Slashdot Mirror
EU Probes TVs Over Energy Test Scores
joesreviewss writes: The European Commission says it will follow up on evidence that Samsung and another TV-maker use software that alters their screens' power use during tests. The BBC reports: "One study indicates that some Samsung TVs nearly halve their power consumption when a standardised test is carried out. Another accuses a different unnamed manufacturer of adjusting the brightness of its sets when they "recognise" the test film involved. Samsung has denied any wrongdoing. It acknowledged that it used software that altered its televisions' performance during tests, but said this was the effect of a general energy efficiency feature that came into effect during normal use and had nothing to do with the testing process."
When there are money, power or sex involved people are willing to lie and cheat.
When 1person suffers from a delusion,it is called insanity.When many people suffer from a delusion,it is called religion
This is the same thing that VW did. It's rampant. Anyone who thinks that VW is the only culprit here is crazy. And anyone who thinks that these two companies are the only ones cheating energy regulations are crazy, too. I propose we solve this problem with some very effective deterrent. You intentionally cheat regulatory testing? Then all of the patents you hold are automatically invalidated. Kill two birds with one stone.
In other news, teachers awarded for teaching kids to perform at standardised tests.
Seriously. As soon as there is a standardised test, it's going to get optimized for.
Government sets absurd limits then companies cheat ...film at 11.
I'm pretty sure homosexuality is nothing like corporate dishonesty and greed.
Not limits, it is rating. You will simply be rated lower for consuming more energy. You have no reason to cheat... Film at 11.
It is of course easier to do with technology.
But the problem with standardised testing is more elaborated on here.
https://www.youtube.com/watch?v=J6lyURyVz7k
From 5 minutes onward.
Here's the thing, its not absurd if SOME of the TV pass the test. When they do pass, governments tighten the tests for the next generation of machines.
So car emissions wouldn't be so strict if every car failed them, and TV tests wouldn't be so strict if every TV failed them.
We know from the emissions test that BMW passed the *road* test and VW only passed the *rolling*road* test and failed the real road test. So the emission levels set isn't absurd, and the test is too artificial to properly check these limits.
THEY TEST BECAUSE THEY DON'T TRUST! The duty is on the test to catch the cheat, because if they trusted companies not to cheat, they wouldn't have the compulsory test!
VW *could* have passed the test if they'd also optimized their road driving to pass that emission test because BMW did pass it, even when it was done on the road test.
What the problem here is the test is not realistic enough. VW were caught out when the test was made more realistic, and then the emission authority played word games about 'subverting their test'. But really the emissions authority should have changed the test to be more realistic to catch all the car companies out.
BECAUSE THEIR JOB IS TO CATCH THE COMPANIES THAT DON'T PASS THE TEST!
Likewise this test video is not a proper test. They should have a suite of videos, only *some* of which are given as examples to the TV companies, to prevent any real or fake optimization effect.
You see my point, testing emissions and performance is an adversarial process. The job of the tester is there because companies might cheat, so the testers job is to make the test more realistic till the company cannot cheat.
VW was a failure of the tester to catch the cheating! They need to up their game.
I'm still surprised that anyone is surprised over this kind of "news". After all, PC and mobile hardware manufacturers have been tweaking their drivers for years to show better results on benchmarks.
But if company A is honest and company B cheats, then company A will be at a disadvantage until company B gets caught. If everyone cheats, then they are at a level playing field, even after getting caught (what if it turns out that all car and TV manufacturers cheat?).
I've a TV from Samsung (a non EU manufacturer) that senses background room light and can turn the set's brightness down as the room gets darker; I find the reduced contrast between room and TV improves the picture. Philips (an EU manufacturer) make TVs with "Ambilight", a set of lights on the TV to increase room illumination behind the set, which achieves the same effect, but of course at much increased energy consumption.
It's not just manufacturers that game the tests. ;-)
Volkswagen, Samsung? Anybody beginning to think that companies are acting like a bunch of fucking homos, acting straight when mum and dad come round?
Beginning? Lots of systems detect when benchmarking software is running and alter their performance. I think that a few Android phones do this as well, here is a Slashdot story about it:
http://hardware.slashdot.org/s...
It is dangerous to be right when the government is wrong.
