WebP is obsolete. It's still based on VP8 codec, which in video has been replaced by VP9 long time ago. AVIF is based on AV1, which is a successor to VP10. So WebP is a few generations behind in the VPx lineage, and is no match for modern codecs.
AVIF was a quick hack originally by Netflix by placing an AV1 frame into a HEIC container. I believe it was done in a few weeks of work.
AV1 was largely based on VP9/VP10 and was developed by a team working in Chrome organization.
JPEG XL main mode (VarDCT) and the JPEG recompression is largely developed by Google Research.
WebP as a format was based on VP8, a video codec built by On2 Technologies. On2 was bought by Google in 2010 -- a year before Google published WebP. The transparency and lossless encoding as well as non-video keyframe-by-keyframe animation were designed at Google. The On2 VP8 codec used initially in WebP lossy was not that suitable (too many artefacts) for photography transmission. Jeff Muizelaar wrote a great blog post about this. The codec for WebP were redesigned (without format) changes at Google, and kept improving significantly until around 2015 when it reached pretty good maturity.
(Personally, I don't like what it does to highly saturated dark colors, such as dark forests or dark red textures, but it is much much better than it was.)
WebP was the classic Google “ship the prototype” move - they were hoping Chrome could muscle it through but it delivered only modest compression improvements (10-15% real world - the marketing papers promised 30% based on comparisons to unoptimized JPEGs) but it was missing features and had very primitive software support, making it harder to produce, deliver, or share (when Facebook switched, a common complaint was someone downloading a picture and finding it didn’t work in another app or when they sent it to a friend).
Very few sites pay for so much outgoing image bandwidth to make that compatibility cost lower than a 15% savings.
it's still a version of jpeg that supports transparency, and it's actually well supported (down to ios 14) to use without also having to deliver a fallback format now. it's not as good as it's successors, but if you are chosing one format and care about image size, it's the best choice.
I’m not saying it was terrible but that it took a long time for it to be worth the trouble unless you really needed transparency. It’s only been the last year or so that you could expect to be able to use it for anything non-trivial and not spend time dealing with tools which didn’t support it.
it did, but mostly because safari dragged its feet for years. Thank god they didn't take this long for AVIF (though I would have loved it if they had shipped AVIF in io15, since a bunch of devices won't get ios16)
Safari was the least of it - most image processing tools didn’t support it (e.g. PhotoShop got support last year, Microsoft Paint was the year before) or you had to do things like recompile the versions of common tools to add support (again, better now but it takes a while for support to spread through Linux distribution releases), and now you have more security exposure. That was a lot of hassle for very modest compression gains.
AVIF has gone better because it wasn’t based on a video codec which was never really competitive, was developed collaboratively, and didn’t have feature regressions from JPEG. As with tool support, that last matters a lot at many organizations because the edge experience tends to decide things - even if 95% of your usage is boring 8-bit 4:2:0 the institutional memory tends to be shaped by the times you hit something which can’t be used without more work. If it compressed as well as AVIF, more people might have decided WebP was worth it but since it only marginally outperformed JPEG the case was never that strong.
Part of what I meant by “shipping the prototype” was this kind of stuff: someone at Google wanted to find another use for the On2 IP they’d purchased so they tossed it into a 20 year old container format and shipped it. As with WebM, the benchmarks were fast and loose which meant that anyone who replicated them saw substantially lower performance, which is another great way not to build confidence in your format.
I switched to WebP about a year and a half ago. I’d been watching for a long time and it had finally reached the point where support was universal enough that I could not publish a JPEG.
WebP has the big advantage that the quality setting is meaningful, you can set it at a certain level and then encode thousands of images and know the quality is about the same. This is by no means true about JPEG, if you are trying to balance quality and size you find you have to manually set the compression level on each image. Years back I was concerned about the size of a large JPEG collection and recompressed them which was a big mistake because many of the images were compressed too hard.
In 2023 I think you can just use WebP and it will work well, my experience looking at images is that AVIF does better for moderate to low quality images but for high quality images it doesn’t really beat WebP.
> Years back I was concerned about the size of a large JPEG collection and recompressed them which was a big mistake because many of the images were compressed too hard.
Distortion metrics[0] such as MS-SSIM, MS-SSIM*, SIMM, MSE, and PSNR can be used to define a cut-off or threshold for deciding the point at which the image is "compressed enough" by using one or more of those algorithms and predefining the amount of acceptable/tolerable distortion or quality loss. Each of those algorithms has some trade-offs in terms of accuracy and processing time, but it can definitely work for a large set if you find the right settings for your use-case. It is certainly more productive than manually settings the Q-level per image.
