[214324]

Parallax can be used to one's advantage, as it can be an offset to help you see the tip better. What people do is use custom nibs that they make themselves, and they make it longer or shorter than the standard nibs that came with Wacom pens. I know there's that funny article about using spaghetti for nibs but honestly that's a terrible idea. The grinded wheat will go everywhere. People have been using nylon trim wires or toothpicks for ages, especially as Wacom nib shapes are easy to mimick (just a cylinder that tapers to a round end) and they were very expensive. When the Intuos4 came out, Wacom added the new experimental paper texture to the surface. It was more like sand paper. My old Bamboo tablet was from the same time period. Many people complained that half the nib was gone after a painting session. For me, 3 days of very light use shaved my nib's round tip into a flat plane. Wacom quickly fixed that and offered to change the surface of people who bought the Intuos line. Can't remember if it was free or charged, but it was a very big deal and many people were not happy. Anyways, some carefully polished their tablet surface until it was smooth. Others used cheap replacements for nibs. I personally wrapped a paper on top of my tablet and sanded a toothpick to the shape of the nib. That one single nib lasted many many years! Only ditched it a few months ago. People discovered that the length of the custom nib affected the feel of the tool. Some could make it quite long and drag it like a brush. One person I know had a large Cintiq with a smooth glass surface. He cut his longer than usual (I believe. This was many years ago so memory is fading) with a nylon trim wire after I told him about it, and he rounded the tip by melting it slightly on top of a lighter. He was pretty happy, as the nylon tip didn't give him the annoying plastic-clicking-on-glass feel, and the extra distance increased the parallax effect, so he can see the brush tip in Photoshop without the stylus blocking anything, quite important for line art that requires a very small brush.

Apple's digitizer is honestly amazing. Still can use some polish here and there, but with the power in those tablets and the almost full migration of some very powerfull desktop art applications, they are seriously looking to become mobile workstations for those who just need one or two specific apps.

I think there's a bit of a difference here. Parallax is bad, period. If the system by default can't match pen tip to surface precisely it's just bad in a user experience setting. Nobody wants to draw somewhere on the screen only to have it appear elsewhere without notice. Edge accuracy is probably one of the worst user experience problems with EMR since it's basically variable, uncontrollable parallax.

That being said, offset is useful. That's why I included the small bit on people who do actually prefer separate displays and pen tablets (e.g. Intuos/Bamboo, Intuos Pro) because they don't need to see their hand or pen in the way while drawing. I don't do art (wish I could or had the time) so I'm not sure if there's software that introduces pen-to-cursor offset voluntarily but as far as I'm aware there isn't. Interchangeable surfaces are also nice for separate pen surfaces/tablets because they usually don't affect the performance of the device so long as the material is of the right thickness and doesn't interfere with electromagnetic radiation. For pen displays your only solution is stuff like matte screen protectors which degrade image sharpness.

Interesting. I haven't been following coffee lake news until yesterday night so you can say I'm clueless. I had a feeling this was a response to AMD but still, only started reading up what's up yesterday, and it's mainly to see what's inside the newest 2-in-1's. Thanks for the write-up. Appreciate it!

I mainly posted that section out of personal excitement regarding how things may go in a few years. The main program I use is Photoshop CS6. Others are some 3D program that are quite light if used in optimized workflows. In Photoshop, the only real tool I need is the brush tool and all those that surround it. They are nearly all CPU based, but really only use up to 4 cores. And they really really like high clock speeds.

Say a filter takes 12 seconds to complete. If I shave off 25%, it will be 9 seconds. I will notice it, but I'm already in a mode where I'm waiting, stretching, or sipping my coffee. I don't care about those 3 seconds. Now lets say my processor is struggling with keeping up with the brush, so there is a small lag. This is a common problem with weaker mobile devices that are working with complex brushes. If I shave off 25%, it will feel very responsive, as strokes are done quickly and consistently. Vise versa, if I increase the lag by 33%, it will feel terrible. It's one of those few workflows where when you need it, 2 cores turboing up to 4 ghz temporarily is extremely appreciated once in a while. 4 cores at lower speeds are acceptable for other stuff when Photoshop isn't chewing up a single core (and all the RAM).

