Rambalac

Lives in Japan Japan
Works as a Software developer
Joined on Mar 18, 2009

Comments

Total: 49, showing: 1 – 20
« First‹ Previous123Next ›Last »
In reply to:

Alexander Raemer: If google and alike made it imposible to right click and copy, download and save as etc, it would go a long way to protect everyone's copyright...

Do you mean stop displaying pictures? To show a picture you have to download it first.

Link | Posted on Aug 19, 2017 at 08:38 UTC
On article Advertising vs reality: microSD memory card speed test (73 comments in total)

Test is extremely stupid and absolutely useless. Most important minimum writing speed with mostly full card is not tested. That guys are just incompetent.

Link | Posted on Aug 8, 2017 at 10:37 UTC as 10th comment | 2 replies
In reply to:

Cary Knoop: From the article: "4:2:2 10-bit codec is 64x as much data as before"

No it isn't!

10 bit is 4 times as much data as 8 bit and 4:2:2 is 1 1/3 times as much data as 4:2:0, so it is 5 1/3 times as much not 64 times as much.

There is only one explanation - incompetence. Which itself is ok if not ignorance. 4 bit is not 16 numbers, only possiblity of 16 different numbers. Amount of data is irrelevant to processing or storing. Data can be reduced during processing but amount of processing output is the same amount if you store it. 1 bit is minimum possible amount of information (presume maximum entropy), so 4 bits is twice bigger than 2 bits by amount of data. If you cut two cakes 1kg each into 4 and 16 pieces it does not mean the second cake become 4 times bigger.

Link | Posted on May 9, 2017 at 08:19 UTC
In reply to:

Cary Knoop: From the article: "4:2:2 10-bit codec is 64x as much data as before"

No it isn't!

10 bit is 4 times as much data as 8 bit and 4:2:2 is 1 1/3 times as much data as 4:2:0, so it is 5 1/3 times as much not 64 times as much.

Wait, subtract? Indeed...
How many disks do you need to store 10Mbit of data if one disk is only 1Mbit? 9 million disks.
What is the name of uni for such idiots?

Link | Posted on May 9, 2017 at 05:34 UTC
In reply to:

Cary Knoop: From the article: "4:2:2 10-bit codec is 64x as much data as before"

No it isn't!

10 bit is 4 times as much data as 8 bit and 4:2:2 is 1 1/3 times as much data as 4:2:0, so it is 5 1/3 times as much not 64 times as much.

How much did you pay for your PhD? Or are you female?
If you even graduated high school you would know that amount of data/information is measured by number of bits and related to maximum possible number logarithmically.

Link | Posted on May 9, 2017 at 05:18 UTC
In reply to:

Cary Knoop: From the article: "4:2:2 10-bit codec is 64x as much data as before"

No it isn't!

10 bit is 4 times as much data as 8 bit and 4:2:2 is 1 1/3 times as much data as 4:2:0, so it is 5 1/3 times as much not 64 times as much.

It does not. You just have no idea what is data and how to measure it.

Link | Posted on May 9, 2017 at 04:59 UTC
In reply to:

Cary Knoop: From the article: "4:2:2 10-bit codec is 64x as much data as before"

No it isn't!

10 bit is 4 times as much data as 8 bit and 4:2:2 is 1 1/3 times as much data as 4:2:0, so it is 5 1/3 times as much not 64 times as much.

10 bits of data is 10/8 times more than 8 bits. But you will learn it at school.

Link | Posted on May 9, 2017 at 04:21 UTC
In reply to:

Cary Knoop: From the article: "4:2:2 10-bit codec is 64x as much data as before"

No it isn't!

10 bit is 4 times as much data as 8 bit and 4:2:2 is 1 1/3 times as much data as 4:2:0, so it is 5 1/3 times as much not 64 times as much.

Regardless compression and quantinization 10 bits is only 10/8 more data than 8 bit. 4:2:2 is 8/6 more data than 4:2:0. So it's only 60% more data. And in many cases 10bit video can be compressed more than 8 bit with the same quality.

Link | Posted on May 8, 2017 at 22:44 UTC
In reply to:

ProfHankD: Technically, DFD can be better than PD. The catch is that it is computationally intensive and highly lens dependent for best performance. I don't know how lens dependent this implementation is, nor how they get the lens OOF PSF modeled (it could be lens data or it could be automatically determined, but I suspect they don't make it very lens dependent).

@zeratulmrye It does not work and Panasonic never said that. Panasonic said DFD requires lens calibration data stored in lens and only Panasonic does calibration and writes it to lens.

Link | Posted on May 8, 2017 at 12:13 UTC
On article Panasonic Lumix DC-GH5 Review (1180 comments in total)
In reply to:

John Koch: Panasonic should offer a firmware patch that includes a submenu with standardized modes for video tracking AF: "follow the ball" for sports, "largest face" or group of faces, backlit face, dark background, dark face / light shirt, dancers (back-lit, spot-lit, or sun-lit), etc. The multiple bar customization options are too difficult to select by trial and error.

