RecliffD hat geschrieben:DSLR´s haben ca. eine Verzögerung von gerade mal 0,002 s!
Das ist leider falsch, denn das wären 2ms. Richtig wäre, dass eine sehr schnelle DSLR etwa bei 40-60ms liegt, eine Trägere würde etwa 90ms und mehr brauchen.
Es gibt deutliche Unterschiede. Am besten mal ein paar DSLRs ausprobieren mit AF und Stabi ausgeschaltet und bei iso1600 damit man eine schnelle Verschlusszeit bekommt, und dann auf jpg höchste Quali, dann eine schnelle Speicherkarte rein, und dann mal einfach ein wenig rumballern. Wirst Dich wundern. Ich würde die Reihenfolge mal so definieren (die Olys kenne ich zu wenig):
lahm....
Pentax istD
Dynax 5D
Alpha 100
Pentax istDs
----------
mittelmässig, auch ohne AF:
Pentax K100D
Nikon D50
Nikon D40
Nikon D80
Nikon D70(s)
================ ab hier machts Spass
schnell, auch mit AF:
Canon 350D
Canon 400D
Pentax K10D
Canon 30D
Canon 40D
Fuji S5 pro
Nikon D200
Nikon D300
--------
sehr schnell, auch mit AF
Nikon D1x
Nikon D2xs
Canon 1D Mk III
Nikon D3
Wenn man Tests liest mit ganz genauen Millisekundenangaben mit eingeschaltetem AF: Das ist Schwachsinn. Die Unterschiede sind viel grösser, je nachdem wie und mit welcher Linse man fotografiert. Wer es eilig hat mit dem Abdrücken der wird erst unterhalb meiner Mittellinie wirklich glücklich werden, wer nicht der kann drauf verzichten.
hier mal ne Site zum Vergleichen, hab die eben erst entdeckt, aber ich sehe schon dass ich damit nicht konform gehe. Die Angaben sagen praktisch nichts aus was man in der Praxis bestätigen könnte:
http://www.impulseadventure.com/photo/shutter-lag.html
Also eine Cybershot mit angeblich 9ms ist sicherlich nicht schneller als eine Nikon D2hs mit 37ms (shutter lag) und eine Canon 10D auch nicht schneller als eine 1D MkIII (total lag). Und eine Cybershot ist auch mitnichten schneller beim AF-lag als eine D1x. Das ist schon etwas peinlich sowas zu veröffentlichen. Aber so sind sie halt, die Messtechniker. Sollten auch mal rausgehen und Sport fotografieren
weiter unten hat sich auch ein Skeptiker gemeldet, der das genauer beschreibt was ich meine:
Roger
I have looked at your chart and the numbers do not agree with side by side real-world experience. For example, your chart indicates the Canon 10D, 20D and Rebel XT have faster total shutter lag than a ID Mark II. I have a Canon 10D, 20D and 1D Mark II and I can assure you there is a world of difference, with the 1D Mark II being impressively faster than any of these cameras.
I looked at your references on some cameras, and see that you used data from http://www.imaging-resource.com, and while looking at that data the description seems reasonable, the data in fact must be in error. One possibility is that the lens used in the tests (not an L lens) is a limiting factor, not the camera.
Example: the 1D Mark II is astoundingly fast compared to the 10D using the same lenses on the same subject, (comparing multiple lenses). Currently, I do wildlife photography with a 1D Mark II with a 10D as a backup, often switching lenses (e.g. 500 mm f/$ L IS on a tripod and 300 mm f/4 L IS hand held). To rate the 10D total lag at 0.189 second and the 1D Mark II at 0.235 second is just plain wrong, unless you put the fastest autofocusing lens on the 10D and the slowest possible lens on the 1D Mark II. I shoot a lot of wildlife action, and the response time on the 1D Mark II is well under 0.1 second (total lag time in my experience in multiple shooting conditions). The 10D feels like a slow point and shoot in comparison.
In real world action conditions, I would have two problems with the 10D: 1) erratic action (e.g. bird in flight) with a complex background (e.g. distant trees) has trouble locking onto the subject and not the background, and 2) while tracking a subject, if the focus point is moved off the subject (e.g. due to my inability to track erratic movement), the camera would never regain focus until the subject stopped.
On the 1D Mark II, I don't have these problems. Reports from people in the field say the 20D has the same problems as the 10D. But with the 1D Mark II, I can lose and reacquire the focus point on a moving subject in what seems like well under 0.1 second. The focus accuracy is much better on the 1DII also (having had more than 50% out of focus action shots on a 10D, almost all in good focus with the 1DII) with typical large birds in flight (e.g. eagles, cranes, egrets).
So, your table http://www.impulseadventure.com/photo/shutter-lag.html/ is highly suspect, regardless of the source of the data.
Roger (Photos at: http://www.clarkvision.com)
ANTWORT:
Hi Roger -- Thank you for some excellent insight into performance outside of what the numbers would seem to imply. I think the only fair measurement for comparison purposes may be the shutter lag, not total lag as derived from an average of test results. This is the reason that I list the resulting lower and upper bounds along with the average to show the deviation. Including the autofocus in the total lag time is highly dependent upon AF mode, lens selection and scene contrast as you rightly point out!
You'll note that the review sites (3 of them, not just imaging-resource) all provided measurements in the range 230 to 240 ms for the Canon 1d Mk II. However, you can see that the results given for the 10D and 20D have a much wider spread, from 146 up to the 240 ms listed for the Canon 1D mk II. So, while the average might show as 189, the fair comparison value might be the 240 ms (identical with the Canon 1D mkII).
I don't necessarily agree that the numbers are incorrect. Taking the largest total lag times in the comparison would yield similar results, but they are likely captured in a very synthetic setting (high-contrast static target in a well-lit environment). Of course your real-world experience would highlight the weakness in the 10D and 20D autofocus times, where motion servo tracking and low contrast conditions may really degrade the 10D or 20D's best-case performance. That being said, the shutter lag is indeed reported as being significantly faster on the 1d Mark II when compared to the other prosumer cameras.
Ideally, we would have an objective, reproducable comparison of total shutter lags with the same setup and lens, but with typical real-world scenarios! But this is out of the scope of the setup for most camera reviewers. They would need a mechanized rig that could reproduce the motion, with some digital counter in the image scene (to assess the real delay accurately) and a remote release trigger. If someone set this up for most of the major dSLRs and with comparable lenses (e.g. L on Canon), this would be fantastic. Unfortunately, I think we'll have to settle for a sampling of results from different reviewers, all with slightly varying setups. The hope is that with enough reviews and results, the averages will be useful as a rough starting point of comparison.
Again, thank you very much for providing some real-world insight into the performance you've observed between the cameras. This feedback is often much more useful than another synthetic data point! Great gallery, BTW!
Letztendlich verteidigt der Autor der Seite seine Zahlen als solche, gibt aber zu dass sie in der Praxis nicht haltbar sind.
Fazit: Nett zu lesen, aber reine Zeitverschwendung wenn man wissen möchte ob eine Cam schnell auslöst oder ne lahme Ente ist.
fibbo