Regarding the 6 core part, I'm not trying to argue. You are completely right. It's just looking up the current CPU lines made me think that Intel is actually doing something (albeit they are probably just reacting to AMD), and the CPU market will finally get interesting. The current gen stuff is set, but in 5 years, mobile workstations might be very interesting for digital painting, even if it is a fairly niched market. For what I do, some 3D renderers take advantage of GPU rendering very well (saw a watercooled 11 Titan XP setup as a dedicated Octane Render setup a while back), but most rely still on the CPU. Seeing higher core counts being brought to the mainstream, especially on portable devices really makes me wonder about how things will be like in a few years. A light device, that can turbo very high on 2 cores when needed but also packaged with a good 6 core design can be pretty nice for Photoshop and a 3D program on the side. Also really excited for AMD's stuff, especially with power efficiency as you mentioned. I guess I kinda got ahead of myself with daydreaming.

Mmm I wouldn't exactly describe it as the CPU/GPU market getting more interesting, rather more of it's getting to the point where it should have been. Intel's been just sitting on their butts doing nothing as usual. I expected them to have something proper planned in response to Ryzen but I guess they stopped making good Plan B's after the several years of Bulldozer so that's not surprising.

I'd pin it more at 3 years for GPGPU to take off. AMD, as far as I'm concerned, is still very serious about HSA and from what I gather, Intel is getting serious about it too after poaching Radeon Technology Group's head of design, Raja Koduri, for their own graphics department. GPUs have basically been sitting at the front of AI and self-driving cars as well. One of the big themes of electronic chips is integration. The most obvious one is SoC's (system on a chip) which have a CPU, GPU, modem, etc. all integrated into a single chip. x86 CPUs are finally starting to move into this direction, with AMD creating the infinity fabric to help them tie different chips together on a logical level and Intel's EMIB to help create smaller, thinner packages, connecting dies on a physical level. With AMD trying to forcefully gain marketshare in their partnership with Intel, which I can only expect to continue for at least the next 3-4 years before Intel can come up with competing integrated graphics products, we'll see more optimizations for GCN-based architectures (NVIDIA's marketshare needs to fall. They play very dirty and somehow get away with it, but I can't take away credit from their recent architectures which are quite amazing honestly) and alongside it, hopefully, more support for GPGPU stuff like OpenCL and even more HSA pressure.

The more exciting products I would expect to come next year with Zen 2. Zen right now is a bit held back by 1. AMD's relative inexperience with SMT and the newly developed Infinity Fabric and 2. GlobalFoundry's 14nm process which currently limits the maximum clockspeed on Zen chips to basically 4GHz for the most part. GlobalFoundries has been on a roll recently and their 7nm process sounds like it's going to be a winner. Intel has had enormous issues with their 10nm process, now even on 10nm++ which is still having problems. The original 10nm and I believe 10nm+ process as well were actually performing worse than the 14nm processes so Intel is hard stuck on the lithography side.

Anyways with Zen 2 we should probably see the single core IPC gap between Intel and AMD to shrink to something that's quite negligible which will basically make Zen the better buy, period, unless Intel can come up with something good. Hopefully at that point we can see high performance APUs much like Kaby Lake G.

Hopefully we also get to see something more promising regarding Navi. I'm not exactly sure how Vega is actually. From what I understand, at the correct operating point, Vega is actually very, very power efficient despite how Vega 56 and Vega 64 look. Vega GL and GH for the Kaby Lake G processors look fantastic, giving us 65W chips and 100W chips with performance near 1050-1050Ti and 1050Ti-1060 levels with a CPU to boot as well. We'll see. Anyways, this year will not be that exciting with Zen+ (Zen on GlobalFoundries 12nm+ process made specifically for Zen to allow for better clockspeeds and power efficiency). Next year with Zen 2 will be the truly exciting year. If AMD can hold up on it's 7-8% performance increase per Zen generation, I will be very happy, especially since they also have to deal with Spectre fixes (Meltdown is one of 3/4 Spectre attacks that's only possible on Intel systems. AMD, Intel, and ARM architectures are vulnerable to varying degrees to the other kinds of Spectre attacks) which can hurt IPC slightly.

