...

Again, what can you buy with 2cents???

[Gary Ferguson]

The articles on "Luminous Landscape" regarding full frame sensors are fascinating. As someone who switched from a Nikon D1x to a Canon 1DS I guess I'm in the full frame camp.

But I'm puzzled by one of the rebuttal points.

Michael, you said there's no quality advantage with a smaller sensor. I'd agree with respect to the sensor itself, however I'd have thought that a smaller sensor enables a lens with a smaller image circle, and wouldn't this promote better image quality?

I've recently been looking at the performance of the new generation of "digital" large format lenses designed to cover the 645 and 6x9 formats, they're certainly delivering a useful quality advantage over my modern 4x5 lenses by virtue of more restricted image circles. And at the other end of the scale I recall one of the key justifications for Minox sub-miniature cameras was that their tiny lenses were setting new performance highs, the Minox problem of course was that a wide range of film stocks to take advantage of this performance was never really available!

Incidentally Michael. you mentioned recently that you had trialled the P25 and would shortly publish a practical review. Any idea when it will appear? I'm expecting to take delivery of a P25 sometime in November. As I could only try out the H25 before placing an order I'm eager to get some second opinions from outside a studio!

[Scott_H]

I think Michael's rebuttal is primarily focused on image quality, which is probably his primary consideration.  I think there are other considerations though.  

It is going to be a question of volume and economies of scale.  Not everyone needs the image quality that a full frame sensor will provide, and not everyone will be willing to pay a premium for it.  There always will be a premium, even if the gap is getting smaller.  The computing power, and memory card sizes to deal with higher pixel counts add to the premium, something a lot of people don't account for when they are looking at higher resolution cameras.

If image quality were the only consideration, then 35 mm would not be so popular.

[didger]

Personally, it is only cost that prevents me from getting a full-frame DSLR from the EOS line.

A candid admission.  I think a lot of the folks that claim there are other reasons not to get a full frame DSLR are sour graping and twisting logic to the breaking point.  

Since almost all my shooting is on strenuous backpacking trips, I hope someday that hi res full frame sensors will be available for something much lighter than what's available now.  I still hope (with waning expectations) that some day there will be really good ultrawide zooms.  In the meantime I'll continue to be happy carrying my 1ds and 5 prime lenses on my backpacking trips; well, sort of happy.  My back and my old climbing injuries are not always so happy with this abuse.  Now something like a Digital Rebel and the new Canon 17-85mm IS lens, but with full frame sensor and optical quality to match; I'd probably sell my soul for that.  My gear weight for the most extreme backpacking trips could be reduced to 10% of what it is now.

[BJL]

Having more carefully read the whole "two Michaels" debate, I have a couple of comments.

There seems to be no disagreement that larger formats like 645 will be the choice for some niche of high end digital photography, just as larger than 35mm film format cameras survive now, albeit greatly outnumbered by 35mm film SLRs even in professional usage.

MJ seems to be addressing where the SLR mainstream will settle (forget the overall mainstream, which is solidly settled somewhere below 2/3" format, and maybe heading towards the camera phone.)

I agree with MJ that lenses will likely be the persistent "cost and convenience" factor causing smaller format DSLRs from 4/3 through EF-S up to DX to be far more numerous than digital cameras in 35mm format and up.

However, as I have said before, the fairest comparison is at equal effective aperture diameter, not equal aperture ratio: 200/2 for DX vs 300/2.8 for 35mm, 300/2.8 vs 400/4, 400/2.8 vs 600/4.

Such comparison gives equal DOF wide open, and about equal speed for a given noise level once the speed/noise advantage of a larger sensor is accounted for. The price, weight and size comparisons still seem to favor the smaller format with these shorter but "brighter" lenses, but not as nearly as much as if one compares at equal maximum aperture ratios.

Matching lenses this way does elimimate at least a number of common arguments, including one that MR repeats, based on the lower noise of larger sensors at equal ISO speed settings. For example, if when going from 35mm to true APS-C format (a factor of 1.4x) you choose lenses with focal length and maximum aperture ratio both reduced by that factor of 1.4x, the APS-C lens gives
- about the same or lower price and weight
- the same DOF wide open, and in general the same DOF at one stop faster
-  twice the exposure level at the same shutter speed and DOF (or same shuter speed and wide open), allowing you to use half the ISO speed setting unles you ar already at minimum. This at least potentially balances the higher noise levels seen if you compare at equal ISO speeds.

