Kiril Karaatanasov

Lives in Bulgaria Bulgaria
Has a website at http://www.karaatanasov.info
Joined on Mar 31, 2006

Comments

Total: 353, showing: 1 – 20
« First‹ Previous12345Next ›Last »

Pointless ......

Link | Posted on Mar 26, 2017 at 07:41 UTC as 10th comment
In reply to:

ttran88: Zeiss needs to stop monkeying around and make some EOS M lenses 😉

May be Canon need to make a camera first and also license the protocol to Zeiss? I mean that the Batis line of lenses is full frame and M5 is crop also M5 is bit of yesterday's tech like 4 years behind Sony and Fuji.

Link | Posted on Mar 23, 2017 at 14:27 UTC
In reply to:

princecody: Hope it's autofocus 🤓👍🏻

Batis line has AF with Sony E mount cameras.

Link | Posted on Mar 23, 2017 at 14:24 UTC
In reply to:

chbde: It's an interesting approach google choose - in principle they omit some of the image data (mostly by zeroing DCT coefficients) in some educated-guess-way, evaluate the image with their Butteraugli perceptual metric to simulate a human viewer, and do this iteratively until they are close to an defined quality goal. I guess it's not likely that their method will find widespread usage, as it's probably inherently much too slow and memory hungry, albeit for high-volume web sites it may be very interesting. If your image is requested a million times a week, it's of lesser importance if the encoder needs an hour to make it 30% smaller.

BTW, apart from spying on people (what they certainly do), Google actually finances a lot of valuable research, and I can see in the case of Guetzli the benefit for them to save money on bandwidth, but no additional user tracking opportunity.

One risk coming out of the breach of privacy through extensive data analysis is erroneously associating individuals as part of some group. It may be fine if the difference is seeing ads for a product you are not keen to buy. It get pretty nefarious if Goolge are legally pressed to share this group with a national security force as part of counter terrorism investigation. All of of a sudden one may not be allowed on to planes or denied access to services or privileges, be subjected to extensive searches and monitoring etc. Further things can get uglier if all of a sudden potential employers from Google(?) can get their hands on some of that data.....

Link | Posted on Mar 22, 2017 at 18:32 UTC

Color seems quite badly distorted in the example

Link | Posted on Mar 17, 2017 at 22:46 UTC as 42nd comment | 2 replies
In reply to:

chbde: It's an interesting approach google choose - in principle they omit some of the image data (mostly by zeroing DCT coefficients) in some educated-guess-way, evaluate the image with their Butteraugli perceptual metric to simulate a human viewer, and do this iteratively until they are close to an defined quality goal. I guess it's not likely that their method will find widespread usage, as it's probably inherently much too slow and memory hungry, albeit for high-volume web sites it may be very interesting. If your image is requested a million times a week, it's of lesser importance if the encoder needs an hour to make it 30% smaller.

BTW, apart from spying on people (what they certainly do), Google actually finances a lot of valuable research, and I can see in the case of Guetzli the benefit for them to save money on bandwidth, but no additional user tracking opportunity.

Google are tracking people on and off line all the time...if you use android Google now where you are, where you have been, who you have met with, have been speeding.

When you are on the internet on pretty much any device and browser Google can track pretty much everything you see. So they know your political bias, your fears, your speed of reading, your interests .....they pretty much know you better than you know yourself...

facebook are close but not as all encompassing....

Link | Posted on Mar 17, 2017 at 22:45 UTC
On article CP+ 2017: Hands-on with new Voigtländer E-mount primes (116 comments in total)

There was no chance to pop these lenses on a camera?

Link | Posted on Feb 24, 2017 at 14:50 UTC as 15th comment | 1 reply
On article CP+ 2017: Hands-on with Sigma's newest lenses (198 comments in total)
In reply to:

Flashback: If you have a minute Sigma, could you knock up some E-mount versions?
Don't bother with the stabilisation, if that helps!