What I do want to see are updated VAIO Z Canvas and VAIO Z Flip machines. I do wish VAIO (no longer under Sony for those who don't follow) would consider using Wacom EMR much like Samsung does for their pen products. I'm personally hoping for a VAIO Z Flip with a 65W Kaby Lake G chip.

 

[U-3C]

It's been almost half a decade bruh. I'd seriously question Intel if a chip one tier below another one could not match and beat the others after close to five years.

Well, it's more like just under 4 years. Just find it funny that there has been so little progress, with Skylake and Kaby Lake barely offering anything and all of a sudden Intel goes in panic mode. Like, I'm still taking to people with overclocked Sandy Bridge cpus who just stare at Kaby Lake and shake their heads, "Nope, not with upgrading. Will wait another generation."

A Chinese artist did a very detailed video serious benchmarking the impact different components made for digital painting in Photoshop. He went from a Sandy Bridge i5 2500 (not sure if it was overclockable or not) to a Skylake i7 6700k and said he literally could not feel a difference. Even going from the igpu to a 1060 made a consistent difference in benchmarks, but a 4 generation leap in cpu tech, despite Photoshop loving single core cpu speeds, doesn't do much.

I think there's a bit of a difference here. Parallax is bad, period. If the system by default can't match pen tip to surface precisely it's just bad in a user experience setting. Nobody wants to draw somewhere on the screen only to have it appear elsewhere without notice. Edge accuracy is probably one of the worst user experience problems with EMR since it's basically variable, uncontrollable parallax.

I see. Parallax indeed was always bothersome. It's just when people had to deal with it in the past and had no solution, everyone eventually adapted to it. Earlier Cintiqs were horrible. Terrible screen, overheating issues and parallax, are just some of the problems. Using the stylus for small movements like clicking things were always difficult unless you have a huge screen. People survived, but to this day I still never want a Cintiq. Just don't like having my hand blocking a section what I'm painting on.

If I remember correctly, the first gen surface pro had serious parallax issues. So did the Cintiq Companion. Having a small screen really didn't help with that matter Ave nobody could really adjust to those very well. Aside from inconvenient button and port layouts, battery life was a big issue too. Haswell cpus helped a bit for the Cintiqs but they were still not that great. It was the Cintiq Companion and the Surface Pro lines at war with each other for some time, and the art community was just in the side watching. Neither of the first gen products were perfect, but they were both respected and became better with time. Everyone was looking for more mobile setups and the were nothing really like these two products at the time.

I do want to say that the Cintiq's stand is still garbage to this day, and the Surface Pro's stand is still awesome, to this day... -_-

Not sure about software offset. I know certain mobile painting apps have an offset feature. It's required as finger painting on an Android phone with no digitizer isn't fun. With the Cintiqs, some people like my friend made custom nibs to create that offset. I never tried it but this person was very happy with it.


Matte screen protectors are decent for casual use, but people who do any graphical work probably need a good display. Aside from image sharpness, screen protectors can also affect the colours accuracy quite a bit; a big no-no if that's important.

Talked to someone who asked his company to get him a Cintiq. He tried it for a few hours and hated it, so now it's in the hands of done IT guy I think. XD The financial department might have poundered how the company just wasted 3k on a fancy display.

Honestly though, simple graphics tablets have come so far. For less that 100 dollars, you can grab an amazingly designed tablet that even the wealthiest pros could only dream of a decade back. Accurate, amazing texture, portable, comfortable, and durable with lots of support. Wacom has barely done anything with their monopoly over the years but finally Chinese companies have caught up, both in hardware and software.