In more detail, with the same pixel count in both sensors and hence pixels of half the area in APS-C, this doubling of "lens speed" balances the halving of pixel area and hence gives the same amount of light gathered at each pixel. This seems likely to give similar noise levels.

The main potential disadvantage for the smaller sensor is blown highlights. Not a problem when using elevated ISO settings (underexposing the sensor and then amplifying more), but it gives larger sensors a possible advantage in dynamic range in situations where the minimum ISO setting can be used.

A final disagreement with MR; only a few new lenses are needed when changing to an APS-C format DSLR. It should be clear by now that there is no need for Nikon, Pentax, Konica-Minolta etc. to introduce entire new lens lines for smaller DSLR formats, or for lens owners to replace their whole 35mm format lens collections.

The only new lenses needed are those reaching wide angle focal lengths (below about 50mm). Nikon has already produced most of the DX lenses they need; most photographers only need to add one or two lenses to their existing collection, maybe trading away one or two 35mm "wides". Specifically, the majority of current Nikon 35mm film SLR users probably only need to add the 18-70 f/3.5-4.5 DX when changing to a Nikon DSLR.

[Scott_H]

So, given the higher cost of manufacturing the 35mm size sensor, which Camera makers would ever put the 35mm size sensor into their entry line?  It simply will not happen.  So, I see that consumers willl always get the APS size sensor on their Camera, while the 35mm size sensor will strictly be reserved to Pro-line camera ($3000 and up).

This is the point to me.  A larger sensor will mean better image quality, but will it be worth the premium in price.  Especially considering the needs of the average consumer.

I don't think this site is neccesarily a good sample space when it comes to the needs of the average consumer.  I consider myself a fairly involved amature, and I would have a very difficult time justifying (rationalizing would probably be more accurate) the purchase of a $3k full frame camera (if one existed).

The question is whether or not the cost differential is small enough to be justified by improvements in image quality that most people will not even be able to see.

[BJL]

> As soon as a proportionately lower aperture ratio is available on the shorter focal length lens used with the smaller format, you can get the same DOF as in the larger format, but with a "faster" f-stop

That is where the aberration problem creeps in.

Agreed: the lower limit on usable aperture ratios sets an upper limit on performance in a given format, and that limit seems to be about f/1.4. I even note a trend towards a f/2 limit. Aperture ratio is similarly limited by sensor characteristics; beyond f/1.4, light reaches the sensor in a very broad cone, and light from near the edges of that cone comes in at such an off-perpendicular angle that the sensor cannot detect it well.

So, for people who still want extremely large apertures, for extremely shallow DOF or extemely high shutter speeds, even with the far higher ISO sensor speeds possible with digital cameras compared to film, larger formats will continue to be attractive. But this is unlikely to influence the DSLR mainstream; look at the relatively small apertures being offered on all recent DSLR oriented lenses like the Canon EF-S, Nikon DX and Pentax DA models. The Nikon 200/2 and Olympus 50/2 and 150/2 and the 7-50 f/2-2.8 lens of the Sony 828 are the "fastest" recent lenses, and they are still only f/2.

[BJL]

Didger,

actually there seems to be no disagreement about the facts, just a question of which facts one considers more relevant.

No-one seems to disagree with the claim that, if you use sensors of the same pixel count, camera at the same distance from the subject, lens giving the same angular field of view, and the same shutter speed, but the sensors are of different sizes (so that the focal lengths are different),

a) if you use the same exposure index ("ISO setting") with each sensor, and hence the same aperture ratio, the larger sensor will have less noise (due to the greater light gathering power of the larger pixels) and less depth of field (due to the longer focal length)

 if you choose the apertures in each case to get the same depth of field, the larger format will need to use a higher exposure index, so the noise level comparison becomes far less clear.

Rough theory suggests that noise levels will be about equal, but as you say, it is better to do experiments at this stage.


To me, shutter speed and depth of field are direct characteristics of the image, whereas exposure index is not, so comparison ( seems of far more practical relevance.


Note well: I am not claiming any advantage in this repect for smaller sensors, just arguing that a claimed advantage for larger sensors is of little practical significance. In that sense, I agree with Michael; the main advantage of smaller formats is cost and convenience, while larger formats have some advantages, like more resolution at a given dynamic range.


