Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test

Started Jul 6, 2016 | Discussions
howieb101 Contributing Member • Posts: 633
Re: AFC test was redone!

Rishi Sanyal wrote:

LightBug wrote:

Tatouzou wrote:

joergensmi wrote:

...

It can be clearly seen that the central AF-point has been selected and none of the upper ones.

For me it is clear now that this test fails due to faults by dpreview.

http://www.dpreview.com/reviews/pentax-k-1/6

There has been an update of the exiftools by Phil Harvey The version is 10.22 dated 07.07.2016 12:26 7.226.089 exiftool.exe

The result is now
Focus Mode : AF-C (Focus-priority)
AF Point Selected : Center; Single Point

I wonder why Rishi Sanyal didnt reply this post about the AF point he used: these are facts, and they show incoherences in his assessments of how he conducted the review.

First of all, I didn't write this review. Do you see my name on the author list?

There was a miscommunication about which set of images the author ended up including in the rollover we published (we had 20+ runs), which was ultimately irrelevant to the conclusions of these tests, because the center point was still over the bike, and the bike/cyclist were still out of focus in most of the shots, with focus far behind the entire bike (playing catch up).

All you've managed to do is nitpick something that has absolutely no relevance or bearing to the conclusion or review. But, out of respect for our audience, we redid it today again, to get the same exact result.

DPR obliged and re-did the straight-line AFC test. Results look much better than before, I could count 4-5 frames in focus, while previously I remember only 1 or 2 was in focus.

There is literally no difference in the results. We're publishing the rollovers side-by-side just for you. Why would the results be any different to begin with? As I'd noted before, Richard's face was within the DOF of what the AF points were focusing on (the bike in the last rollover prior to today), so the only reason we re-did the test was simply to avoid the nitpicking criticisms that we already knew had nothing to do with the end result. There is no change in the results.

-Rishi
-------------------------
Rishi Sanyal, Ph.D
Deputy Editor, Technical Editor | Digital Photography Review
dpreview.com (work) | rishi.photography (personal)

Within the depth of field isn't the same as the AF point no matter what you say.

If you say in the review you focussed on a persons face it implies you focus on their face. If you say you focussed on their singlet  people would expect that you focussed on their singlet.

Are people going to have to guess?

If the original text said the focus was on the cyclists singlet DPR wouldn't be getting all this feedback.

Rishi Sanyal
Rishi Sanyal dpreview Admin • Posts: 814
Re: AFC test was redone!
5

howieb101 wrote:

Rishi Sanyal wrote:

LightBug wrote:

Tatouzou wrote:

joergensmi wrote:

...

It can be clearly seen that the central AF-point has been selected and none of the upper ones.

For me it is clear now that this test fails due to faults by dpreview.

http://www.dpreview.com/reviews/pentax-k-1/6

There has been an update of the exiftools by Phil Harvey The version is 10.22 dated 07.07.2016 12:26 7.226.089 exiftool.exe

The result is now
Focus Mode : AF-C (Focus-priority)
AF Point Selected : Center; Single Point

I wonder why Rishi Sanyal didnt reply this post about the AF point he used: these are facts, and they show incoherences in his assessments of how he conducted the review.

First of all, I didn't write this review. Do you see my name on the author list?

There was a miscommunication about which set of images the author ended up including in the rollover we published (we had 20+ runs), which was ultimately irrelevant to the conclusions of these tests, because the center point was still over the bike, and the bike/cyclist were still out of focus in most of the shots, with focus far behind the entire bike (playing catch up).

All you've managed to do is nitpick something that has absolutely no relevance or bearing to the conclusion or review. But, out of respect for our audience, we redid it today again, to get the same exact result.

DPR obliged and re-did the straight-line AFC test. Results look much better than before, I could count 4-5 frames in focus, while previously I remember only 1 or 2 was in focus.

There is literally no difference in the results. We're publishing the rollovers side-by-side just for you. Why would the results be any different to begin with? As I'd noted before, Richard's face was within the DOF of what the AF points were focusing on (the bike in the last rollover prior to today), so the only reason we re-did the test was simply to avoid the nitpicking criticisms that we already knew had nothing to do with the end result. There is no change in the results.

-Rishi
-------------------------
Rishi Sanyal, Ph.D
Deputy Editor, Technical Editor | Digital Photography Review
dpreview.com (work) | rishi.photography (personal)

Within the depth of field isn't the same as the AF point no matter what you say.

