Does A-DEP work reliably for you?

[steve_dunn2]

<p>I didn't use DEP mode much on my film bodies, but I found it worked

well when I did. Now I have a 20D, which only has the crippled A-DEP

mode, and it seems really flaky.</p>

 

<p>The manual implies that the camera will set the subject distance

and aperture so that anything covered by any of the 9 AF points will

be within DOF. But what I found, doing some playing with it, is that

often only <em>some</em> of the AF points seem to be used. I can

compose so that the AF points cover a wide range of distances,

half-depress the shutter release, and get f/5.6; do it again, without

recomposing (handheld, but with the AF points clearly still covering a

wide range of distances), and get f/32. I can get a virtually

unlimited number of different apertures if I keep going long enough.

If I take several pictures this way, there are huge differences in DOF

between pictures.</p>

 

<p>DEP, at least, was repeatable; if I picked the same near and far

points twice in a row, I'd get the same aperture twice in a row.

A-DEP doesn't seem to have this characteristic. Does anyone else find

this?</p>

[bobatkins]

A-DEP is pretty useless unless you're a total newbie who wouldn't know an f-stop from a hole in the ground.

 

For maximum DOF with acceptable sharpness, just stop down to f16.

 

DEP was much better. I don't know why they dropped it. I guess selecting focus twice was more than total newbies could handle.

[paul -]

A-DEP is so bad it should be on the other half of the dial with the other "Idiot Modes" like Portrait, Landscape, etc.<p>The few times I've tried it, after letting A-DEP select the focus and exposure, I took the usual forum advice of applying those settings in Manual mode (but adjusting the shutter/aperture to get the same exposure value at a narrower aperture than the reading, to be more certain of adequate DOF) and switching off AF before taking the shot. Too much hassle, when simply focusing 1/3 to 1/2 way into the desired areas of focus and stopping down the aperture in other modes usually works just as well.<p>As far as using the 9 AF points, the manual points out:<i><br>

"The camera uses the 9 AF points to detect the nearest and farthest subjects to be in focus"<br></i>but follows that with:<i><br>"All the subjects covered by the AF points <b>which flashed in red</b> will be in focus."</i><p>In other words, the camera tries find something to focus on in each of the 9 points, but only selects near and far focus points from those AF points which find something to lock onto, ignoring the rest.<p>The ever-changing exposure/aperture values are a mystery to me too.

[andrew robertson]

It's because each time you focus in A-DEP mode with the camera in the same spot, it

chooses a different set of focusing points and f/stop and shutter speeds.

 

I sometimes use it to figure out the minimum aperture I need in Av mode to cover the area I

need to be sharp. In that sense it's not totally useless.

[mark u]

Somehow things were easier with lenses that had a long focus travel, a good distance scale, and consequently well spaced DOF aperture markings that allowed you to work by focussing on each end of the field, and then moving focus so that the markings bracket the distances. Moreover, you could decide on your own sharpness criterion rather than relying on the one implicit in the markings, which was likely too slack for larger prints. I believe the criterion for those bodies that don't have a shiftable DEP mode (which allows you to adjust the aperture/shutter speed combination as if you were in P mode) is also based on a rather large circle of confusion (?35 microns

for full frame IIRC).

 

Another issue with DEP or A-DEP is that the focus point is set 7/17ths between the near and far points, which is a reasonable compromise when working with fairly close distances, but is perhaps less ideal as you approach hyperfocal conditions, when the rear DOF becomes infinite.

[steve_dunn2]

<cite>DEP was much better. I don't know why they dropped it.</cite>

 

<p>Well, Chuck Westfall gave us the answer to that: they needed to free up space in the firmware for other programs. I think they'd have been better off adding slightly more flash memory so they could keep DEP, but hey, I don't get to design the thing, nor do I get to decide whether the extra couple of bucks it would cost for the extra flash is worth it (obviously, my answer would be that it is, and Canon's answer is that it isn't).</p>

 

<p>I agree that DEP was so much better. I didn't use it much, but at least if you did the same thing five times in a row, you'd get the same DOF, or darn close to it, each time.</p>

 

<cite>In other words, the camera tries find something to focus on in each of the 9 points, but only selects near and far focus points from those AF points which find something to lock onto, ignoring the rest.</cite>

 

<p>Sure, and I understand what you're saying: if the camera can't figure out how to focus on one or more of the AF points, it ignores it when figuring out what things should and shouldn't be used to set DOF. But it picks different things each time. That's what puzzles me. If it used object X this time, why did it completely ignore it last time? If its behaviour is predictable and/or repeatable, I could deal with it.</p>

 

<cite>I sometimes use it to figure out the minimum aperture I need in Av mode to cover the area I need to be sharp.</cite>

 

<p>Which is exactly what it's for. But if you have to try it several times before it actually guesses which AF points you want it to light up, it's not that much use.</p>

 

<p>Thanks for the answers, folks. I knew I wasn't the only one who thought A-DEP was a big step backwards from DEP; that's been discussed here a few times. It's nice to see I'm also not the only one who finds it puzzling how the camera decides which AF points to use for it.</p>

[bobatkins]

<em>DEP was much better. I don't know why they dropped it. <p>

Well, Chuck Westfall gave us the answer to that: they needed to free up space in the firmware for other programs</em>

<p>

Much as I respect Chuck, and even accepting that as the true reason, it's a cop-out. Maybe "back in the day" when memory was more expensive it was the original reason. If Chuck says it was, then I'm sure it was.

<p>

However I don't see it as a reason not to put it back now. Memory is cheap. It's not like you need 100GB to add DEP back. A couple of hundred bytes of machine code is probably enough.

