Root Access

bwDraco

18:00

On the APS-C side of the Alpha system, speed is the main emphasis, and the α6300 is loaded with some of the most advanced AF tech I've seen from Sony yet.

qasdfdsaq

Sony full-frame sensors are still miles behind even mid-range traditional DSLRs

bwDraco

Not sure how?

Sony is the #1 manufacturer of image sensors for a reason.

Canon's sensors don't quite match up especially in dynamic range. Several Nikon cameras use Sony sensors, while Pentax uses Sony across its lineup.

qasdfdsaq

Higher noise. Lower resolution. Lower dynamic noise.

Canon sensors are superior, and always have been. The lead is smaller these days, sure but still. Meh.

bwDraco

You must be kidding me. The numbers don't lie. Just check DxOMark.

qasdfdsaq

Sony also supply the crappy sensors in low-end smartphones. Just because you ship a large volume of junk doesn't mean it's any good.

bwDraco

18:03

@qasdfdsaq Those sensors are tiny, that's why.

What you're asserting is the exact opposite of what my experience tells me. Canon may have better image processing but that doesn't make up for sensor technology that falls behind the curve.

qasdfdsaq

> The 5DS R currently boasts the highest resolution sensor in a full-frame camera. In Raw, detail is properly resolved, and not aliased, to around 3900 lines. This is slightly lower than the Nyquist / √2 value of 4100 lines, possibly due to lens limitations (with the Otus 85mm, the 5DS R resolves to at least 4000 lines). Aliasing kicks in for the Sony around 3750 lines, and the Nikon D810 begins to alias soon after 3470 lines.

Canon has terrible image processing.

JPEGs out of the camera goes straight to the bin.

bwDraco

@qasdfdsaq That's sensor resolution. That says nothing for the noise at higher ISOs.

The EOS 5DS has the best sensor among Canon's lineup (dxomark.com/Cameras/Canon/EOS-5DS) but that pales in comparison to Sony's 42MP BSI-CMOS full-frame sensor (dxomark.com/Cameras/Sony/A7R-II).

qasdfdsaq

Well let's see, Sony's sensitivity range goes up to 1/4 of that of Canon.

bwDraco

@qasdfdsaq Irrelevant if you cannot get usable images at such high ISOs.

qasdfdsaq

And just as far behind high-end Nikons

Better an unusable image than no image at all.

And noise is almost always better than motion blur.

bwDraco

18:09

@qasdfdsaq Wrong. Nikon uses Sony sensors for the D750 and D810.

The D4 and D5 use in-house sensor designs but that's a different matter; these are designed for low noise over resolution.

qasdfdsaq

So what you mean is right, the exact opposite of wrong.

bwDraco

Also: dxomark.com/Reviews/More-pixels-offset-noise

qasdfdsaq

Nikon and Canon sensors go over ISO 320,000, Sony doesn't come close.

bwDraco

@qasdfdsaq Again, irrelevant if the image is unusable. Have you looked at the α7S II?

qasdfdsaq

There's a reason professional low-light photographers don't use the A7s II

bwDraco

18:12

Explain?

qasdfdsaq

10 mins ago, by qasdfdsaq

Higher noise. Lower resolution. Lower dynamic noise.

bwDraco

I'm not familiar with any issues specific to the α7S II only that it is optimized for sensitivity over resolution. Each camera in the α7 line is optimized for its intended application: sensitivity, resolution, or a compromise between the two.

@qasdfdsaq "Higher noise" is not an argument that can be made against the α7S II. This assertion is the exact opposite of what the camera is designed to do: excel in low light.

qasdfdsaq

Yes it is.

Though since you're determined to ignore everything I say, I don't see why you're even bothering to ask.

bwDraco

@DavidPostill (or whoever starred that last message): care to explain?

qasdfdsaq

Somehow I don't think Canon and Nikon would have both just released new DSLRs at twice the price of the Sony if they weren't confident they were going to sell.

bwDraco

18:20

@qasdfdsaq They have other factors in their favor. Camera systems are subject to vendor lock-in (switching systems is very expensive), and they have well-established and proven AF systems. While mirrorless cameras continue to make major advances towards DSLR-speed subject tracking (and faster-than-DSLR one-shot AF), there's no substitute for a system that you know just works for what you shoot.

Bob

@DavidPostill Attempted an answer shrug

qasdfdsaq

I somehow don't think 75% of people are buying twice as expensive cameras just because of vendor lock-in.

Especially when half of those are new, first-time buyers with no vendor lock-in whatsoever.

