[rmtschanz]

For anyone who's considering the K-3 for purely live view mode, please consider that it is a tad slower than the K-01 when using live view. Is it considerably slower? No, but noticeable if you've used them both. test notes below.

Live View Test:
-Shooting a white wall 10ft away; the room has natural sunlight coming in. This setup forces the lens to zoom to each distance extreme before obtaining focus.
-Pentax FA 50mm f/1.4 lens.
-I start the stop watch when I press the trigger, and stop it when the live view point turns green (indicating focus lock).
-This was averaged out over about 20 shots for each camera. I say "about" because sometimes it would not lock focus.

K-01 will obtain focus time in 1.6 seconds fastest, 2.1 seconds slowest
K-3 obtained focus time 1.8 seconds fastest, 2.2 seconds slowest

---

Last edited by rmtschanz; 03-29-2015 at 04:35 PM.

[Adam]

Hate to play devil's advocate, but:

For anyone who's considering the K-3 for purely live view mode, please consider that it is a tad slower than the K-01 when using live view.

I don't think this conclusion is correct.

-Shooting a white wall 10ft away; the room has natural sunlight coming in. This setup forces the lens to zoom to each distance extreme before obtaining focus.

This isn't a realistic scenario; you need to use a contrasty target that can reliably focused upon and start from minimum and infinity (in addition to another initial distance if desired). Since the K-3 has a more powerful screwdrive motor, it will be able to turn screwdrive lenses with a longer focus throw much more quickly. In this case, since the K-3 turned the lens faster, it's possible that it was missing the mark due to the already-low contrast. If you'd like to level the playing field and look at the focusing algorithms alone, I'd recommend using an SDM/DC/HSM lens.

Also, where you in spot AF or auto AF? CDAF speed is greatly increased in auto mode when multiple AF areas are used at the expense of a little bit of guesswork by the camera.

[stevebrot]

-Shooting a white wall 10ft away

With a white wall, you are luck that either camera was able to manage a lock.


Steve

[stevebrot]

K-01 will obtain focus time in 1.6-2.1 seconds, on average
K-3 obtained focus time 1.8-2.2 seconds, on average

I hate to be persnickety, but a range is not the same as an average. Average (arithmetic mean) would be the sum of all the measurements divided by the number of measurement, one number. The other consideration would be the scatter around that mean (variance). If the variance is greater for the K-01 (which it appears to be), there is a good chance that there is no statistical difference between the two sets of times.


Steve

[Johndav]

Given that the K-01 only have Liveview I would hope it excelled, so I can't imagine K-3 owners losing too much sleep over this

[shaolen]

yay I love the K-01

[clackers]

At a cursory glance, it can't be stated that there is any advantage when the mean difference is 0.1s or thereabouts with a much larger variation of about 0.4s.

This is like someone complaining a coin is loaded because they just threw three tails in a row.

You *could* attempt some academic rigour by doing an audio recording of the lenses focusing and getting the exact times from looking at the file in Audacity. The stopwatch suffers first of all from about 0.1s of human reaction time plus the latency in the action.

You'd also provide a contrasty target ... this is Contrast Detection being measured, right?

---

Last edited by clackers; 03-28-2015 at 03:23 PM.

[rawr]

I'll just leave this here (from a Nikon manual). See the first point

[stevebrot]

I'll just leave this here (from a Nikon manual). See the first point

I do wish Pentax would put this little cheat sheet in with each camera they sell. It sets the ground rules and tempers expectation. Thanks for sharing.


Steve

[rawr]

I do wish Pentax would put this little cheat sheet in with each camera they sell.

Nikon put it everywhere - it's in the manual for the D4, D810, D750, D610, D7100 etc.

Poor Nikon - it must have some real problems with its AF if it needs to include such warnings with every camera

[rmtschanz]

All good points taken. A bit of background to help understand why I did this. My wife prefers using the K-01 because it's easy. Just turn it on, look at the screen, and snap a pic. I've tried convincing her to use the K-3, so we can sell the K-01. But she claims that using the K-3 is 'slower', so I wanted to prove or disprove her. I lost the battle, until more testing is done....

Yes, I suppose I should have said, "fastest/slowest" instead of "on average". I'll make that change to the OP. I like the idea of using audacity to get an accurate reading beyond my slow human hand . If I have some time in the next couple of months I'll redo the test in a more controlled manner.