I would believe they ought to focus on lenses where E mount brings inherent advantages e.g. size. I would hope some wide angle lenses like 35/2, 16/2.8, 12/2.8, 12-24/4 could be done smaller when designed for E mount. I also believe normal zooms like 18-55/2.8-4 for APS-C and 24-105/4 full frame could benefit from native mirrorless design. Sigma did very well with the 30/1.4

135/1.8, 100-400 and 14/1.8 would be similar size regardless of the mount. As long as the MC-11 & Metabones offer first class support for focusing I am happy. 24-70/2.8 seems to also be one of those lenses that will remain huge regardless of the mount and I am happy with an adapter.

Link | Posted on Feb 23, 2017 at 20:41 UTC
On article CP+ 2017: Hands-on with Sigma's newest lenses (198 comments in total)

What is the MO mode of focusing?

AF/MF is pretty usual. I see MO for first time on these new Sigma zooms - 24-70 and 100-400

Link | Posted on Feb 23, 2017 at 20:34 UTC as 32nd comment | 1 reply
On article Fujifilm XF 23mm F2 R WR sample gallery (284 comments in total)
In reply to:

Kiril Karaatanasov: Why is rendition so poor? Is this typical Fuji? This is terrible. Shadows are completely killed. It looks as if the camera is defective. Colors are ...bizarre to avoid saying awful?

Rishi I did not mean to troll.

From photographic perspective the samples are bad. Shadows are crushed and lack detail. This I believe to not be defect of the sensor but the JPG engine.

As you can see people that are NOT fanboys of Fuji JPEG are iterating the same concern. Weather you want to hear or not is your choice.

To me the test is flawed whether by Fuji's poor default JPEG implementation or your choice of JPEG processing is irrelevant. The outcome is poor quality JPEGs that fail to demonstrate the lens capabilities. The samples do have retro appeal as photographic product and I do not object to your photography skill or Fuji decision to implement instagram filter in the camera. Only point I am making is that the over-processed look of the samples is obstructing the review.

Link | Posted on Feb 20, 2017 at 09:19 UTC
On article Fujifilm XF 23mm F2 R WR sample gallery (284 comments in total)
In reply to:

gunlbum: Anyone else notice a pattern in the noise? Specifically look at the mother's face in the 5th photo. Is this typical with Fuji? I'm sure it could be taken care of but interesting for me.

Lettermanian it seems A6300/A6500 use very similar sensor to Fuji X-T2/X-T20 subtract the Bayer vs X-Trans and both offer ample photographic capability

Link | Posted on Feb 17, 2017 at 15:47 UTC
On article Fujifilm XF 23mm F2 R WR sample gallery (284 comments in total)
In reply to:

Aroart: Fuji has some sort of magic going on.. Even the images that people see as flawed has a unique characteristic that is beautiful...

Too bad this is lens review and the samples do not show any of the lens qualities. Instead we can enjoy the built-in camera filters on extremely soft, poorly exposed and over-processed images....

Link | Posted on Feb 17, 2017 at 15:37 UTC
On article Fujifilm XF 23mm F2 R WR sample gallery (284 comments in total)
In reply to:

Jason Haven: The negativity around here is insane.

A couple notes, since I have this lens:

- It's sharp enough wide open, unless your subject is pretty close.
- By f/4 it's sharp (pixel peeping sharp) across the frame
- Lens is absolutely silent, and incredibly fast.
- Build quality is top notch.

For my uses, it's a fantastic lens, since I don't intend to attempt much in the way of 'up close at f/2' photography with it.

I also own the 23mm f/1.4. The f/2 is a better all around lens. The f/1.4 is better for astro work, and portraits (if your goal is to blur the background). Both tools have their uses. :)

@Jason Haven you would do everyone a favor if you can share on Flickr or other site some images with this lens with all art filters turned off and as much as possible accurate exposure and white balance. A lens review should not be obfuscated with camera effects.

DPR messed up big time these shots.

Link | Posted on Feb 17, 2017 at 15:35 UTC
On article Fujifilm XF 23mm F2 R WR sample gallery (284 comments in total)
In reply to:

Kiril Karaatanasov: Why is rendition so poor? Is this typical Fuji? This is terrible. Shadows are completely killed. It looks as if the camera is defective. Colors are ...bizarre to avoid saying awful?

@The Name is Bond lenses at least from reputable manufacturers have certain characteristics. If you go about showing the qualities of a lens it is best to switch off all the "art filters" a camera can do.