Frankly, I think this is yet another of those hangups where people are used to comparisons of different options when working with the same format, and do not take account of all the things that change when trying to make a relevant comparison between different formats. Most of these issues come down to comparing at equal aperture ratio even though image properties are very different at equal aperture ratio in different formats.

[mikepak]

I have new evidence to prove my case.
Please check this out.

http://forums.dpreview.com/forums....0854751

[drew]

MF digital as represented by FF and above sensors will grow to the same size as the old MF film market. Digital backs will dissapear in favour of fully integrated cameras which will either take lenses for 35mm (1Ds MKII, DCS Pro N/C etc) or MF (Mamiya ZD, Hasselblad HD).
4/3rds and APS-C sensor cameras will take over from the quality 35mm film market and this is where the bulk of sales will be for the foreseeable future. Just like it used to be for 35mm, these cameras represent the ideal compromise between image quality and camera size and weight. Compatibility with existing lens technology is a bonus.

[didger]

I believe whe are going to have APS size for quite a while.

Yes, no one is arguing this.  We've had 35mm film, MF film, and large format for quite a while too.  It's not a question of APS vs full frame 35mm.  They can and will co-exist in exactly the same way that 35mm and MF have co-existed, with similar price and customer base profiles.

[Dan Sroka]

We've had 35mm film, MF film, and large format for quite a while too.  It's not a question of APS vs full frame 35mm.  They can and will co-exist in exactly the same way that 35mm and MF have co-existed, with similar price and customer base profiles.

Just been lurking on this topic, and I believe that your statement here sums it up. The discussions about "quality" are rather moot, since "quality" is a highly relative term, dependent on the style of photography you do. A landscape photographer may value pixel-count and accuracy
to capture the integrity of the scene. A street photographer may value speed and stealth. (I met a very cool photog who uses the camera in his cell phone to make some amazing images... yes, blown up to 20x24). A fine art photographer can fit anywhere along the spectrum. To make the claim that one camera is inherently "better" incorrectly assumes that all photographers value exactly the same kind of tool.

Personally, I hope both full-frame and small-frame continue to exist alongside each other.

There's nothing inherent in a smaller image circle that produces higher image quality. The trade off is, of course, the need for greater magnification.

I ordered a P25 in Contax mount last week. It should be here within about 2-3 weeks. I'll have a full field report not long after it arrives.

Michael

[didger]

If image quality were the only consideration, then 35 mm would not be so popular.

Well, cheap point and shoot cameras are even more popular than 35mm and most popular of all are disposable 35mm cameras like you can buy in convenience stores.  So what?  I think it's fairly safe to assume that on a landscape forum with a lot of very serious amateurs as well as semi-pro and full on pro members, it's fairly safe to assume that almost all of us want the best image quality we can possible afford.  I know I do.  Image quality is THE consideration.  Economic factors are hurdles to overcome in the eternal quest for improving image quality.

[BJL]

I can just one rather indirect way that a smaller sensor can allow better image quality, and that is when we are faced by by price and weight limits on the kit we can own or carry.

Downsizing the focal lengths and image circle to get the same field of view with a smaller format might allow increased resoluton in the sense of line pairs per millimeter, but no more than in proportion to image size, so I see no likely gain in "line pairs per image height", which is what counts for print resolution.

However, that smaller format lens can need less of the rather expensive optical glass to make it, reducing the price and weight. For example, the new Nikon 200mm f/2 seems intended to do with DX format what a 300mm f/2.8 does in 35mm format, but is set to be less expensive, even
with its extra stop.

That gives us budget limited users of smaller formats the chance to put those savings into higher quality lenses, faster lenses, or more lenses, possibly improving the quality of some of our images.

Also, when weight is limited, the smaller format can allow carrying more lenses, or lenses of lower minimum aperture ratios (like 150/2 in 4/3 or 200/2 in DX vs 300/2.8 in 35mm), again possibly improving some images.

For example, Michael explained his choice of a 500/4 over a 600/4 in terms of weight: in an "APS" format, something like a 400/4 becomes an even lighter option, with "600mm equiv. FOV".

[Edward]

> However, as I have said before, the fairest comparison is at equal effective aperture diameter, not equal aperture ratio: 200/2 for DX vs 300/2.8 for 35mm, 300/2.8 vs 400/4, 400/2.8 vs 600/4.

Great point!  Might not matter for rocks and trees, but if you use selective DOF, it makes a huge difference.

The convergence of aberrations and diffraction as the sensor gets smaller is also an issue, but not yet critical at APS size.