Actually this isn't accurate, Kraken.io doesn't use any SSIM-related algorithms, it just blindly applies some standard compression regardless of the image's content.
You can use jpegli for better jpeg compression heuristics. It uses custom heuristics that were originally used in JPEG XL, then copied over and further optimized using nelder-mead to minimize distortion metrics (butteraugli and simulacra 2)
There is plenty of wrong or misleading information here:
- What is the cross supposed to mean for PNG compression of photographic images? PNG can compress photographic images just fine and for some applications (where you want lossless) it used to be a good choice.
- PNG has animation support, even if everyone except Firefox tried to
- While WEBP and AVIF support lossless compression and animation, those features are not available in all browser versions that support static lossy webp/avif images.
a) This is a property of the encoder, not of the image format.
and
b) Like any other video codec-based format, webp overcompresses dark areas in images so no, you can't rely on consistent quality across collections of arbitrary images.
France has been mismanaging its fleet and signally for years that they want to move away from it, under-investing in it, etc. It's not the technology's fault.
Germany had some of the best managed plants in the world -- until they decided to shut them down, leading to more coal being burned and more dependence on other countries like Russia...
Germany is exiting both nuclear and coal. In the time from 2010 to 2022 where 14/17 of nuclear plants were shut down generation from coal was reduced from 263 TWh per year to 181 TWh. Renewables increased from 105 TWh to 254 TWh. I also would have preferred to leave the nuclear plants running longer and exit coal faster, but in the overall scheme of things it does not matter too much. Nuclear is basically irrelevant. It is too expensive and slow to build. In reality, renewables will take over everything very quickly.
Gas and coal imports from Russia stopped completely. But guess what still depends on Russia: The nuclear industry in Europe and the US.
Also, the reason this is only possible at all is that Germany uses the rest of Europe as a giant battery to manage the non-dispatchability of renewables. The import-export balance often changes by as much as one third of Germany consumption in 12 hours [1].
The electricity prices are also becoming zero[2] in Germany during parts of the day, which is a great outcome only on the surface. As this progresses, the consequence will be that renewable electricity producers aren't getting paid during their prime generating hours. This means even more subsidies will be required going forward to bring additional production. It will become more apparent once the reserve of easily dispatchable electricity sources is fully tapped to balance renewables across Europe. We will see very high prices during mornings and evenings and whenever it's cold and dark. The fossil fuel plants that are turned on during these periods will need to earn enough to address the additional wear due to quick power cycling and to keep being maintained for the rest of the time when they are unused.
The grid is for trade, so I am not sure what sense this complaint makes. Also France relies on imports sometimes. Electricity prices changing with production and demand is also exactly what market is for. As long as the market is working, this is beneficial to buyer and seller and changing prices are signals that allow the market to optimize production and consumption. Even when the price becomes zero sometimes, that does not mean than renewables need more subsidies if they earn money at other times where the price is higher. Also Germany still has enough of conventional generation capacity to ramp up production if needed, so the "is possible at all because" comment is wrong. If there is trade, then because this is cheaper overall (and in general this helps buyer and seller).
> As long as the market is working, this is beneficial to buyer and seller and changing prices are signals that allow the market to optimize production and consumption.
The market will always look to extract the highest possible price from the consumer.
> If there is trade, then because this is cheaper overall (and in general this helps buyer and seller).
Strange how "cheaper overall" fantasy is "consumer energy prices have quintupled in the past few years" in reality
> dispatchable electricity sources is fully tapped to balance renewables across Europe. We will see very high prices during mornings and evenings and whenever it's cold and dark
The prices in summer were so high mainly because half of France nuclear plants were offline. Look, even the article you cite mentions this: "The largest Nordic nation became the region’s top exporter in the first half after France suffered problems at its aging reactors"
Again: total numbers mean absolutely nothing on a night when there's zero production from renewables.
That "shill" is showing data. Just the fact that you don't like this data doesn't make it invalid.
Right now it's night in Germany, and even though the wind is blowing, it's only at 40% generation. And look, there's coal, supplying 22%, and gas supplying another 9%: https://imgur.com/a/3bYudyd
So what? You are complaining that Germany still uses coal. This is a fair complaint (and one should complain about this), but Germany is in the middle of the transition to renewables and that in the middle of the transition you still use coal does not tell you that the transition is not working. In fact, there are plenty of simulation studies which show that it will work. We also know that nuclear will have no meaningful impact in fighting climate change because it is too expensive and too slow to build.
> but Germany is in the middle of the transition to renewables and that in the middle of the transition you still use coal does not tell you that the transition is not working.
