tesilab

tesilab

Lives in United States United States
Works as a Software Engineer
Joined on Feb 4, 2011

Comments

Total: 123, showing: 1 – 20
« First‹ Previous12345Next ›Last »
On article Panasonic Lumix DC-G9 Review (901 comments in total)
In reply to:

Stejo: These are good results. Panasonic's DFD technology is terribly underappreciated. What they have achieved with CDAF is nothing sort of miraculous. It will inevitably surpass PDAF given enough time and RnD. I only hope they keep investing in it.

@Stejo Off-sensor traditional PDAF requires elaborate calibration--if it is available--and cannot approach CDAF accuracy. On-sensor PDAF has been limited to whatever could be squeezed onto the sensor without compromising image quality. (Dual-pixel system I guess helps with that). But I don't think DFD as a technology doesn't really have very far it can realistically go from here, with the exception of faster processing cycles. And all technologies are benefitting from those.

Link | Posted on Jan 5, 2018 at 16:12 UTC
On article Panasonic Lumix DC-G9 Review (901 comments in total)
In reply to:

Stejo: These are good results. Panasonic's DFD technology is terribly underappreciated. What they have achieved with CDAF is nothing sort of miraculous. It will inevitably surpass PDAF given enough time and RnD. I only hope they keep investing in it.

However impressive it is (kudos Panasonic) I don't think it can ever surpass PDAF. The very nature of DFD seems like heuristic reverse-engineering of depth from OOF pixels, vs PDAF which is very straight-forward out-of-phase measurement. I think Canon' s dual-pixel approach is the one with the brightest overall future.

Link | Posted on Jan 4, 2018 at 23:42 UTC
In reply to:

JerryWilliam: Another "Fuller Frame" camera. The original Pentax 645 took a 56×41.5mm image on 120 roll film. When will a REAL 645 sensor be produced?

Real MF in this case would be 6x6 (56x56 mm), not some puny 645 medium format wannabe.

Link | Posted on Nov 15, 2017 at 01:16 UTC
On article Demo: Sony a7R III's impressive Eye AF (134 comments in total)
In reply to:

sharkcookie: Can someone test what happens if you do a close up portrait with a fast lens wide open. In those cases you have to get the focus on the eye ball rather than the eye brow. That has been an issue with almost any AF even when the AF point selected is right on the eye. Almost all AF systems are going for the closest object which is usually the eye brow. Would be cool if Sony had that down.

I guess you haven't been paying much attention the last couple of years. It nails the eyeball wide open.

Link | Posted on Oct 27, 2017 at 00:24 UTC
In reply to:

Hinder: I'm confused, a5000, a6000 a 63000 a6500, a7 a7s a7r a7l a7ll a7lll a7rll a7sll a7r lll a9rll...............Your names are confusing. Why do they come out with a new one every week and minor improvements that seem to be able to be fixed with a firmware update but keep jacking up the price? The reason I have not bought a Sony lately is I'm afraid it's going to be replaced or outdated every three months. I don't care about video, I just want to take a picture. Sony, can you please just make a nice camera that won't be outdated, overheat or doesn't focus properly and stick with it for a bit?

@Hinder: How exactly does it hurt you that Sony keeps making new cameras? Your issue is that you can't decide when a camera is what you want or whether you should wait for something better. There will always be something better, so deal with it.

Link | Posted on Oct 26, 2017 at 01:13 UTC
In reply to:

RPJG: Asking for a friend *cough*:

What are the downsides of using electronic shutter? When would you *not* want to use it?

Electronic shutter can cause banding artifacts, especially in fluorescent lighting.
Also motion captured at a "fast" shutter speed will be distorted, since the exposure time may be very brief, but it is not exposed at the same time all across the sensor.

Link | Posted on Oct 26, 2017 at 01:10 UTC
In reply to:

ProfHankD: Interesting approach. Cameras often have custom hardware acceleration for processing images, so why not use it tethered as a coprocessor? The catch is this shouldn't be the ONLY raw conversion available, especially since they suggest each camera can only process raws from the same model.