Bob

@qasdfdsaq That's actually an interesting question. What percentage of (bodies|$) are purchased by new buyers?

Hm.

There's obviously (?) far more amateur photographers than professionals. But the latter are also likely to purchase more equipment. And the former are more likely to pick a point & shoot or just a phone camera...

bwDraco

@Bob Compact cameras are scarce these days. The main (and usually only) reason you would want to use an ordinary small-sensor compact (ignoring high-end large sensor models) is zoom range.

For everyday shooting, a smartphone camera will do the job.

In fact, my Nexus 5X easily outperforms my Nikon COOLPIX S9900 in low light.

Bob

shrug I haven't touched a DSLR in over half a year. Too much setup work, lack of time and will to go take photos. Meh.

bwDraco

18:32

My DSLR is now mainly a camera for mission-critical images, when I need more image quality and speed than my other cameras can provide. (I'm a sports photographer by trade.)

Well, a key issue with mirrorless cameras is that unless you use an electronic shutter (which tends to degrade image quality), the live view feed is disrupted whenever you take a picture because the shutter must close and recock, release to actually take the picture, then recock and reopen again to restart the live view feed.

During continuous shooting, the shutter is recocking between shots which means that there is no live-view feed between shots—you only get to see the last shot taken.

Electronic shutters avoid this issue, requiring nothing more than a brief interruption between shots, but increases noise levels in low light (at least with current technology).

Pentax K-3 II: Continuous shooting in Live View

►

Notice that the vast majority of the time spent during the sequence is simply for recocking the shutter, moving the shutter curtains back into place for the next exposure.

qasdfdsaq

@Bob I haven't bought a new camera in 4? 5? years?

I have a low-end DSLR and have lost more lenses than the total cost of all DSLRs I've bought -_-

DavidPostill

@Bob Good answer (after having gotten the extra info from the OP) ;)

qasdfdsaq

@DavidPostill Lol XY problem again I see

Turns out all he wanted was... 5v power.

Sure he could just connect up a pin/cable to the 5VSB, or just connect any old USB device -_-

Cutting the cable completely is a bit extreme

Anyway, I've gotta go rescue some Kerbals I left in limbo last night cause I had a headache

bwDraco

18:48

Alright, have fun gaming. See ya.

qasdfdsaq

Urgh my mouse is sticking

DavidPostill

@qasdfdsaq You needs to cleans its balls ;)

qasdfdsaq

It has no balls

I've not had mouse balls for a decade

allquixotic

@Bob Is it wrong that I am seriously considering Intel QSV for a work project? :P

bwDraco

@allquixotic QSV has limited control over the video quality. I'd much prefer to use a proper GeForce or Radeon card for video encoding; dedicated graphics cards have more powerful video encoding hardware that allows higher quality settings.

allquixotic

18:56

@bwDraco so would I, but our work computers have either no hw-accel video encoding at all (old, old boxes), or QSV (the latest-gen boxes).

bwDraco

For maximum flexibility, I'd just do CPU encoding.

Slow but generally produces the best results.

allquixotic

fidelity isn't very important

bwDraco

@allquixotic In that case, QSV is likely to be usable but do make sure the quality of the output is satisfactory. You may need to make some adjustments, and it won't be as optimally compressed as, say, x264/HandBrake's "Very Slow" preset.

allquixotic

yeah, but you can't compress using the CPU and libx264 or libx265 using the "Very Slow" preset in real-time on a laptop Haswell i7, either

given the real-time requirement, probably the best compression efficiency we can manage on these systems will be via Intel QSV

bwDraco

@allquixotic Yeah, unless you have a super-expensive Xeon or HEDT processor. <15 fps is not unusual for 1080p video (I get about 10-12 fps on my system).

If it's live video, than quality isn't the biggest concern. I'd use QSV.

Just make sure you have the right quality-speed balance. Monitoring with GPU-Z can help ensure you're not overworking the Intel HD Graphics GPU and dropping frames (which is worse than a low-quality feed).

allquixotic

19:03

the GPU's other work besides the QSV encoding will be to render standard "enterprise IT" style webpages in IE 11 or Chrome

nothing flashy; 2010-era web "1.9" / "2.0.0" type crap

bwDraco

I'd tune it to let the GPU run to about 75-80% of capacity, leaving enough headroom for light Web usage.