No, it shows that people believe in bullshit and actively hurt their own clean power generation capabilities due to decades of FUD.
And yes, however many renewables you build, you still have no answer for intermittent power generation.
> We also know that nuclear will have no meaningful impact in fighting climate change because it is too expensive and too slow to build.
Strange how some countries build them relatively fast (China can builds 1 reactor in 6 years, building several reactors in parallel, see Fuqing Nuclear Power Plant) while e.g. in Europe after three decades of FUD, scremongering and underinvestment, we not only believe they can't be built fast enough, but can't build them either.
I don't get this bogus argument. Germany is using the least coal in its history for electricity. The data is very easy to find, but it's a knee jerk reaction that they got rid of nuclear and so they must've replaced it with coal. Not true at all, and it's getting tiring reading this nonsense.
50% of the German electricity are now renewable, wind, solar, biomass, water. In 2022, Germany was creating and exporting a considerable chunk of electricity to help plugging the holes left by the switched off French reactors.
So it matters when wind is providing 20%, but it doesn't matter when it's providing 100%. Thus, even though coal use is at a historic low, it's actually increasing, and in other news water is dry and grass is red.
> So it matters when wind is providing 20%, but it doesn't matter when it's providing 100%.
Of course. Because when there's no electricity, there's no electricity.
> even though coal use is at a historic low, it's actually increasing
You pretend this is a contradiction when it's not. Germany has just shut down its last reactors. So yes, the usage of coal will increase because when renewable generation is low you still need to provide electricity. Guess what provides that electricity.
> Your claim was that Germany replaced nuclear with coal. Here are the hard numbers:
Funny how these hard numbers don't answer a simple question: where does Germany get its power on a quiet night like April 15th (answer is simple: buring coal, gas, and biofuels).
> BTW with what did France replace its failing reactors last year?
Last year it's "failing" generators were taken of for planned maintenance. The actual failures accounted for 0.18% loss in power.
Did you had a look at what Europe didn't stop importing from Russia due to lacking alternatives... cough cough.
I also don't get how an article can claim a global solution if nuclear currently is at 10% and we cannot keep up with supply and plants and and and.. this wont scale to 100%, not even 50 or even 20% of future global needs, would deplete cheap enough enable resources too quick... and what also is always forgotten: our world will likely have more, not less conflicts, unfortunately. Have fun managing this plants with wars all around and rivers going empty.
Yeah yeah all the issues happening and brought up again is just stupid people, mismanagement, etcetc.. But that unfortunately is humanity :/
Just no :/ especially as there are good enough and more sustainable other alternatives.
> Germany had some of the best managed plant
Better than France, likely agrees.. but still awful and that sentence is a joke. If that is your bar, good night. German infrastructure is currently rotting at record pace, happy we got those plants out there.
I am cautious on technology vs management. It's like the people who thinks death penalty is great in theory but the justice system is imperfect in practice: maybe, but you can't have one without the other. Nuclear plants have to be managed and that's an issue too.
The US Navy could be quite quick building reactors and innovating, because it didn't have to go through the same amount of regulation. They've operated lots and lots of reactors with very little problems for decades.
Go figure when the government owns the reactor and hires technicians to maintain them, they can be built "quickly".
In an earlier post here
> The truly unique feature of US nuclear power is the unlimited power that was given to federal regulators.
> [...] Congress had effectively told the regulator make the rules up as you go. This meant the regulator had no problem changing the rules. A design that was legal at the start of construction, could be declared illegal any time thereafter.
The solution may be that the federal government owns the plant and contracts out the work. Ultimately, the government would have to answer to itself about regulation changes.
It's because the US Navy can just decide to do it without years of red tape and mountains of paperwork. National security cuts through the red tape.
I don't think the government owning civilian reactors would help if they don't reform the process that means it takes 10+ years and millions of pages of documentation to get anything done.
This particular proposal is based on the idea that if renewables are cheap and getting cheaper (which they are) then you should build a lot of them.
They claim the lowest cost model for generating 100% renewable involves 5x the capacity being built.
However, they recommend rolling out 12x because that would provide another 500Twh of energy for only a 20% increase in spend (while reducing the amount of battery needed by 50%). A concept they have cheesily named "superpower".
While nuclear is one of the most expensive electricity sources, with surprisingly limited fuel reserves given the energy density, it's incredibly safe (despite the reputation) and I think there is inherent value in a diverse supply that can make up for the sticker price.
Also, China is able to build it at a reasonable price (~$40/MWh). It's not destined to be expensive, the price is just a consequence of excessive safety precautions.
Note that this article says could not will, so I’ll answer a related question of yours that better reflects the article:
> How could they deal with variability of all these renewables?