Homosexuality may not be compatible with certain interpretations of christianity, but how would it be immoral? That makes no sense at all. Nobody else is being hurt or disadvantaged in any way by someone being homosexual. Furthermore, it is not a choice, but an inherent property of some human beings. Even if they would want to they could not change it, so by definition it cannot be moral or immoral.
In other news, teachers awarded for teaching kids to perform at standardised tests.
Seriously. As soon as there is a standardised test, it's going to get optimized for.
I used to work at the national Vehicle Engineering Lab, and one of the main complaints that we would get is that we do not publish the tests that we do to decide if a custom vehicle or modification is legal. This is precisely why, people will conform to all the published standards yet find workarounds for the intent. Make a vehicle safe, and we'll allow it. Overpower it with no safety features and we won't. Sometimes, wise decision makers are better then well-defined laws.
Of course, finding wise decision makers is very difficult. We were lucky to have a very wise, experienced engineer lead the team and the process, and I really don't know how it will work out when he needs to be replaced. This is why rule of the people was experimented with two thousand years ago, and hereditary government was found to be problematic three hundred years ago.
It is dangerous to be right when the government is wrong.
And it already impacts the quality of the devices. I have a plasma TV (bought a plasma because of better black levels) and it limits the white level if there is too much white on the screen, I guess to reduce the power consumption (whichi s actually very similar to a LCD TV of the same size). The thing is, I do not really care about the power consumption - as long as the breaker does not trip, I'm fine, after all, I do not watch TV that often. That, combined with some other quirks (huge input lag, unless set to game mode for example) makes me wish my car was bigger so I could have bought a used TV of older model - while that would have used twice (or more) the power, it would have most likely had better image quality.
How dare these energy-efficiency wonks want some kind of better world for our kids. How *dare* they!
I think that you hit the nail on the head! There is unfortunately no incentive to be honest. It's the same as in politics. If you are honest then you get no votes because you make no promises (that you know that you cannot honour) and when they find out that your opponent lied then they assume that everyone lies which rubs off on you.
Or sports for that matter, what good is it to compete without steroids when all that will happen is that people will assume that you take it anyway and if you don't you will have a disadvantage towards the ones who actually do.
The big question would be how we as a society can move from this absurd reality into a world where honesty actually would pay off. I don't know.
Since there's never going to be a better world, we're way past the "as good as it will ever be" stage and it's long since the downhill race started, why even bother?
Your negative outlook is noted.
For much of the world, things are getting better. Wars are down and poverty is down. It is only really for some of Western Europe and the USA, i.e. those subscribing to neoliberalism, that life is becoming tougher, and that's only been for the 35 years since the ideology started setting the scene.
Most people don't care that the industrie as a whole games test or uses substances that are cheap but unhealthy.
That is the real problem.
You're acting like people actually read the energy use sticker on their TV. Most people can't even find it under the size 300 font sticker on the front which says OMG Ultra-HD! 4K! SPECIAL! BUY ME NOW!
It makes sense in retrospect but I didn't actually realise TVs were energy tested. It made more sense to do it to fridges and things which are always on and can't be switched off easily. That's about the only time I look at ratings.
Volkswagen, Samsung? Anybody beginning to think that companies are acting like a bunch of fucking homos
I've got a significant number of videos of fucking homos. None of them use cheating on NOx emissions tests as a major part of the plot.
How dare these energy-efficiency wonks want some kind of better world for our kids. How *dare* they!
They *don't* want a better world for our kids.
They want us to use less electricity, because they want to cut generating capacity, because they won't let us us nuclear -- which even the president of Greenpeace has stated is the best option, given our energy needs -- and the only other options to meet demand during time when "green" energy is not being productive (and we don't have full capacity for it built, in any case, and building it will take electricity) are fossil fuels.
But they don't want us to use fossil fuels, either (I agree with them on that), so the only other option is to front-load energy consumption on these products in the manufacturing process, so that they take less energy while in actual use. Which is how most actually energy efficient televisions achieve their efficiency, even though, in net terms, they're less efficient. But hey, they get the little "EnergyStar" sticker because they pass the tests.
So our kids aren't actually getting a better world than they'd have gotten by simply increasing generating capacity in the first place, instead of shifting the energy costs (and environmental pollution) to coal burning plants in China.