No one's stopping you from viewing the full image, all of which we provide. If you examined them, you'd see the exact same story: the camera can't keep up with the subject in these runs no matter where the AF point is placed.

If you say in the review you focussed on a persons face it implies you focus on their face. If you say you focussed on their singlet people would expect that you focussed on their singlet.

Are people going to have to guess?

If the original text said the focus was on the cyclists singlet DPR wouldn't be getting all this feedback.

Yes we would. And we did. Have a look at our page where we collate all of the rollovers we published here .

That first run, shot on a much sunnier day, had the center point right over Sam's face. I know, because I shot that run. Still <50% hit-rate. But the initial complaints were that we didn't go in and change the AF Hold setting to 'Off'.

Out of a desire to make sure we'd covered all our bases and given the camera as much chance as possible, we shot sequences with AF Hold Off. There was a miscommunication here as to which point was used as I didn't shoot this run, but the end hit-rate was largely the same: below 50%.

After the new round of complaints that the face wasn't what was focused on (though what *was* focused on, under the center point, still wasn't focused and showed the same exact pattern of focus lagging behind we saw with the face, and in the previous run which *was* focused with a point over the face), we did everyone the favor of shooting the sequence yet again, despite the previous two runs (one of many, many runs since we do each test in at least triplicate) showing the exact same thing.

And we see the same behavior.

At this point, we have tried our best to get the most out of the system, yet see the same behavior we initially suspected having shot many sequences. The crusade to try and nitpick some fine point that supposedly invalidates all our tests may continue, but at this point we are more confident in our results than we ever have been, as the results now are exactly what we had, and shared, days/weeks before.

We're happy to be proven wrong, though, if you have convincing data - not irrelevant nitpicking - showing otherwise.
-------------------------
Rishi Sanyal, Ph.D
Technical Editor | Digital Photography Review
dpreview.com (work) | rishi.photography (personal)

 Rishi Sanyal's gear list:Rishi Sanyal's gear list
Sony RX100 IV Nikon D810 Canon MP-E 65mm f/2.5 1-5x Macro Nikon 85mm F1.8G Sigma 35mm F1.4 DG HSM Art +3 more
Rishi Sanyal
Rishi Sanyal dpreview Admin • Posts: 814
Re: AFC test was redone!
5

LightBug wrote:

Rishi Sanyal wrote:

There is literally no difference in the results. We're publishing the rollovers side-by-side just for you. Why would the results be any different to begin with? As I'd noted before, Richard's face was within the DOF of what the AF points were focusing on (the bike in the last rollover prior to today), so the only reason we re-did the test was simply to avoid the nitpicking criticisms that we already knew had nothing to do with the end result. There is no change in the results.

-Rishi
-------------------------
Rishi Sanyal, Ph.D
Deputy Editor, Technical Editor | Digital Photography Review
dpreview.com (work) | rishi.photography (personal)

Rishi,

I have to disagree with your statement above that redoing this test makes no difference. I do see a increase in hit rate from first test to second test.

Here are crops from the first straight-line AFC test:

http://www.dpreview.com/galleries/0725780996/photos/3477783/imgp2504?inalbum=dpr-k-1-afc-test-1-crops

From the first set, I would consider IMGP2504, IMGP2513 in focus. That's 2 out of 14 when the focus point was not set on the face.

Here are crops from the second straight-line AFC test that was redone:

http://www.dpreview.com/galleries/0725780996/albums/dpr-k-1-afc-test-2-crops

From the second set, I would consider IMGP2813, IMGP2814, IMGP2819, IMGP2821, IMGP2822 in focus. That's 5 out of 14 when we give K-1 a chance to actually focus on the face.

So I think the redo was worthwhile, don't you think?

Joey

No, not particularly, but I thought it was doing to address reader concerns. After all, this site is largely for you, our audience.

Here are all our results, collated. As you can see, our latest run is no better than the last, The last 2 runs had ~4 images in focus, with the latest one just having more in focus at the beginning because the set of 14 we chose started a bit earlier, so the cyclist hadn't moved much in the beginning (and therefore didn't require any refocusing on the camera's behalf to remain in-focus). Why did we choose a shifted sequence? Because the subsequent ones were very much out of focus.