I’m guessing a mix of pumped hydro, purchased hydro and nuclear, home storage, robust grid, increased energy efficiency, smarter usage periods for heavy users such as industry or EV charging, etc. And then there is emerging technologies such as on-site carbon capturing of natural gas power plants, liquid metal/molten salt batteries capable of robust large scale grid storage, etc.
Solutions do exist, and no one of them will solve it for all, but together they will.
A large variety of places, but not by replacing 100% of electricity and only electricity in one region at a time.
You replace 30% of electricity. That's easy.
Then you replace a bunch of dispatchable non electric loads and use the times the new generation can produce electricity to replace another 20.
Then you add a bunch of 4 hour storage and dirt cheap thermal storage. You get another 20.
Then if there's finally a breeder reactor it can join the party, otherwise you round out the rest with storage as the price plummets. Burner reactors are irrelevant.
I feel like variability we can engineer around. What I’m more concerned about as someone generally pro wind/solar is how we’re going to discover and build supply chains for all the extra materials we need for this transition. We need a renaissance in resource exploration, discovery and extraction.
They actually rule that out in their model. It should make things slightly cheaper and easier if the EU nations all do this and trade energy, but it's not required.
They exclude a few other notable things:
> Our limit scenario makes a number of severely constraining assumptions
for the purpose of emphasizing what is possible for 100% SWB systems.
The bar for clean energy will not be nearly so high in most locations.
Assumption 2: no conventional operating reserve
Assumption 3: no other renewables
Assumption 4: no distributed generation or storage
Assumption 5: no impacts from electric vehicle energy storage
Assumption 6: no demand response, load shifting, energy arbitrage,
or peak shaving
Assumption 7: no technology breakthroughs
Assumption 8:
no subsidies, carbon taxes, or other financial innovations
These are all good things, they're not predicting or recommending against them, they're just saying they've assumed they don't exist when running the numbers to prove it would work everywhere.
I used to think about this as being a problem, but more and more I'm wondering if this is the wrong way to think about renewable capacity? For comparison, for years ISPs have been saying "what would you do with 1 gbps internet speeds?" And they've been putting of making the change to that because it's difficult. I think sometimes the problems that we expect to happen are a lot harder to be certain of when we're entirely speculating.
I bet some of this would sort itself out if they had sufficient renewables, and that the rest would be an easier incremental problem if they made the switch.
Well-known? French nuclear reactors are doing load following every day.
What nuclear energy is bad at is as a backup for renewable energy when there is no wind. Because nuclear costs the same whether you use it or not, so you pay for an energy source you don't use most of the time so that you can use wind instead. If you use nuclear, scrap wind.
Nuclear should be heavily subsidized regardless. Whenever they aren't needed for immediate demand, these power plants can be put to use for so many other tasks: desalination of water, energy storage in terms of hydroelectric pumping, direct carbon capture. Chemical fuel creation (hydrogen, ammonia, methanol, methane, etc).
If we want to unscrew ourselves of the ticking time bomb we set, we need all the "clean" energy we can get, and then some. Nuclear is potentially an existential requirement to the mix.
Synthesize and bottle carbon-neutral fuels like hydrogen or methanol with surplus off-peak nuclear power. Use fuels for peak on-demand electric, transport, etc.; methanol burns in standard internal combustion engines. Solves both the variable output and battery problem. Cost isn't too pleasant, but it scales and all the tech for this cycle exists now.
Pumped hydro is the obvious answer here, and today accounts for by far the majority of stored renewable electricity. There are geographies in Germany (natural and artificial) that can easily accommodate a lot more pumped hydro.
Second option (and also a boring one) is distributed home storage using traditional led-acid or lithium-ion batteries. This is already rolled out and can be increased to scale pretty easily.
A more exciting answer is liquid metal (or molten salt) batteries that works as large scale grid storage by heating the batteries elements (Calcium and Antimony) to very high temperature that keeps them separated while charged slowly mixing into a new liquid alloy as it discharges. You can read more about the technology here https://ambri.com/technology/ However I think before 2035 this will at best be a distant third from the two (boring) storage options above.
(but my understanding is that there are many other battery technologies in development that work too, e.g. older chemistries and/or mechanical/gravity/heat batteries)
But with the other logic, oil is not an energy source either, it is rather an energy storage that just so happened to store energy created by the sun millions of years ago.
I think it is OK to be loose on the technicalities in this instance, we all know what was meant.
Why do you think that's a useful question? Policy proposals like this are at the level of encouraging people to make fast-tracks for investments and passing planning permission for the factories to build the batteries and the mines to get their feedstocks; governments aren't generally even in the business of directly building the power stations themselves.