But hey, at least it's not in our back yard, right? If we don't have to see the pollution where we live, then it's not really there. Just like when we were babies playing peek-a-boo, and new we could become invisible by pulling the blanket up in front of our face, or we know that, when we are on a date, if we eat desert off the other persons plate, they get the calories instead of us, and it's therefore less fattening.
In the electronics industry it is time honored practice to game the benchmarks. Slashdotters know of so many benchmarks and slowly the industry games them and makes them useless, for graphics cards or chip throughputs etc. It would be a surprise only if TV makers did not game the benchmarks. It is not just the hardware companies.
For example NASA tested the radar cross section of well known shapes as a benchmark to evaluate the correctness of simulation software. Shapes like the NASA almond, or icecream-cone. Users are expected to run the simulation software to test how accurate the simulation results are. Curious thing, create the model exactly with same dimension well aligned with XYZ axes you would get dead bang accurate answer. Rotate the geometry by say, 1.23 degrees about X axis and then 4.56 degrees about Y axis. The solution should remain invariant under a rigid body transformation, (remember to rotate the excitations in synch). Some software would fail horribly. Much longer run times and much lower accuracy.
But the free market comes into play here. Some vendors who don't cheat educate the customers about it. They will brag, "our solution will remain invariant under rigid body xforms, test our competitor! just test it and then get back to us". But when the testing is done by government the feedback loop takes so much longer to close. Like this VW incident.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
Here's the thing, its not absurd if SOME of the TV pass the test. When they do pass, governments tighten the tests for the next generation of machines.
So car emissions wouldn't be so strict if every car failed them, and TV tests wouldn't be so strict if every TV failed them.
We know from the emissions test that BMW passed the *road* test and VW only passed the *rolling*road* test and failed the real road test. So the emission levels set isn't absurd, and the test is too artificial to properly check these limits.
You're missing the point; it's like software tests. Bear with me.
Most software tests in the waterfall model -- including tests in xnu (that's the name of the Mac OS X kernel build component) and in ChromeOS regression tests -- tests for the bugs that have already been fixed. In other words, they are regression tests, making sure we don't repeat the same mistakes that we made previously.
There are a couple of problems with this kind of testing:
(1) You are unlikely to "un-fix" a bug that was bad enough to report, spend effort tracking down, and fix
(2) There's probably a big comment in the code to keep you from doing that anyway
(3) You always miss new bugs with this kind of testing, since new bugs aren't regressions
(4) It makes the build validation step slower and slower over time
(5) It has an incredibly low probability of improving the quality of the product delivered to the customer
So why in the heck do we test software this way?
Because it's something we can test; we had to write a unit test for it anyway, to verify that the problem existed in the old code, and was not present in the new. So we throw that unit test onto the pile of build validation tests. Because we can. Because it makes us look diligent. Because management wants us to. Because marketing doesn't want the black eye of a repeat of a previous problem, especially if it was an egregious one (new egregious problems are OK, though).
What does this have to do with emissions standards on cars? Why do we test for particular emissions levels? What research has linked pollution of a particular type to specific environmental problems?
We test to these standards ...Because We Can.
The NOX standards for Tier 1 were set at 1.0 g/mi. This was because, at the time Tier 1 standards were enacted, that was about the accuracy of our equipment. Tier 2 standards were phased in at 0.07 g/mi. This was because the equipment was improving, and that became our limit on accuracy.
In other words, the standards are what the standards are, and we test for those for the for the same reason we regression test software, rather than testing whether or not it meets the written specification:
Because We Can.
In other words, we don't set standards because they are environmentally reasonable, or technologically achievable on the engine side of things, we set them because they are testable on the emissions testing side of things.
BTW: Correct way to do software testing is:
(1) Write a spec
(2) Write tests to verify the software meets the spec
(3) Write the software to pass the tests
(4) Run the tests as build validation that the product meets spec following changes
You might run the individual bug verification tests before a major release, or if you suspect something, but it's a stupid way to run waterfall.
Test something because it's useful to test it. Don't just test Because You Can.