Which is why we stress that we look for trends/patterns in these tests, not absolute numbers. 3/14 vs 4/14 is not a statistically significant difference - the error in this test is high.

If anything, the first test we originally published with the review had the best hit-rate, yet there were immediate complaints about that when we published (go figure?). Perhaps because of the extremely sunny weather? Though, again, we're looking at trends here, and even that initial rollover of 5/14 falls far behind what we expect from DSLRs, or even good mirrorless cameras, today.

Subject tracking consistently always gives up, leading to a 15% or lower hit-rate for a subject that's small enough relative to the small AF area to actually *require* some form of subject tracking to begin with (if your subject eats up most of the AF area to begin with, it won't require any subject tracking at all - which is partly why MightyMike had better results relative to ours).

I'll look into publishing a separate home page news story clarifying these points, and points of confusion, shortly. But if you think, after looking at all these results holistically, that there's some vast differences between any of these results that are indicative of our failures on any of these runs, I'd encourage you to think a little more carefully about this test, what it means, and what you can extrapolate from it.

-Rishi
-------------------------
Rishi Sanyal, Ph.D
Deputy Editor, Technical Editor | Digital Photography Review
dpreview.com (work) | rishi.photography (personal)

 Rishi Sanyal's gear list:Rishi Sanyal's gear list
Sony RX100 IV Nikon D810 Canon MP-E 65mm f/2.5 1-5x Macro Nikon 85mm F1.8G Sigma 35mm F1.4 DG HSM Art +3 more
MightyMike Forum Pro • Posts: 38,915
Re: AFC test was redone!
2

Rishi Sanyal wrote:

LightBug wrote:

Rishi Sanyal wrote:

There is literally no difference in the results. We're publishing the rollovers side-by-side just for you. Why would the results be any different to begin with? As I'd noted before, Richard's face was within the DOF of what the AF points were focusing on (the bike in the last rollover prior to today), so the only reason we re-did the test was simply to avoid the nitpicking criticisms that we already knew had nothing to do with the end result. There is no change in the results.

-Rishi
-------------------------
Rishi Sanyal, Ph.D
Deputy Editor, Technical Editor | Digital Photography Review
dpreview.com (work) | rishi.photography (personal)

Rishi,

I have to disagree with your statement above that redoing this test makes no difference. I do see a increase in hit rate from first test to second test.

Here are crops from the first straight-line AFC test:

http://www.dpreview.com/galleries/0725780996/photos/3477783/imgp2504?inalbum=dpr-k-1-afc-test-1-crops

From the first set, I would consider IMGP2504, IMGP2513 in focus. That's 2 out of 14 when the focus point was not set on the face.

Here are crops from the second straight-line AFC test that was redone:

http://www.dpreview.com/galleries/0725780996/albums/dpr-k-1-afc-test-2-crops

From the second set, I would consider IMGP2813, IMGP2814, IMGP2819, IMGP2821, IMGP2822 in focus. That's 5 out of 14 when we give K-1 a chance to actually focus on the face.

So I think the redo was worthwhile, don't you think?

Joey

No, not particularly, but I thought it was doing to address reader concerns. After all, this site is largely for you, our audience.

Here are all our results, collated. As you can see, our latest run is no better than the last, The last 2 runs had ~4 images in focus, with the latest one just having more in focus at the beginning because the set of 14 we chose started a bit earlier, so the cyclist hadn't moved much in the beginning (and therefore didn't require any refocusing on the camera's behalf to remain in-focus). Why did we choose a shifted sequence? Because the subsequent ones were very much out of focus.

Which is why we stress that we look for trends/patterns in these tests, not absolute numbers. 3/14 vs 4/14 is not a statistically significant difference - the error in this test is high.

If anything, the first test we originally published with the review had the best hit-rate, yet there were immediate complaints about that when we published (go figure?). Perhaps because of the extremely sunny weather? Though, again, we're looking at trends here, and even that initial rollover of 5/14 falls far behind what we expect from DSLRs, or even good mirrorless cameras, today.

Subject tracking consistently always gives up, leading to a 15% or lower hit-rate for a subject that's small enough relative to the small AF area to actually *require* some form of subject tracking to begin with (if your subject eats up most of the AF area to begin with, it won't require any subject tracking at all - which is partly why MightyMike had better results relative to ours).