I see in these samples terrible colors that may well be caused by extremely poor optics. So DPR are not doing favor to Fuji by applying artsy filters in lens review. The presented shots if ruined by art filters as some people imply do not demonstrate well lens contrast, lens colors, lens resolution and lack of optical deficiencies. I really see crappy photos that hardly manage 5 stops of dynamic range possibly due to lens that has extremely high amount of internal reflections and extremely low quality optics.

Link | Posted on Feb 17, 2017 at 15:32 UTC
On article Fujifilm XF 23mm F2 R WR sample gallery (284 comments in total)
In reply to:

misspiggy01: this looks like a vintage lens from the 60ies with a native mount and AF. what´s not to like?

This is terrible both the colors and the lens. It really looks like made in the 60s....like so bad,

Sony 24/1.8 can run circles around this Fuji rendition.......

This Fuji look is nothing short of terrible

Link | Posted on Feb 16, 2017 at 20:56 UTC
On article Fujifilm XF 23mm F2 R WR sample gallery (284 comments in total)

Why is rendition so poor? Is this typical Fuji? This is terrible. Shadows are completely killed. It looks as if the camera is defective. Colors are ...bizarre to avoid saying awful?

Link | Posted on Feb 16, 2017 at 20:53 UTC as 32nd comment | 7 replies
In reply to:

Kiril Karaatanasov: In the bojeh examples the GM lens performs vastly better. Sigma is clearly squashing the bojeh towards the corners in cat's eye shape. The Sony retains circular bojeh all over the field. This allows for stitching bokeh panoramas

So really if you shoot Sony camera this sigma is not usable at all

It is not nice if you do bokeh panorama/brenizer.

https://learningdslr.quora.com/The-Brenizer-Method-Bokeh-Panorama

I use the older non Art Sigma 85/1.4 and panoramas stich ok. Also the Canon 135/2 is doing well.

As you can see in my test the Zeiss Batis and the other lens I tested produce similar cat's eye effect on the bokeh making stitching harder if not impossible.

So "not usable" for bokeh panoramas. I believe the Sony from these examples fares batter but am not sure if it is usable in real life too as I wrote in my second post. The test is really not showing bokeh in extreme corners where stitching will be hardest...

For other things the Sigma is probably well worth it.

Link | Posted on Feb 10, 2017 at 20:36 UTC
In reply to:

Kiril Karaatanasov: In the bojeh examples the GM lens performs vastly better. Sigma is clearly squashing the bojeh towards the corners in cat's eye shape. The Sony retains circular bojeh all over the field. This allows for stitching bokeh panoramas

So really if you shoot Sony camera this sigma is not usable at all

I did some home tests for this characteristic on my lenses and the old Sigma 85/1.4 is may be similar in this regard or a bit better. Canon 135/2 is magnificent in this regard.

You can see my test here https://photos.google.com/share/AF1QipNfLXlzuPkTtsf50ckxhExOzuFZYwt5WZlnx6nvJCzETkWvp4rI9pTM4QnAZaqdLw?key=MGhyc0REZlBCOFh5d2ZLSVVDbWc4dGkwbmg4LVpB

Would have been nice if DPR had the lights a bit further in the background and lined up such that some of them fall closer to the corner. From the provided samples this Sigma looks pretty dismal

PS In the real world sample the Sony does not look promising either.....Just check how the old Canon 135/2 works and....

Link | Posted on Feb 9, 2017 at 11:06 UTC

In the bojeh examples the GM lens performs vastly better. Sigma is clearly squashing the bojeh towards the corners in cat's eye shape. The Sony retains circular bojeh all over the field. This allows for stitching bokeh panoramas

So really if you shoot Sony camera this sigma is not usable at all

Link | Posted on Feb 9, 2017 at 10:49 UTC as 67th comment | 4 replies
In reply to:

villagranvicent: These G Masters would be a perfect match for my Casio Master of G.

....And you need to go to the photo shoot in a G class Mercedes

Link | Posted on Feb 8, 2017 at 08:42 UTC
Total: 353, showing: 1 – 20
« First‹ Previous12345Next ›Last »