<p>

Perhaps they're saving it for the 30D or the 47D or whatever. Wouldn't that be a surprise!

<p> It never was perfect. It used the "7/17th rule". Given that the camera knows focus distance and aperture, they could actually calculate DOF with a lot better accuracy than that.

[jeff_conrad]

I got the same explanation from someone else in Chuck's group. I got the

feeling that many of the folks in Lake Success weren't too pleased with the

decision, but he did mention that relatively few people used DEP, which

probably is why the folks in Japan considered it expendable. Admittedly,

DEP is useful primarily for landscape and architectural work, and this

probably accounts for a small fraction of Canon DSLR users.

<p>

For years, I had reasonably good luck using DoF scales on MF lenses,

although the process could be slow, and I found the mental interpolation

between marked distances tedious. This all changed with the introduction

of AF. Most AF zooms lacked DoF scales, and the focusing ratios on most AF

primes were such that the DoF scales were almost impossible to use.

Moreover, the need to use MF to set DoF made AF useless. For a long time

Canon were the only manufacturer to offer the same ability to control DoF

with their AF lenses as one had with a $150 Pentax K1000. With the demise

of DEP, no DSLR affords a systematic means for controlling DoF. Something

is wrong when new technology gives less capability than previously was

available.

<p>

I agree with Bob that, given that most current EF lenses transmit distance

information, Canon probably could have improved the "7/17" rule so that DEP

would be usable for closeup work as well. Even as it was, I found that DEP

worked as well as I could do using lens DoF scales on MF lenses, and it was

much easier and faster. When first introduced, DEP was much derided as an

"amateur" feature, although the true rank amateurs were the ones who did

not understand why DEP was needed--the same ones who probably never learned

how to operate the DoF scales on MF lenses. If setting focus and f-number

from the focus spread were a dumb idea, it would not remain standard

practice for most LF users.

<p>

The absence of DEP on the EOS-1D Mk II was a bad sign, and the continued

absence on three subsequent models suggests that without a massive protest,

it's not likely to come back in the near future. At the time of the 1D Mk

II's introduction, I sent Canon a detailed write-up on why DEP should be

resurrected, but unless they get thousands more, I don't think anyone will

pay much attention.

[steve_dunn2]

<p>I think the answer is genuine but I disagree with the rationale behind it (which is not Chuck's responsibility but rather the responsibility of the engineers and project managers and marketers and bean counters back in Japan), as others do. IIRC, the EOS 3 can have its firmware reprogrammed by a Canon service centre using some kind of hookup to the hot shoe, so we can deduce that its firmware is stored in flash. Flash was vastly more expensive several years ago, when the 3 came out, yet they didn't feel it necessary to excise this feature to shave cost. Flash is much cheaper now and it surely wouldn't add much to the cost of, say, a 20D to give it enough flash to handle true DEP.</p>

 

<p>We do have to recognize the reality of Chuck's employment with Canon. Yes, his job was to give us technical information. But his job was also to be loyal to Canon. So if he had a personal opinion that this, or any other design decision, was dumb, one would expect him to explain the decision in an appropriate way for a Canon employee, and keep his personal opinion to himself.</p>

 

<p>There have been reports that the distance scales in some of the lenses which report this information are rather lacking in precision. Just to make up numbers out of thin air, a lens may be able to tell 2m from 3m but have nothing between them. That would render distance information useless for DOF calculations, and perhaps explains why E-TTL II reportedly uses distance information more as a sanity check than as a main part of its calculations. Hopefully, now that the bodies actually use the distance info for something, newer lenses will have much higher-precision distance encoding hardware (along with the capability, if it doesn't already exist in the protocol, to report not only the distance but how precisely it has been measured), and maybe if Canon ever sees the DEP light and resurrects it, they will use distance information from such lenses. But I'm not holding my breath.</p>

[lester_wareham]

Its rubbish. Go to a spreadsheet and run some hyperfocal distance tables off.

 

I have noticed that although the primes have a DOF scale its often not that useful. The 50mm f1.4 only has f22 or something.

 

I guess this is because of the limited rotation under AF. Canon had no problem engraving a decent DOF scale on the old FD lenses.

[ted_marcus1]

Answers about lack of memory in newer cameras never made sense to me. The very first EOS camera (the 650, introduced in 1987) included DEP mode; that was back in the days when processors were slow and memory was very expensive. I just can't believe that A-DEP requires less code or less processing than DEP.

 

I used DEP quite a bit on the 650 and the Elan II. With my Rebel XT, I can use A-DEP but it's not convenient at all. The best way I've found is to twist and turn the camera so that a focus point is over both the near and far point of the image. Then I let it set the focus, after making sure that the points both light up. If they don't, I twist the camera and try again. Once it's set the focus, I switch the lens to manual and stop down (if it's not already at f/16 or smaller). Not convenient, but (somewhat) better than depth of field tables or the "DOFMaster LE" on my PDA-- the distance markings on my lenses are too sparse to make accurate manual settings. I relied on a similar trick with the old DEP mode using wide angle lenses, whose depth of field was such that the camera would often choose a wide-open aperture.

 

If you're one of those supposedly rare individuals who frequently used DEP on "legacy" cameras, it's probably a good idea to write to Canon and complain about A-DEP. Canon's executives and marketeers have made assumptions what customers will use; unless customers inform then otherwise they'll just continue to rely on their assumptions. Unless, of course, those executives are so arrogant that they believe they Always Know Best regardless of how many contrary complaints they receive (in which case their loyal gatekeepers will make sure the complaints never get anywhere near them).