The companies cheat anyway, so the number on the product is meaningless (ranked first merely for topicality)
People buy a [thing] for its inherent qualities. A TV is chosen for its picture quality, size, features and price first. Energy consumption? The only person I've ever known to choose a TV based on energy consumption was a guy with a boat who wanted to keep his house battery consumption down.
Maybe if all products were essentially equal, energy consumption would matter. You could argue that fridges are basically the same, but when we bought ours size was the most important quality (biggest that would fit the opening) followed by the interior arrangement.
And does the energy consumption vary that much? Maybe the cheapest possible fridge uses a lot more energy, but once you narrow the field and eliminate the tails of the curve they seem awfully close.
If that is your experience then it says more about your experience than wider practice. Regression testing is simply ensuring that functionality that is already in place is not compromised when new versions/extensions/etc are added. Tests from the original specification for the system would be part of ongoing regression testing as the system is expanded/updated.
Somewhere a porn producer is taking notes.
But cheating on NOx testsvis so taboo! So very hot!
Not ratings, it is limits. Your television company will be barred from sales if it doesn't meet standards. You have every reason to cheat. Film in 2011:
http://www.energy.ca.gov/appliances/tv_faqs.html
This is why you (almost) can't buy a plasma TV anymore, BTW. I'm glad I bought my last TV in 2010. Looks great. I don't use it enough for it to matter, and even if I did, where I live, 50% of our power is nuclear, and more than half of what's left comes from hydroelectric, and about 10% comes from solar and windmills. Almost none comes from gas/oil/etc. And we have a surplus of power.
Sometimes, wise decision makers are better then well-defined laws.
Of course, the problem with government by unwritten rule is that the "wise decision makers" will then use that rule to punish their enemies simply for being their enemies.
Look at the IRS, which targeted people and groups they didn't like for harassment. That's allegedly illegal, but nobody was ever punished for it because "wise decision makers" didn't think it was necessary.
Look at the Secret Service, who did not like that a member of Congress was investigating them, so they leaked that a decade earlier, he applied for a job with them and they decided not to hire him. Allegedly illegal, but no one will be punished for that because the "wise decision makers" don't think it's necessary.
To turn it back to the automotive world, the "wise decision makers" in government decided to PLAY UP an unintended acceleration issue in Toyotas and COVER UP a similar bug in GM cars because the government owned GM at the time.
Part of the reason we have written laws is that they're supposed to protect us from government by the whims of "wise decision makers."
Increasing nuclear capacity is good for one interest group: nuclear industry. For the rest of us, though...
But only the diesel ones
I hadn't the slightest objection to his spending his time planning massacres for the bourgeoisie... (P.G. Wodehouse)
I do not know how it is whee you live, but in my country the energy class (A-G with some really efficient devices getting multiple As) is displayed on a sticker on the device and the sticker is quite large.
Also, isn't power consumption the reason plasma is less popular than LCD? It seems to me that plasma has better picture quality, but uses more power.
Part of the reason we have written laws is that they're supposed to protect us from government by the whims of "wise decision makers."
I agree with you completely. There is no form of rule which protects the people from the ruled, and also does not prescribe methods for causing harm legally.
A good ruler would balance the needs of the people against his own ego and family interests. I can think of very few modern rulers who I would nominate as being fair and just, the King Hussein of Jordan comes to mind. He's not perfect, but he's doing a better job than any other ruler that I can think of in modern times. He does an outstanding job of protecting his kingdom from threats inside and out, and additionally keeps his neighbors happy (ISIL notwithstanding). All other modern leaders who were good for their respective nations are considered as monsters by other nations. Ahmadinejad and Hitler come to mind. They were both fanatical about advancing their own nations, and did not care about what damage to other people was done in the process.
I venture to say that a "fair, just ruler" is a matter of circumstance no less than it is a matter of personality and values.
It is dangerous to be right when the government is wrong.
There will be a better world, just not for (99.99% of) us.
What absurd limit was set?
I prefer the "u" in honour as it seems to be missing these days.
Holy crap, I hope you have some QA people stepping in before you release software.
And it already impacts the quality of the devices. I have a plasma TV (bought a plasma because of better black levels) and it limits the white level if there is too much white on the screen, I guess to reduce the power consumption.
Maybe, you'd have to ask the engineers, but didn't Plasmas used to have serious problems with burn-in? That is another possible reason for limiting white levels. I can also see the possibliity of eye strain issues as a reason.