Once again, I'd strongly recommend that you monitor your GPU load using GPU-Z.

allquixotic

probably not going to capture at 60 fps, either -- to save bandwidth and because it doesn't really matter -- I'd say 30 fps max, probably even lower for some people, depending on their preference

we'll never be encoding data that resembles video or games or 3d rendering of any sort

it should compress fairly efficiently, too, with lots of solid blocks of color

bwDraco

Let me guess. You're streaming a live presentation to colleagues?

allquixotic

will have to make sure the fonts are legible though, that'll be the biggest challenge

@bwDraco recording Selenium running through automated tests (clicking and typing through a web UI like a user would) so the tester can see "what happened" if they get up from their chair and find that the test failed

(and to provide better defect write-ups, videos attached can help dev troubleshoot)

bwDraco

Oh, so you're screencasting.

allquixotic

19:08

well it's writing to disk

bwDraco

That should not require a lot of compute to stream efficiently.

Bob

@allquixotic Heh. Spoon? :P

allquixotic

capturing from GDI+ (the ugly old layer in Windows for capturing the UI Without a custom driver) and writing to disk

Bob

(I think they offer pretty much exactly that as a service...)

bwDraco

You're doing software user acceptance tests?

allquixotic

19:08

@Bob damn, didn't realize you were still up / probably asleep, sorry to wake you

bwDraco

(I've been taught this in software engineering class.)

allquixotic

@bwDraco UAT's a separate division ;p

Bob

@allquixotic nah, I should be asleep but shit happened

@bwDraco This is more like regression testing...

allquixotic

@Bob that implies that Spoon could get to our apps. trust me, they can't. too much security

@Bob ^

Bob

@allquixotic Yea, I think the Selenium stuff is hosted :\

allquixotic

19:10

Functional testing

Functional testing is a quality assurance (QA) process and a type of black-box testing that bases its test cases on the specifications of the software component under test. Functions are tested by feeding them input and examining the output, and internal program structure is rarely considered (not like in white-box testing). Functional testing usually describes what the system does. Functional testing does not imply that you are testing a function (method) of your module or class. Functional testing tests a slice of functionality of the whole system. Functional testing differs from system testing...

Bob

The browser sandboxes run locally though.

bwDraco

@Bob: Remember that allq is working for a government contractor, so security is paramount.

allquixotic

that's my main job function ^

bwDraco

I'd better get an internship to learn the skills I need :\

Bob

@bwDraco I am well aware. But that does not automatically mean you can't use external services.

allquixotic

19:11

@Bob without going into too much detail, there are diverse opinions around using "cloud services" on government projects within the US Federal IT sector... meaning, some organizations categorically disallow it; some conditionally allow it after careful review; some embrace it willy-nilly.

I've been in the Federal IT sector long enough to have worked with clients all across that spectrum.

Bob

@allquixotic Ya, it's the same in private.

Though usually things like testing are more likely to be approved than hosting. Much lower risk.

allquixotic

my "main" client is the sort that allows it after careful review (high level of bureaucracy and cost involved in determining the security requirements are met)

Bob

Anyway, just threw it out there as something you reminded me of :P

allquixotic

meaning, to just bring in something like Spoon would cost boku bucks

Bob

@allquixotic Turns out it's not really cloud anyway, apparently :P

> All browsers and test scripts run on your local machine

allquixotic

19:14

even to get it hosted internally on the LAN, if they were to sell the software or a hardware appliance, would be months worth of tedious effort, business justifications, etc

@Bob oh huh

Bob

> The security benefit of this is that no browser data or network traffic passes through Spoon servers. The only data stored on the server is the test results, which can be viewed in your online account. You can turn off test result storage by unchecking the Save test reports check box in the top-right corner of the spoon.net/selenium.

allquixotic

in that case, if it were free then I could probably just use it... the de facto security posture towards FOSS seems to be that if it doesn't show up in Add/Remove Programs, it's not a program, so it's fine

Bob

So, it still depends on how much you trust third-party software, like any other. But it's probably much easier than actual 'cloud'.

allquixotic

(which is silly, but whatever)

Bob

@allquixotic Yea, nah. Proprietary, paid, subscription.

shrug

Anyway, you'd probably be fine with just 15fps. Maybe even lower.

And there's minimal changes frame-to-frame so compression should be pretty good.

Dropped frames, etc., aren't much of an issue.

allquixotic

19:17

I was thinking of giving the user free reign over dictating the FPS, with min and max bounds checking between 1 and 30 fps ;p

Bob

If you're just trying to see the test in action... meh.