I believe I've effectively proven this to be a rather incorrect assessment of my test, The camera was only using a single point to lock focus and then by design of the AF system and everything that is written in the companies literature the rest of the AF points followed the subject i initially locked onto. The head was the same size at the beginning of both of our tests so how can my subject be much larger? Did you misunderstand the term SEL-33? The only thing the camera attempted to track was the head. and its been demonstrated as equal in size relative to the frame in both of our tests.

I'll look into publishing a separate home page news story clarifying these points, and points of confusion, shortly. But if you think, after looking at all these results holistically, that there's some vast differences between any of these results that are indicative of our failures on any of these runs, I'd encourage you to think a little more carefully about this test, what it means, and what you can extrapolate from it.

-Rishi
-------------------------
Rishi Sanyal, Ph.D
Deputy Editor, Technical Editor | Digital Photography Review
dpreview.com (work) | rishi.photography (personal)

-- hide signature --

Mike from Canada
"I am not a great photographer! God is a great creator! All I do is capture His creation with the tools He has provided me."
'I like to think so far outside the box that it would require a telephoto lens just to see the box!' ~ 'My Quote :)'
http://www.michaelfastphotography.com/galleries/VP-BDI_3a.jpg
http://www.airliners.net/search/photo.search?sort_order=views%20DESC&first_this_page=0&page_limit=180&&emailsearch=mighty_mike88%40hotmail.com&thumbnails=

Petroglyph
Petroglyph Veteran Member • Posts: 6,085
Re: Focus point

Pretty smart to go through the exif.  Can you determine what was set for camera FW release, Lens FW release and focus limiter setting?

Mostly, I'd just like to know whether the lens is reporting the focus limiter setting back to the camera.  If it is, that implies the camera uses this information to fine tune its AF algorithm.  If it does, that means all of this AF/C is totally customizable by the folks writing the camera FW (firmware).

 Petroglyph's gear list:Petroglyph's gear list
Sony a7R II Zeiss Batis 85mm F1.8 Rokinon 135mm F2.0 Samyang 20mm F1.8
Petroglyph
Petroglyph Veteran Member • Posts: 6,085
Re: AFC test was redone!

Rishi Sanyal wrote:

Petroglyph wrote:

Rishi Sanyal wrote:

LightBug wrote:

Tatouzou wrote:

joergensmi wrote:

...

It can be clearly seen that the central AF-point has been selected and none of the upper ones.

For me it is clear now that this test fails due to faults by dpreview.

http://www.dpreview.com/reviews/pentax-k-1/6

There has been an update of the exiftools by Phil Harvey The version is 10.22 dated 07.07.2016 12:26 7.226.089 exiftool.exe

The result is now
Focus Mode : AF-C (Focus-priority)
AF Point Selected : Center; Single Point

I wonder why Rishi Sanyal didnt reply this post about the AF point he used: these are facts, and they show incoherences in his assessments of how he conducted the review.

First of all, I didn't write this review. Do you see my name on the author list?

There was a miscommunication about which set of images the author ended up including in the rollover we published (we had 20+ runs), which was ultimately irrelevant to the conclusions of these tests, because the center point was still over the bike, and the bike/cyclist were still out of focus in most of the shots, with focus far behind the entire bike (playing catch up).

All you've managed to do is nitpick something that has absolutely no relevance or bearing to the conclusion or review. But, out of respect for our audience, we redid it today again, to get the same exact result.

DPR obliged and re-did the straight-line AFC test. Results look much better than before, I could count 4-5 frames in focus, while previously I remember only 1 or 2 was in focus.

There is literally no difference in the results. We're publishing the rollovers side-by-side just for you. Why would the results be any different to begin with? As I'd noted before, Richard's face was within the DOF of what the AF points were focusing on (the bike in the last rollover prior to today), so the only reason we re-did the test was simply to avoid the nitpicking criticisms that we already knew had nothing to do with the end result. There is no change in the results.

Sorry to bring this up on a difficult day. How was the 3-way focus limiter set on the side of the lens again? I didn't see it in the article. I could have missed it.

Full range.

I'm frankly flabbergasted anyone things the results changed. The results haven't changed since Day 1 (when we'd initially published the review with a different AF Hold setting, since default is *not* 'Off'). 3 iterations later, the results are exactly the same, but we were definitely wrong and incompetent the previous times, right?

Desire to believe is a powerful thing I suppose...

I can't wait to hear the next excuse for why our latest run also has a dismal ~30% hit-rate...