Get people all upset about asian tv makers so we can go light on VW and save those jobs.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
Volkswagens spewing filth: Not good.
TV sets spewing filth: I'm OK with this.
Have gnu, will travel.
When Samsung gets busted, they'll say they had no idea they did that... and they're probably right because it's probably something they stole from Sony haha
I didn't say it's not on the device. I said no one gives a shit.
Plasma wasn't just a bit energy inefficient. It was HUGELY energy inefficient consuming more than double the power of a typical LCD and in some cases triple.
By comparison most LCD TVs differ only slightly in consumption these days maybe +/- 20% and the ones that score highest score lowest on other features such as total brightness.
It comes down to the similar comparison of energy efficient light bulbs. People buy energy efficient bulbs over incandescent when they need to save power, but few if any people actually then look closely at the package to see which one ends up with the best lm/w. Going from 100watt to 18watt is a big difference compared to going from 18watt to 16watt.
Just as well, or there'd have to be an exception to rule 34.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
On the plus side, Samsung TVs still meet mandated limits for NOx emissions no matter what mode they are in.
I've got a significant number of videos of fucking homos. None of them use cheating on NOx emissions tests as a major part of the plot.
Rule 34. You pointed it out, someone has to make it happen. It might as well be you.
How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
Burn-in happens locally - so, having the entire screen black except for a 2cm by 2cm completely white square would burn the square just as much as having half the screen completely white. But having half the screen completely white lowers the white level, unless it is set low enough in the first place (took me a while to figure that out when setting the picture parameters). As for eye strain - well, I could be watching TV in a lit room, I also doubt that LCDs do that (where the lamp has to be on anyway independent of the picture on the screen).
The sad thing is that for a lot of people the image quality does not matter. What matters is the thickness, weight, features (fullHD, 4K, 3D , but not for everyone) and power consumption. The number of TVs set to stretch a 4:3 source proves it.
That was one of the reasons. I think potential burn in (though really image retention, since it's not permanent burn in) is the other major cause.
Plasmas at the time were LARGER, CHEAPER, and looked better. I think it's only now that we're getting into gigantic TVs (70 inch and above), that plasma can't do.. I also think it can't do 4K, or not easily. Those seemed to be reasons companies have stopped making plasma.
Also, from what I've read (popular press, online reviews), the more recent plasmas (not VERY recent, as in the last several years at least) were much more energy efficient.. The huge discrepancy compared to LED was, I think, based on early plasmas.
(I have a 42" plasma that I got when my CRT died several years ago. So I guess I'm sort of biased, but I basically went out and bought a TV very quickly, and plasma seemed to be the best and cheapest at the time.. I am someone who cares about power usage, but my overall electricity bill is low anyway -- $20-$30.)
I bought a plasma because I wanted image quality comparable to that of a CRT. As for 4K - no point in that until the source material can handle it. Last time I checked, BluRay was still 1080.
You forget one important thing: fixures are limited by how much power (heat) they can handle, so the more efficient the light is, the brighter light you get.
Forget magic. Any technology distinguishable from divine power is insufficiently advanced.
If that is your experience then it says more about your experience than wider practice.
Given that the regression tests used for both ChromeOS and the Mac OS X kernel are available in the publicly published source trees for each, I assure you that my experience is not unique.
Regression testing is simply ensuring that functionality that is already in place is not compromised when new versions/extensions/etc are added. Tests from the original specification for the system would be part of ongoing regression testing as the system is expanded/updated.
I said as much, when I made the post to which you are replying.
One of the process problems that both projects have is that a single test failure is considered a "build breaker"; and there is no distinction between:
* Tests which fail because they previously passed, and have regressed
* Tests which fail because they were written to verify product acceptance criteria, and that code is not yet written
By not having any way to allow a test failure to be acceptable during the development process, every test failure breaks the build, and shuts down progress for the rest of the team, while that breakage is dealt with.
This is, in general, one of the major dangers of utilizing a bazaar-model iterative process to approach a goal through successive approximation. The other danger is transiently flakey tests that are not disqualified as build breakers because they are considered important enough to break the build, but not important enough to track down (and fix) the reasons for them being flakey (usually differences in timing, e.g. cache creation on initial vs. subsequent boots, etc.).