Nowhere near as demanding as full video.

bwDraco

@Bob Skype is 15fps, and that's never really an issue for some of the collaborative work I've done with my classmates.

allquixotic

the biggest criterion is really "Can I read the text?"

if yes, then everybody's happy

Bob

So your requirements are pretty minimal and well within the capabilities of QSV. Or even software encoding if you wanted.

@allquixotic Well, end of the day you're looking for something similar to RDP, VNC, Citrix, etc.

allquixotic

well, I'll be falling back to software on our old Core2 boxes; I'll have to note to those specific users that they should definitely keep the fps at 10 or 12 or so

@Bob except that the result gets written to disk in hopefully a small file

Bob

19:19

IIRC RDP can manage some stuttery ~5 FPS for full video, though this was back in 2009.

Some tricks to get it much smoother for partial changes.

qasdfdsaq

RDP can easily manage 60fps these days.

Bob

@qasdfdsaq I've still not managed to actually watch a video smoothly over it :P

allquixotic

I've gotten native-looking youtube videos playing over RDP from Windows 10 to Windows Server 2012 Standard on a 100 Mbps local uplink at work, connected to a gigabit uplink on my dedi

qasdfdsaq

Lossy encoding, UDP, and native stream passthrough

bwDraco

On to another subject... memory cards are so much cheaper than they once were. About 3.5 years ago, a 16 GB pro-spec memory card costed $45. The same amount of money today gets you a 64 GB card that is just as fast.

Bob

19:19

Last tried that in early 2015? I think.

qasdfdsaq

But requires Windows 8.1+

Bob

Ah.

qasdfdsaq

That's the main reason I bought Windows 8.1 in fact, as the RDP is a huge improvement over W7

Bob

Yea, the only 2012 servers I have access to are over a slow internet link.

bwDraco

@allquixotic I've used RDP to access my laptop from college over the Internet with a 1.1 Mbps uplink. It's slow but usable. That shows how efficient RDP is these days.

Bob

19:21

@bwDraco It's always been capable of that.

qasdfdsaq

It makes a very big difference over slow connections too,

bwDraco

RDP lets me keep my laptop at home when I don't need to take it with me.

Bob

I've seen RDP used with <1 Mbit up ADSL with 3+ streams at once. Not sure what the actual bandwidth usage was, but they were fairly smooth.

And that was back with XP.

qasdfdsaq

<1Mbit ADSL up is why I got it

I was used to 20Mbps up in my last place, so W7 RDP was fine

bwDraco

With 5 Mbps, RDP will run very smoothly, almost as good as if you were sitting right at the console. At that point, latency is the main limiting factor, not bandwidth.

Even 3 Mbps will give you an excellent RDP experience.

Bob

19:24

@allquixotic Actually - we end up recording lots of tests&samples as GIF and it works pretty well (easy to chuck on an issue tracker). A proper video codec should be even better.

8 FPS, 5 FPS, all fine... should minimise the resulting file size

You'd probably be fine with software encoding :P

bwDraco

@allquixotic: Are you using something like OBS to stream the video? You could just capture the whole desktop surface and stream it efficiently at 15 fps, and it doesn't take high bitrates to do it. 5 Mbps is probably more than enough for the application you describe.

Bob

(granted, we don't record a whole 1920x1080 viewport... more like <800x600)

bwDraco

@Bob You can delta-encode animated GIFs in a fashion similar to proper video codecs. It's not the best format but it's universally viewable.

Bob

@bwDraco Still nowhere near as efficient as a proper video codec.

allquixotic

@bwDraco again, not streaming...... recording to disk.... but doesn't OBS just use ffmpeg under the hood?

Bob

19:28

Though you do gain some size reduction simply from the reduced palette...

bwDraco

@allquixotic OBS supports QSV.

Bob

@bwDraco GIF rendering is actually subject to all sorts of weird quirks.

bwDraco

@Bob 256-color palette causes all sorts of issues, I know.

Bob

@bwDraco Far more than that.

Timings are often way off.

bwDraco

@Bob Yep. GIF only supports 10ms precision for frame timing.

Bob

19:29

@bwDraco Nup. Not even that.

22

A: Why is this GIF's animation speed different in Firefox vs. IE?

Your GIF animation contains 40 frames, with a 0.03 second delay specified between each frame. That should give it a nominal runtime of 1.2 seconds per loop, which matches what I'm seeing here in Firefox. The problem seems to be that many browsers will ignore very short per-frame delays and repl...

Have fun.

And there's not even any guarantee that future renderers will not enforce higher timings.

bwDraco

...and most GIF decoders will slow down instead of dropping frames if they can't keep up.