Funny, if a lens has a focus limiter on it, it's the first thing I set before doing an AF/C session.  Often, it can make the difference between so-so & golden on a Canon lens.  I admit I know nothing about the particular lens, Pentax DFA* 70-200 f/2.8 except I do know (or think I know) it has a 3-range limiter on the side.  If you bike rider was, say, totally in range-2 for the test, I bet I could change your results with that lens/camera combo. 

 Petroglyph's gear list:Petroglyph's gear list
Sony a7R II Zeiss Batis 85mm F1.8 Rokinon 135mm F2.0 Samyang 20mm F1.8
Petroglyph
Petroglyph Veteran Member • Posts: 6,085
Re: AFC test was redone!

MightyMike wrote:

Timbukto wrote:

Rishi Sanyal wrote:

Full range.

I'm frankly flabbergasted anyone things the results changed. The results haven't changed since Day 1 (when we'd initially published the review with a different AF Hold setting, since default is *not* 'Off'). 3 iterations later, the results are exactly the same, but we were definitely wrong and incompetent the previous times, right?

Desire to believe is a powerful thing I suppose...

I can't wait to hear the next excuse for why our latest run also has a dismal ~30% hit-rate...

Seattle weather!

The phase of the moon?

Need an excuse to get that lens.

-- hide signature --

Mike from Canada
"I am not a great photographer! God is a great creator! All I do is capture His creation with the tools He has provided me."
'I like to think so far outside the box that it would require a telephoto lens just to see the box!' ~ 'My Quote :)'
http://www.michaelfastphotography.com/galleries/VP-BDI_3a.jpg
http://www.airliners.net/search/photo.search?sort_order=views%20DESC&first_this_page=0&page_limit=180&&emailsearch=mighty_mike88%40hotmail.com&thumbnails=

 Petroglyph's gear list:Petroglyph's gear list
Sony a7R II Zeiss Batis 85mm F1.8 Rokinon 135mm F2.0 Samyang 20mm F1.8
DuncanM1 Contributing Member • Posts: 817
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
6

I'm reporting this here:

Here's another carefully done lab test that shows the K-1 to be competitive with comparable FF DSLRs:

http://www.chassimages.com/minimags.php?fic=Pentax-K1&dos=OT

(pages 8-9 and 18-19)

The review is in French and it uses Flash so I can't quote it directly except to copy it by hand.

Here's a quote from pages 8-9 of the above report:

"la cadence de declenchement n'est pas des plus elevees (4.5 i/s) mais l'autofocus est reactif. Le K-1 parvient a suivre le sujet jusqu'a 2m de distance. C'est remarquable car beaucoup d'appareils ne descendent pas aussi bas. La rafale est correcte et, surtout, les sujets mobiles seront restitues avec precision...a condition d'avoir une optique soit a la hateur. "

which google translates, roughly, as:

"the triggering rate is not higher (4.5 i / s) but the autofocus is reactive. K-1 manages to follow the subject until 2m away. This is notable because many devices do not descend so low. The burst is correct and especially moving subjects accurately be returned ... provided you have an optical or has hightwise." (provided you have a high performance lens - my translation of the last sentence.)

The data on pages 18-19 shows the K-1 to be a middling (IE: better than some and worse than others) AF performer compared to other price comparable FF DSLRs.

More from page 8:

Another quote and translation from page 8 of the above review:

"La difference tarifaire est telle que l'achat de deux objectifs comble largment l'ecart du prix qui separe le K-1 du Nikon D810.
La rafale du K-l plafonne a 4,5 i/s, ce qui le situe au niveau du Canon 6D. Le Nikon D810, qui a le meme capteur, grimpe a 6 i/s. Cette performance n'est pas honteuse, mais elle montrera parfois ses limites face a des animaux ou des sportifs.
Une rafale etriquee est souvent le signe d'un autofocus poussif. Ce n'est pas le cas ici. L'AF est performant, il suit la rafale comme une horloge, meme jusqu'aux distances proches: un remarquable resultat.

The fare difference is such that the purchase of two goals largment fills the gap that separates the price of the K-1 Nikon D810.
The burst of the K-1 caps was 4.5 i / s, which places it at the (level of the) Canon 6D. The Nikon D810, which has the same sensor, climbing (to) 6 i / s. This (K-1) performance is not shameful, but sometimes show its limitations facing animals or athletes.
A narrow burst is often the sign of a winded (slow) autofocus. That's not the case here. AF highly efficient, it follows the burst like clockwork, even up close distances: an outstanding result."