Adam, good point about screwdrive vs HSM. If I had any HSM lenses I would redo the test. I'm using the center spot focus for CDAF. For whatever reason, the white wall forces the lens to go from infinity to minimum then back to 10ft where it locks focus. It's a sand textured wall, and it does lock focus at 10ft. Further, since CDAF always needs to refocus to lock focus, the lens is always starting at the 10ft mark, moving to infinity, moving to minimum, then back to 10ft.

The focus algorithm and attached lens need to work together to obtain a focus. If the K-3 has the fastest screwdrive motor to date, how am I able to focus slightly faster with the K-01 using a screwdrive lens? A couple of scenarios I conjured up to explain: 1.) the K-01 uses a similar motor as the K-3. 2.) The CDAF algorithm changed slightly on the K-3 (to increase focus accuracy or something). or 3.) The K-3 slows the screwdrive motor while in CDAF mode.

[IchabodCrane]

All good points taken. A bit of background to help understand why I did this. My wife prefers using the K-01 because it's easy. Just turn it on, look at the screen, and snap a pic. I've tried convincing her to use the K-3, so we can sell the K-01. But she claims that using the K-3 is 'slower', so I wanted to prove or disprove her. I lost the battle, until more testing is done....

Yes, I suppose I should have said, "fastest/slowest" instead of "on average". I'll make that change to the OP. I like the idea of using audacity to get an accurate reading beyond my slow human hand . If I have some time in the next couple of months I'll redo the test in a more controlled manner.

Adam, good point about screwdrive vs HSM. If I had any HSM lenses I would redo the test. I'm using the center spot focus for CDAF. For whatever reason, the white wall forces the lens to go from infinity to minimum then back to 10ft where it locks focus. It's a sand textured wall, and it does lock focus at 10ft. Further, since CDAF always needs to refocus to lock focus, the lens is always starting at the 10ft mark, moving to infinity, moving to minimum, then back to 10ft.

The focus algorithm and attached lens need to work together to obtain a focus. If the K-3 has the fastest screwdrive motor to date, how am I able to focus slightly faster with the K-01 using a screwdrive lens? A couple of scenarios I conjured up to explain: 1.) the K-01 uses a similar motor as the K-3. 2.) The CDAF algorithm changed slightly on the K-3 (to increase focus accuracy or something). or 3.) The K-3 slows the screwdrive motor while in CDAF mode.

I think you've confounded everyone who's chimed in so far with the white wall test. Surely your wife is not taking pictures of white walls, is she? To compare in a meaningful manner, why not use a set of subjects that your wife would likely make a photograph of? From there, you'll need to configure the K-01 and K-3 to use the same CDAF parameters (Auto area as Adam suggested would be good). Lastly, would your wife use LV/CDAF on a K-3 anyway?

[DDoram]

In my experience the K3 LV AF is faster than the K01. The K01 AF is much better with the latest firmware.

[rmtschanz]

I think you've confounded everyone who's chimed in so far with the white wall test. Surely your wife is not taking pictures of white walls, is she? To compare in a meaningful manner, why not use a set of subjects that your wife would likely make a photograph of? From there, you'll need to configure the K-01 and K-3 to use the same CDAF parameters (Auto area as Adam suggested would be good). Lastly, would your wife use LV/CDAF on a K-3 anyway?

True, my wife doesn't take pictures of white walls. To make an analogy, people don't take pictures of resolution charts, but that does not mean they are not useful in objectively measuring the performance of a lens. The white wall was a quick and repeatable process. When I do the audacity sound recordings next month, I'll use a high contrast focus subject, maybe a resolution chart...

As to the K-3 LV general usage, it's only a matter of pushing a button to get into LV mode. From there the K-01 and K-3 are similar when used in the Green Auto mode. However, the K-3 dual SD slots used in conjunction with the O-FC1 flucard is a definite useability advantage over the K-01. When we take pictures of friends and family, I have it set for a small 2MP image to store on the flucard (slot 2), and the RAW saves on the other SD card (slot 1). Using the Flucard app, we can quickly download the 2MP jpg and share it with friends via email or text. If it's a keeper shot, I still have the RAW file for processing later.

[stevebrot]

True, my wife doesn't take pictures of white walls. To make an analogy, people don't take pictures of resolution charts, but that does not mean they are not useful in objectively measuring the performance of a lens.

Your use of the white wall told you more about the failure rate for focusing on a featureless subject. What you ended up measuring was the corner case of AF time for successful attempts when AF is almost impossible.


Steve


Steve