Once, cheap tablets from Monoprice can get the job done, if you can somehow get over the driver issues. Then came a time that even if the Chinese tablets are very close, Wacom still has the best of you want that extra premium. Eventually, unless you really need ZBrush Core (150 dollar value), which they bundle with their 200 dollar Intuos 3D, I can't recommend Wacom at all, especially the smaller sized Intuos non-pro line. Now, just buy ZBrush Core and a cheap Chinese tablet from a respectable brand in you want to go down the graphics tablet route. The only thing from Wacom that really stands out now is their Cintiq line.

Also, nobody will recommend ZBrush Core over the full ZBrush anyways... :/

Mmm I wouldn't exactly describe it as the CPU/GPU market getting more interesting, rather more of it's getting to the point where it should have been. Intel's been just sitting on their butts doing nothing as usual. I expected them to have something proper planned in response to Ryzen but I guess they stopped making good Plan B's after the several years of Bulldozer so that's not surprising.

Yeah.

That's kinda how I feel too. Didn't use the right words.

As I mentioned it above, the cpu market has just been so bland over the past years, to suddenly see Intel do something so different is amusing. Hope it's a sign that AMD really stepped in Intel's tail. New competition is always fun to look at. Or at least less boring than no competition.

Humor us Intel!

 

[bowei006]

Well, it's more like just under 4 years. Just find it funny that there has been so little progress, with Skylake and Kaby Lake barely offering anything and all of a sudden Intel goes in panic mode. Like, I'm still taking to people with overclocked Sandy Bridge cpus who just stare at Kaby Lake and shake their heads, "Nope, not with upgrading. Will wait another generation."

A Chinese artist did a very detailed video serious benchmarking the impact different components made for digital painting in Photoshop. He went from a Sandy Bridge i5 2500 (not sure if it was overclockable or not) to a Skylake i7 6700k and said he literally could not feel a difference. Even going from the igpu to a 1060 made a consistent difference in benchmarks, but a 4 generation leap in cpu tech, despite Photoshop loving single core cpu speeds, doesn't do much.

!

Bruh, you can't start out saying its a very detailed and serious benchmark of a CPU. Only to then tell me the benchmark was on painting in Photoshop (light to medium workload for most users) and then offer that he subjectively could not 'feel' a difference. I have no doubt that he probably could not feel a difference depending on what he was drawing or what not. There absolutely can be resource intensive use of a CPU by drawing programs but they aren't on the top of the list for benchmarks. Cinebench is another topic.

 

[Nick 214]

Not to step into the middle of this discussion but...

I'm thinking of getting a new plate... Who knows what this is from (easy)?

I kinda like it.

 

[miceblue]

@RERO are you planning to visit CanJam NYC this year?

 

[U-3C]

Bruh, you can't start out saying its a very detailed and serious benchmark of a CPU. Only to then tell me the benchmark was on painting in Photoshop (light to medium workload for most users) and then offer that he subjectively could not 'feel' a difference. I have no doubt that he probably could not feel a difference depending on what he was drawing or what not. There absolutely can be resource intensive use of a CPU by drawing programs but they aren't on the top of the list for benchmarks. Cinebench is another topic.

Sorry.

Couldn't use proper words recently. It should be a "relatively" serious test compared to the zero other tests I could find on the internet. Probably are some in other languages, but I can't find any in Chinese or English.

Shouldn't have used the word benchmark. There are no real benchmarks for this stuff because it can't be benchmarked. This guy was just demonstrating the effects with one of his own pieces to give a basic idea, and it's quite good (relative to zero other videos in the past 7 years) as it focuses on features geared towards digital painting.

The main testing methodology was using Photoshop's built in timing tool to record the time it takes to complete the same action. Not anywhere near ideal, but it demonstrated the basic idea and it's a great starting point for discussion. Doesn't really matter at the end. Just kinda interesting to see certain things, like how the smudge tool at the time takes advantage of gpu acceleration and how the basic brush engine really can't take advantage of hyperthreading. After the three videos, the person briefly summed up his subjective experience, which is where he mentioned in practice, he can't really feel much of a difference upgrading the cpu or gpu.