  • The Chasseur D'Images AF testing was carefully done under laboratory conditions.
DuncanM1 Contributing Member • Posts: 817
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
3

Image Hunter (Chasseur D'images in French) summarizes their AF test methodology:

"Autofocus was quick and precise ... (summary of K-1 AF tests).

One often reads about AF testing in different sites, blogs or magazines. And most often we trusted him.

But what criteria are these tests based upon? On some photos taken at the edge of a road or a railway, on the impressions of a return weekend? Or reliable measurements, accurate, controlled and reproducible

At Image Hunter, technical assessments are always based on a rigorous and identical test protocol from one device to another.

To test autofocus, we use a single test bench in Europe, especially developed for The lab, which simulates the movement of a subject at different speeds. It allows us to know everything about the autofocus of your device: actual rate in AF, number of images made many sharp images, number of blurred images, etc. _

Where others consider an AF "fast" or too slow, the chart shows the real "Image Hunter" results of camera AF performance and facilitates the comparison between different models. And it's all the difference between a real test and simplistic testing."

I wish I could get a copy of the Chasseur D'images report in english, since translating it manually is rather tedious.

SBS
SBS Senior Member • Posts: 2,537
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
1

What about the graph on page 19? The K-1 seems to hold its own very well according to it?

wizofoz
wizofoz Regular Member • Posts: 143
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
1

Wow, that French magazine review looks classy and comprehensive. I only wish I could read what it has to say.

 wizofoz's gear list:wizofoz's gear list
Pentax K-7 Pentax K-5 Pentax Q Pentax K-3 Pentax K-1 +32 more
howieb101 Contributing Member • Posts: 633
Re: AFC test was redone!
3

Thanks for replying Rishi.

I am not really surprised by the results.  I was mainly concerned with the text which explicitly said focused on the face when it was on the jersey.

Given the test results it is kind of to be expected that people would check your findings.  I wouldn't call it nitpicking it's called verification. By your own accounts the focus was on the jersey after all.

Anyway, again thanks for your reply and hard work.

Petroglyph
Petroglyph Veteran Member • Posts: 6,085
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
10

DuncanM1 wrote:

I'm reporting this here:

"La difference tarifaire est telle que l'achat de deux objectifs comble largment l'ecart du prix qui separe le K-1 du Nikon D810.

Above they said one can purchase two large high quality lenses for the price difference between the D810 and K1

La rafale du K-l plafonne a 4,5 i/s, ce qui le situe au niveau du Canon 6D

The burst rate is the same as Canon 6D

. Le Nikon D810, qui a le meme capteur, grimpe a 6 i/s.

The D810 has the same sensor but climbs to 6 fps. (So does K1 in crop mode I guess they missed that.)

Cette performance n'est pas honteuse, mais elle montrera parfois ses limites face a des animaux ou des sportifs.

The 4,5 fps will run into trouble with sports and wildlife.

Une rafale etriquee est souvent le signe d'un autofocus poussif. Ce n'est pas le cas ici. L'AF est performant, il suit la rafale comme une horloge, meme jusqu'aux distances proches: un remarquable resultat.

Slow fps is often the sign of slow AF. That is not the case here. [K1] AF is highly efficient. it follows the frames of the burst like clockwork, even to distances up close: An outstanding result.

  • The Chasseur D'Images AF testing was carefully done under laboratory conditions.

A little different than not good enough to take pictures of your kids running about in the garden I think.

Cheers.

 Petroglyph's gear list:Petroglyph's gear list
Sony a7R II Zeiss Batis 85mm F1.8 Rokinon 135mm F2.0 Samyang 20mm F1.8
CarlosRG Regular Member • Posts: 340
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
1

wizofoz wrote:

Wow, that French magazine review looks classy and comprehensive. I only wish I could read what it has to say.

Chasseur D'images is one of the best technical photo magazines in Europe, if not the best. They have been doing camera and len reviews for decades and they are very competent at it. I have been reading them since mid-80's and I have always found them quite objective and unbiased.

Carlos

Tan68
Tan68 Veteran Member • Posts: 5,778
Re: AFC test was redone!