Bob

Oh, Firefox introduced another fun one a few years back. If it loads too slowly (because GIFs are bloody massive in terms of bytes-per-frame, deltas or not), at some point it just loops early.

You might eventually see the whole thing after a dozen loops depending on the load speed.

bwDraco

Oct 17 '15 at 18:57, by DragonLord

@tereško Sorry, but I have to flag this for mod attention. The animated GIF is 73 MB in size, which is dragging my connection down.

Animated GIF is very inefficient indeed.

Bob

...that's what I said about 10 mins ago.

Its only saving grace was it being playable in browsers without plugins, often after some fiddling. Which is pretty much gone now with <video> support.

Unfortunately, many webapps and sites have not kept up.

... SE Chat, GitHub issues, JIRA, etc.

bwDraco

people.mozilla.org/~dolske/apng/demo.html

Fails in Microsoft Edge but works in Firefox.

Bob

19:39

Ah. APNG. The red-headed stepchild of animated image formats.

s/Microsoft Edge/[^Firefox]/

bwDraco

I believe Firefox is the only major browser with APNG support.

Bob

Huh, Safari does. Fancy that.

qasdfdsaq

> A man died in a house fire caused by the overheating of an iPhone 5 battery which was plugged in to charge overnight, an inquest heard.

Figured it'd happen one day...

bwDraco

Not new. There have been similar incidents involving Galaxy S-series phones.

mashable.com/2014/07/28/samsung-galaxy-s4-fire/#27pZLbG0WEqB

qasdfdsaq

I don't see any mention of anybody dying

bwDraco

19:47

...using the Android 6.0 experimental multi-window mode.

Firefox on top, Chrome on bottom.

qasdfdsaq

So it is new then

Bob

@bwDraco Been in LG and Samsung customisations for a while, IIRC.

Though it's nice that it's expanded to all apps now.

qasdfdsaq

@Bob What, iphones kililng people?

Bob

@qasdfdsaq Wrong message :P

bwDraco

@Bob I believe Samsung was the first to this feature and has the best implementation.

Bob

19:49

@qasdfdsaq Does that include those who drove off bridges?

qasdfdsaq

@Bob Has that actually happened?

Bob

(jab at Apple Maps)

qasdfdsaq

I've heard of close calls but not aware of any actual deaths

People being told to drive onto train tracks, etc. and escaping an actual train only by running out of the car

Bob

@qasdfdsaq Well... not a phone but still cnet.com/au/news/…

qasdfdsaq

o_0

Close enough :-P

Sadly it wasn't the user who died

Bob

19:52

Plenty of idiots texting while driving too.

Unfortunately that often (more often?) results in injuries and deaths of other people.

qasdfdsaq

Urgh £900 a month on food

SleepingGod

Hi

Bob

'lo

SleepingGod

:P

Bob

20:26

qasdfdsaq

21:15

bwDraco

@qasdfdsaq Are these some sort of rodent? They're so adorable...

Michael Frank

@bwDraco Weasels or stoats?

oh, Ferrets!

bwDraco

Ahh. That makes sense.

Always wanted some mice or rats as pets, but I'm not sure if I can commit to taking care of animals on a daily basis.

Michael Frank

TIL Weasel is the greater classification.

qasdfdsaq

Yeah I wanted rats too

But I also like cats too much

Michael Frank

21:24

I have 3 guinea pigs.

and a cat.

bwDraco

Keeping any pet is a big commitment, even though rodents are easier to care for than other pets.

They also don't live very long.

Michael Frank

21:46

For anyone that's interested: ITPro.TV is doing 50% off Premium Annual or Premium Monthly accounts with offer codes DENVER50 or CAROLINA50. Offer expires Superbowl kick-off, 7th Feb 2016.

DavidPostill

Hmm. That looks like spam ;)

Michael Frank

heh...

I just think it's a solid offer for anyone looking to skill up.

Bob

...what is it?

DavidPostill

@Bob "ITProTV blends entertainment, humor and cutting edge technology with IT education and presents it all during a live broadcast. The shows are then organized by exam and exam objective or by technical skill and added to our Course Library."

Bob

22:11

Ah.

bwDraco

23:56

amazon.com/…

Compact 10,000 mAh power bank.

Three high-density cells, unlike competitors which often use four cells to achieve the same.

What I like about Anker is that they're not holding anything back with the quality of the materials they use. They don't skimp by using older, cheaper 2600 mAh cells. These power banks use premium Panasonic 3350 mAh cells.

...and they do this at extremely competitive prices.

Anker power banks have never failed me.

Transcript for