I'm not sure it is using the camera as a coprocessor. I think it more likely that it does *all* the processing in the camera, and uses the computer/screen for a reasonably accessible user interface.

Which by the way would be a very good approach to configuring cameras with bewildering number of options, which could be much better presented/navigated explained as a web application served by the camera.

Link | Posted on Sep 8, 2017 at 12:29 UTC

I tabulated 60 responses from users on this thread, and Sony A9 took first overall place 42 out of 60 times. That's pretty impressive.

Link | Posted on Aug 24, 2017 at 16:29 UTC as 27th comment
In reply to:

sean lancaster: A9 = 28 points
1Dxii = 25 points
D5 = 23 points

Another score that doesn't add up Sean. There should be a total of 72 points, not 76.

Link | Posted on Aug 24, 2017 at 15:50 UTC
In reply to:

wiegerket: Nikon won (YES MY..MY TASTE) 8 out of 10 pictures. One 1 and One 2. I must admit, I own a D5 and I recognize it's output out of anything. I am surprised that the Nikon pictures had the least blow-out. I thought overall, the other two might have an edge on DR.

But there are 12 pictures, so how did you get 8/10?

Link | Posted on Aug 24, 2017 at 15:45 UTC

I really expected Sony to do worse here, though I like the colors I've been getting from Sony just fine. My scores were pretty evenly divided, Nikon came out worst, but I think its all inside the margin of error. This says to me that colors just aren't the factor. Pick ergonomics, or focus accuracy/speed, or reliability/service or lenses or something else.

S=25
C=25
N=22

Link | Posted on Aug 23, 2017 at 23:17 UTC as 59th comment
In reply to:

PhotoDiod: It looks like Sony has figured out the skin colors. I liked them so much, I am ready to switch to Sony from Canon! Now, Sony, if only you can fix the sky color science - it still looks familiarly fluorescent-y:

26:52
28:07
28:47

The sunflower image seems to suffer from similar cartoonish rendition as the sky but I suppose that would be an easier fix in post-processing.

The more accurate skies will show up as cyan. The more pleasing skies have been tweaked, since no one likes their skies to look cyan.

Link | Posted on Aug 23, 2017 at 23:14 UTC
On article Don't buy another lens, buy a flash instead (336 comments in total)
In reply to:

Kaso: "Forget Lenses, Buy a Flash"
"Don't buy another lens, buy a flash instead"

Cheap tabloid title that pitches A against B.

To everything -- turn, turn, turn
There is a season -- turn, turn, turn
And a time to every purpose under heaven

A time to use fast lens, a time to use flash
A time to stay wide, a time to come close
A time for motion blur, a time for action freeze
A time for soft background, a time for everything sharp
A time for crop, a time for larger format
. . .

"There are times when a flash is appropriate"
There, isn't that a much better title?

@Kaso apparently you did not grow up reading the bible. Remove just the "turn turn turn" phrase and it is straight out of Ecclesiastes. https://www.biblegateway.com/passage/?search=Ecclesiastes+3%3A1-8&version=KJV

Link | Posted on Aug 21, 2017 at 20:56 UTC

I think his lawyers must have missed the point. He had a better claim to copyright that had nothing to do with his ownership of the camera. Those pictures were planned and taken with the photographer's intent, not the monkey's. He engineered that whole situation so that a specific type of photograph would result. The physical act of pressing a button is a fraction of what makes a photograph. You could replace that macaque with another one and get a similar result. But the same is not true of the photographer.

Link | Posted on Jul 16, 2017 at 02:46 UTC as 80th comment | 1 reply
On article Fujifilm X-A3 added to studio scene comparison (105 comments in total)

Please Fuji make higher end (APSC) cameras with bayer!