> Only realistic path is nuclear, but it'll take longer than 2035.
Not enough fissile fuel[0] for everyone to do that at western usage levels. And if you're talking that long, you can reasonably build out a global power grid, switching people from mining coal to mining metals, and the cost of making a grid of that scale is about the same (at current metal prices) as we currently spend per year on fossil fuels, give or take a factor of two.
[0] or at least, accessible fissile fuel; if you want to filter the oceans you get all the lithium you could want (and more other goodies like phosphate) as well as the uranium.
We can see how useful it is by comparing the 1970s accident and death rate with todays. Doing that we'll find that we're spending over a billion dollars per life saved from radiation poisoning. Which is ludicrously inefficient.
"a paper [..] provides some empirical evidence that safety changes have contributed to the cost of building new nuclear reactors. But the study also makes clear that they're only one of a number of factors, accounting for only a third of the soaring costs. The study also finds that, contrary to what those in the industry seem to expect, focusing on standardized designs doesn't really help matters, as costs continued to grow as more of a given reactor design was built."
Or build more wind and solar. Already Des Moines gets 80% of its power from wind, and the state is building more all the time. CA is doing some solar steps, but it is taking a lot longer (on a much larger population, so I'm not sure how their progress compares). Most states are not doing much. Europe is also making some steps, but they could do more.
The key to all this is starting ASAP and building constantly. Iowa has been building wind turbines at a rate of about 2/day (a bit over 500 per year) for many years now.
> And they have problems with excessive heat in the summer, because they cannot be properly cooled when water temperatures rise.
Incorrect, the can still be cooled but rivers must be kept below 28 degrees [1]. Because apparently the fish in this one stretch of river are more important than global climate change. Perhaps the takeaway is that we should reconsider the acceptable impact on local environment given the impact on global environment. Similar deal with America lithium mines held up behind environmental review.
Low river flow means more of the water goes into the heat exchanger, and leads to higher temperatures. It's still ultimately due to of heating the river, it's not that there's insufficient water to actually cool the reactor, just that there's insufficient water to cool the reactor while keeping river temperatures below 28 degrees. Say there's usually 100 cubic meters per second of flow, and the heat exchanger requires 10 cubic meters per second. If flow reduces to 30 cubic meters per second that's going to raise the temperature of the river but the plant can still be cooled if the people in charge decided reducing emissions is more important than heating a stretch of river.
Whatever environmental issues caused by heating a river is tiny comparison to global warming. If say I'm handwaving environmental issues, yet you neglect to specify what I'm overlooking. If you have reasons to think that heating a river is more important than averting climate catastrophe, I'm all ears.
And lastly, plenty of nuclear power plants are cooled by ocean water, or by wastewater [1].
You realise there's more to GHG than just electricity generation, right?
I mean it's a big, low hanging fruit because renewables are so cheap now, but even after getting that done we have transport and agriculture and land use chemical feedstocks and old leaking mines and wells and on and on. Plenty to do.
That's why we're also electrifying transportation, should push for heat pumps in buildings, etc, but yeah, there's a lot to do. Not dealing with electricity doesn't help the rest, though.
Though if you take care of car based transport (replace with EVs and/or electric trains), and electric you have done the vast majority. The rest of just a long trail of small things that don't really add up to much together.
massive construction of NPPs, with the last big share of works taking place in the 90s. the large majority of them are still in use today. average NPP age in France is ~30 years.
this is what allows France to produce electricity at ~50-100 grams of CO2 eq. per kWh, a performance that is only rivaled by countries with also nuclear- and/or hydro-heavy grids. for comparison, Germany sits between ~200-600 gCO2eq./kWh (last year's average). Germany's grid is renewable-heavy, but also intermittent, relying on the use of GHG-intensive backups (coal, gas). [1]
The book 'Stumbling on Happiness' by Dan Gilbert is great on this topic.
People are very bad at predicting what will make them happy. Much better to see what people who are happy now actually did, rather than imagine what will make you happy.
Adding to that, I'm reading Antidote by Oliver Burkeman. One of the points that I really related to was that more you think about happiness, more likely you will find yourself unhappy. Essentially the pursuit/expectation of happiness paradoxically will make your brain think of all the things wrong in your life. I don't know how it started, I think that's been true for me. I never thought about happiness when I was young. There were things that were not perfect -- not enough money, no girlfriend, not having clarity on what to do next, failures, envy in general -- but I was still much happier relatively and I was by no means thinking about happiness back then. Just accepted it is what it is, and kept going on with life.
While in practice it won't change my life much, I like the elegance of using a modern standard with this level of performance an efficiency.