Rishi Sanyal wrote:

LightBug wrote:

Tatouzou wrote:

joergensmi wrote:

...

...

... All you've managed to do is nitpick something that has absolutely no relevance or bearing to the conclusion or review. ...

...

... the only reason we re-did the test was simply to avoid the nitpicking criticisms that we already knew had nothing to do with the end result. There is no change in the results.

-Rishi
-------------------------
Rishi Sanyal, Ph.D
...

Amid all this talk of properly designed tests and comments about scientifically repeatable tests (or words to this effect), I wonder about work experience.

Without disclosing details, it has been my work experience that saying 'close enough' (or words to that effect) is not always good enough.

Regardless of who is in the right, there will be times when not having facts and processes in order (ducks as well) opens the door to argument [read: distractions from the point]. Even though the point is valid, the argument can be lost because uncertainty is introduced by wandering ducks...

Anyway, either 'you' or 'they' are correct. The importance of that old saying about ducks becomes more clear. This whole thing offers many lessons for both sides.

 Tan68's gear list:Tan68's gear list
Phase One Capture One Pro +1 more
ser dpreview Regular Member • Posts: 108
Re: AFC test was redone!
1

@Petroglyph, You have all my admiration

Alex Sarbu Veteran Member • Posts: 9,872
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
1

CarlosRG wrote:

wizofoz wrote:

Wow, that French magazine review looks classy and comprehensive. I only wish I could read what it has to say.

Chasseur D'images is one of the best technical photo magazines in Europe, if not the best. They have been doing camera and len reviews for decades and they are very competent at it. I have been reading them since mid-80's and I have always found them quite objective and unbiased.

Carlos

And they're not afraid to "upset" the camera makers - few years ago, they entered into a conflict with Canon, because of an article regarding the 1D MkIII's AF issues.

Alex

 Alex Sarbu's gear list:Alex Sarbu's gear list
Ricoh GR III Pentax K-5 IIs Pentax K-1 II Pentax smc DA 21mm F3.2 AL Limited Pentax smc DA 70mm F2.4 AL Limited +8 more
solarider Veteran Member • Posts: 4,736
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test

Hi, If they (Chasseur D'Images) test for Z axis say on a bike rider how do they comment about the K-1 performance in this?

Thank you.

Nic

-- hide signature --

The intuitive mind is a sacred gift, the rational mind is a faithful servant.
We have created a society that honors the servant and has forgotten the gift. - Albert Einstein
________________
Science... became Scientism, which means it didn't just pursue its own truths, it aggressively denied that there were any other truths at all... Ken Wilber
__________
Intuition will tell the thinking mind where to look next. - Jonas Salk
__________
Coincidence is God's way of remaining anonymous - Albert Einstein

Petroglyph
Petroglyph Veteran Member • Posts: 6,085
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
1

solarider wrote:

Hi, If they (Chasseur D'Images) test for Z axis say on a bike rider how do they comment about the K-1 performance in this?

Thank you.

Nic

They said the fps was slow but the AF capabilities were excellent.  The slow fps could be the entire problem with the bike test.  Better might be to switch the camera to 6.5 fps crop mode and blast away at the bike jpeg only.  See what happens then.  But Chasseur D'Images wasn't fooled by a little confirmation bias into thinking that the AF itself was slow.  This bodes well for future models and firmware updates.

 Petroglyph's gear list:Petroglyph's gear list
Sony a7R II Zeiss Batis 85mm F1.8 Rokinon 135mm F2.0 Samyang 20mm F1.8
ET2 Veteran Member • Posts: 4,110
Re: Analysis of Rishi Sanyals claims on competence and scientific approach for the K-1 AF test
2

Petroglyph wrote:

solarider wrote:

Hi, If they (Chasseur D'Images) test for Z axis say on a bike rider how do they comment about the K-1 performance in this?

They might not have performed the test where the camera has to continuously refocus in z axis. DPR did the test, several times, with various Pentax cameras over last two years and got consistent results.

They said the fps was slow but the AF capabilities were excellent. The slow fps could be the entire problem with the bike test.

As said previously, that makes no sense. It's much harder for the camera to maintain focus at higher fps than it is at lower fps. That's why D500 performance is even more impressive as it does it at 10 fps. Another hole in your argument is K3 that has higher fps but yet performed subpar compared to the competition. So there.

Keyboard shortcuts:
FForum MMy threads