Link | Posted on Jun 9, 2017 at 13:04 UTC as 19th comment | 5 replies
On article Sony a9: more speed, less dynamic range (669 comments in total)
In reply to:

bilcobarnes21: Am I the only one getting a little tired of DP Reviews obsession with the A9 and Sony in general?
It makes me wonder what involvemt Sony have with DPR 🤔

There was a much busier time in the market when more products and innovations from more companies were being cranked out. Things have slowed down quite a bit relatively speaking. DPR even started featuring throwback reviews!

So can they help it if most of the interesting developments are coming from Sony?

Also they've taken a much needed incremental approach to publishing their findings on cameras. Remember when we had to wait months and months for some cameras to be reviewed?

So stop bellyaching about this supposed Sony fanboyism from DPR. We know they'd rather be shooting Fuji. :)

Link | Posted on May 18, 2017 at 13:54 UTC
In reply to:

Kisaha: Being there, done that,

NX500 and NX3000 with 10mm 3.5 fish eye/16mm 2.4f/30mm 2f/45mm 1.8f pancakes.

My friends LX100 is collecting dust inside its lens, and after 3 lens replace in warranty he just gives up. He just bought a NX300, but he already has most NX lenses (he owns a NX1 as well).

Seriously, I do not get why to spend so much money for a compact.

Isn't the LX100 a zoom lens? That would be more likely to suck in dust, I would think.

Link | Posted on May 15, 2017 at 17:35 UTC
On article Alpha-better: Sony a9 versus a7R II (506 comments in total)
In reply to:

Vik2012: Tune in tomorrow, when DPReview compares the new Sony A9 to the Samsung Galaxy S8 because... well... because there's nothing much to write about the A9 between launch announcement and testing.

If comparing the A9 to the a7R II was a parody of lesser sites that use endless filler articles to keep a new product name in "article" titles, then nice one DPReview.

@pazinboise I think you are mistaken about one thing. The group putting out the NX1 in fact did most everything right (amazing tech, super responsive to feedback with updates, etc.) maybe more so than Sony (with respect to updates). But it didn't fail on its merits. It got yanked independently by higher ups/bean counters at Samsung before they could reap the rewards due them.

Link | Posted on Apr 26, 2017 at 16:11 UTC
On article Alpha-better: Sony a9 versus a7R II (506 comments in total)
In reply to:

patrocal: First reviews are coming in and Tony Northrup says the a9 is worst than the A7RII in noise at high ISO and less DR which is surprising cause the A9 has less mpx so you would expect better high iso and the A7rii is 2 years old! So unless you need speed and silent shutter and are ready to compromise quality don't see why you would want A9 at almost $2k more.

@BarnET you misunderstand. I already said any comparisons between different resolution sensors must be done at fixed output size rather than 1:1 (100% crop) for obvious reasons. But the error is the error, and it has nothing to do with 100% viewing. This error can be measured. Really the question in any given case is how noticeable to the eyeball. I didn't claim it was very significant.

Link | Posted on Apr 26, 2017 at 15:44 UTC
On article Alpha-better: Sony a9 versus a7R II (506 comments in total)
In reply to:

patrocal: First reviews are coming in and Tony Northrup says the a9 is worst than the A7RII in noise at high ISO and less DR which is surprising cause the A9 has less mpx so you would expect better high iso and the A7rii is 2 years old! So unless you need speed and silent shutter and are ready to compromise quality don't see why you would want A9 at almost $2k more.

@BarnEt There is a basic phenomenon known as quantization error. (See https://en.wikipedia.org/wiki/Quantization_(signal_processing)#The_additive_noise_model_for_quantization_error )Any time you subdivide a space into smaller wells, you are increasing the cumulative errors, since you must create a discrete value for each sensel. So while yes, it is true that the overall size of the sensor is a major factor in determining the noise, (which is why you should only compare two resolutions at fixed output size rather than 1:1). The cost of smaller wells in terms of additional noise is still noticeable.

Link | Posted on Apr 26, 2017 at 14:45 UTC
Total: 123, showing: 1 – 20
« First‹ Previous12345Next ›Last »