Nice if DPR were to corroborate video AF comparisons to the OM-D EM1ii, a6500, 80D, or even the G80/85. Others have tried this, but more evidence is helpful. The existing tests may contradict because the subjects involve different light conditions or contrasts. Alas, the other models do better or fail less often.

Meanwhile, GH5 buyers who need reliable AF can: a) avoid wide apertures or ambiguous lighting or contrast; b) use MF and chain a close subject in one place; c) use MF, keep the subject at least 4 meters distant, and set the MF near infinity; or d) cut-away to b-roll footage, or any overlay, if all else fails.

Problem with GH5 AF is even if camera recognizes face and selects it with box camera fails to focus to it sometimes.

Link | Posted on Apr 21, 2017 at 07:20 UTC
On article Panasonic Lumix DC-GH5 Review (1180 comments in total)
In reply to:

mermaidkiller: PAL and NTSC ?
These standards from the analog TV age are completely obsolete in a digital age.
So it does not make any sense whether the frame rate is 59.94 or 60.00 or 50 or 72 fps.
So I set my video cameras to 60 / 30 fps (so-called 'NTSC') to get a higher frame rate while I live in the EU where NTSC was not the standard in the analog era.

@Rexyinc, not exactly true. In Japan TV is 60Hz, but electricity 50Hz.

Link | Posted on Apr 21, 2017 at 07:15 UTC
In reply to:

Nukunukoo: Two things that may be a good thing from the older gear 360: (1) At f2.2, the video quality may be slightly sharper and (2) the lower resolution may mean bigger photosites on each sensor for better low light performance and DR, of course with the appropriate bitrate and encoder. Hoping for a good hands on review...

Lower resolution means such 360 video is unwatchable. Even 2x 4K is no enough.

Link | Posted on Apr 2, 2017 at 08:21 UTC
In reply to:

Jefftan: I have heard RT can read data from RAW and correct MFT lens distortion, correct?

any others like Sony E-mount got automatic lens correction?
Thanks

That means it does not support

Link | Posted on Mar 10, 2017 at 19:03 UTC
On article Panasonic Lumix DC-GH5 Review (1180 comments in total)
In reply to:

dutch3dmaster: I pre-ordered mine :-)

One thing I hope for is that the GH5 will come with a more sophisticated time lapse mode. With 4K, 6K and other speedy technology already in place it should not be difficult to increase the frame rate in time lapse mode. Currently 1 frame per second is as fast you can go. A few frames per second would be nice.

Another thing I would like to have is the depth map which the camera seems to generate in combination with an image to turn it into 3D. Again something which should not be a hard thing to do. It could turn the GH5 into an interesting 3D camera!!!

Now you can use VFR in 4K too. It starts from 2 fps.

Link | Posted on Mar 8, 2017 at 20:20 UTC
In reply to:

Jefftan: I have heard RT can read data from RAW and correct MFT lens distortion, correct?

any others like Sony E-mount got automatic lens correction?
Thanks

So basically RT still does NOT support MFT.

Link | Posted on Feb 22, 2017 at 03:51 UTC
On article Extremely dramatic video touts Canon's CMOS technology (196 comments in total)
In reply to:

ecka84: Why is that every time some camera gear company produce a video, it seems like they don't know how to use their own products? The quality is always so low ... so compressed ... some kind of 480p upscale and crap like that. I don't understand ... are they trying to not sell the product, keep it to themselves? Why is that happening?

Not every company. Panasonic videos are now 4K by standard.

Link | Posted on Feb 22, 2017 at 03:46 UTC
On article CES video: the Panasonic GH5 (76 comments in total)
In reply to:

Joe Ogiba: The 6K Photo mode could shoot unlimited 18+ mp images @30fps !
https://youtu.be/rUnpEXaI1S8?t=1016

Use the 100-400mm (200-800mm FF eqiv) handheld in dual IS mode !
6K photo sample : http://lumixlounge.com/wp-content/uploads/2016/12/P1000044.jpg

@Ebrahim Saadaw
It's not limited and it is 18Mp video but in 4:3 or 3:2 aspect.
It does not have raw, but it can use AF.

Link | Posted on Jan 8, 2017 at 19:39 UTC
On article Panasonic Lumix DC-GH5 Review (1180 comments in total)
In reply to:

The Silver Nemesis: Finally: anamorphic 4K capture. Somewhere in 2009, Philips was right. But it was to early for the "consumer" to embrace this...
Panasonic: a sincere big bravo!

So, what is the point of anamorphic? There is no anamorphic TVs and it does not do any better for post processing.

Link | Posted on Jan 5, 2017 at 03:41 UTC
On article Modern Mirrorless: Canon EOS M5 Review (1644 comments in total)
In reply to:

Rambalac: What does "Number of focus points: 49" mean if it's Dual Pixel?

Do you mean it still can focus only in specific points?

Link | Posted on Jan 3, 2017 at 15:28 UTC
On article Modern Mirrorless: Canon EOS M5 Review (1644 comments in total)

What does "Number of focus points: 49" mean if it's Dual Pixel?

Link | Posted on Dec 30, 2016 at 17:34 UTC as 43rd comment | 2 replies
Total: 49, showing: 1 – 20
« First‹ Previous123Next ›Last »