P.S.: Apple and Google typically do not hire many formal Q/A types; tests are expected to be written by engineers working on the product as part of them working on the product, and are added to build qualification relatively ad-hoc. Ubuntu development testing works in relatively the same way; the closest Open Source projects get to formal verification are probably the compiler testing that happens with gcc and llvm.
I wouldn't draw that conclusion at all. Plasma had a major advantage over the LCD screens of the time, but those times changed before the end of Plasma's popularity. At the time LCDs had poor viewing angles poor brightness, and ghosting. These days I'd happily take an LCD.
At least a decent one. Cheap shit is still cheap shit, and cheap Plasmas outperformed cheap LCDs right to the end which didn't help the idea that LCDs sucked.
Yeah and what's written on the box? "40watt equivalent". People in general don't sit down and compare the exact spec on all the boxes trying to squeeze 50lm more out of a fitting.
All other modern leaders who were good for their respective nations are considered as monsters by other nations. Ahmadinejad and Hitler come to mind.
Both of those guys were horrible for their own people. Hitler eventually led Germany to ruin, and Iran spent most of Ahmadinejad's rule isolated on the international stage.
By contrast, I don't think Reagan or Bill Clinton are considered to be "monsters" internationally by anyone other than a group of fringe lunatics. (I'd also posit that both men are more well looked upon by the world at large than by citizens of their own country... ditto Margret Thatcher, actually.)
My point is that it's ridiculous for a standards agency to not publish the standard so as to defeat "attempts to cheat the test." Not having standardized rules understood by all the players allows the standards agency to attack whoever they don't like.
Americans dont read those. We are not conscious about energy usage or fuel usage. Everyone else does.
Americans dont read those. We are not conscious about energy usage or fuel usage. Everyone else does.
This practice is actually quite universal unless you're running your house from batteries. Americans may have a reputation for being worse, but they're really not.
At the time you were right to, these days LCDs are better than Plasmas were.
There are already quickly growing libraries of 4k on Netflix, Amazon, even Youtube has 4k content. Also if you use your TV for something other than videos like another computer monitor, running picture slideshows etc you also get the full benefit.
I'm not saying you're not right. I don't see the value in it right now either. But the library has grown considerably since the last time there was a 4k related discussion on Slashdot and it's growing every day, so you may want to keep a constant check on your assumptions rather than comparing the TV to the Bluray player.
Both of those guys were horrible for their own people. Hitler eventually led Germany to ruin, and Iran spent most of Ahmadinejad's rule isolated on the international stage.
Hitler got Germany out of the Treaty of Versailles. His mistake was commencing Operation Barbarossa. I'm Jewish and I lost a lot of family to that monster, but I realize that until Operation Barbarossa Hitler was doing what was best _for_Germany_.
Due to Ahmadinejad, Iran will have a nuclear weapon in 15 years. He set the stage for the current negotiations, which allow Iran to develop a nuclear weapon in 15 years without sanctions. He did what was best _for_Iran_ despite the short-term consequences.
My point is that it's ridiculous for a standards agency to not publish the standard so as to defeat "attempts to cheat the test." Not having standardized rules understood by all the players allows the standards agency to attack whoever they don't like.
You are right. However, there is no single test. People modify vehicles, from passenger vehicles designed for street use to putting electric motors on skateboards, and we test if they are safe. There is absolutely no way to make rules to cover all the possible scenarios. Put a 200 watt motor on a bicycle and we'll OK it for sale. Put two 550 watt motors on rollerskates connected to a poorly-soldered lithium-ion battery pack in a backpack and we won't.
It is dangerous to be right when the government is wrong.
However, there is no single test. People modify vehicles, from passenger vehicles designed for street use to putting electric motors on skateboards, and we test if they are safe.
I don't understand this part at all. If you have different classifications, and every vehicle has to past the test for their classification (and it's obvious how to figure out what classification your vehicle belongs in) you should publish ALL the tests. If your real goal is to allow manufacturers to pass the certification by making safe vehicles, you should publish your internal test procedures. If you don't do so, people will wonder whether your real goal is to harass manufacturers you don't like for whatever reason.