I honestly need more sleep before I post... :/

 

[RERO]

@RERO are you planning to visit CanJam NYC this year?

Hey man, you just woke me up from my (almost a year-long) Head-Fi coma! Everything... looks different...

Unfortunately, I won't be going this year as I'm on the other side of the world right now. Though I might make it on time for 「London」 this year...

But yeah, how you AAMML guys doing?

Ever since summer of last year, I took a break on the headphone/tech scene and, to a lesser extent, Chinese Cartoons due to work as I rarely get to sit down on my desktop without already being fatigued / sleepy. On weekends I usually just creepily take pictures of random people and try to pass it off as 「Street Photography」. When I get tired, I sit down on some random park and play this Fate kusoge (playing since day 2) and some idol game that no one really plays anymore. Sometimes even shamelessly watching CGDCT cartoons out in public at Wash. Sq. Park.

Last thing I finished watching was 「New Game!」 S1, just this Monday. I like it. It's cute and oddly relatable.
This season I'm only I'm watching Pop Team Epic and Last Encore, with the latter being disappointing so far. Seems like I just can't get into any of the "serious" Fate entrees.

Well, considering I'm on a weird half vacation / half part-time (?) job thingy on my hometown with 3 Mbps internet, might as well funpost on HF.

Spoiler: Hopefully my poor R70x doesn't get lost somewhere on the Pacific Ocean.

よろしくねぇー。

 

[Smugsie]

https://www.hear-it.org/Your-ears-differ-

I was about to RMA my headphones but I think I might ask my mom to RMA my ears. I hope the first paragraph of that website is true, I was going nuts for an hour trying to determine if my headphones really do have a channel imbalance.

Also, Kamidori Alchemy Meister is a good game. I started playing it a few weeks ago and I clocked 100 hours on the first route, not bad.

 

[bowei006]

Not to step into the middle of this discussion but...

I'm thinking of getting a new plate... Who knows what this is from (easy)?

I kinda like it.

Just the first three letters and I could tell it was some mecha oriented thing. Didn't realize it was Outlaw Star haha

Sorry.

Couldn't use proper words recently. It should be a "relatively" serious test compared to the zero other tests I could find on the internet. Probably are some in other languages, but I can't find any in Chinese or English.

Shouldn't have used the word benchmark. There are no real benchmarks for this stuff because it can't be benchmarked. This guy was just demonstrating the effects with one of his own pieces to give a basic idea, and it's quite good (relative to zero other videos in the past 7 years) as it focuses on features geared towards digital painting.

The main testing methodology was using Photoshop's built in timing tool to record the time it takes to complete the same action. Not anywhere near ideal, but it demonstrated the basic idea and it's a great starting point for discussion. Doesn't really matter at the end. Just kinda interesting to see certain things, like how the smudge tool at the time takes advantage of gpu acceleration and how the basic brush engine really can't take advantage of hyperthreading. After the three videos, the person briefly summed up his subjective experience, which is where he mentioned in practice, he can't really feel much of a difference upgrading the cpu or gpu.

I honestly need more sleep before I post... :/

Depending on what he was doing, it could have been very CPU un-intensive (relatively) allowing him to not feel a difference.

Hey man, you just woke me up from my (almost a year-long) Head-Fi coma! Everything... looks different...

Unfortunately, I won't be going this year as I'm on the other side of the world right now. Though I might make it on time for 「London」 this year...

But yeah, how you AAMML guys doing?

Ever since summer of last year, I took a break on the headphone/tech scene and, to a lesser extent, Chinese Cartoons due to work as I rarely get to sit down on my desktop without already being fatigued / sleepy. On weekends I usually just creepily take pictures of random people and try to pass it off as 「Street Photography」. When I get tired, I sit down on some random park and play this Fate kusoge (playing since day 2) and some idol game that no one really plays anymore. Sometimes even shamelessly watching CGDCT cartoons out in public at Wash. Sq. Park.

Last thing I finished watching was 「New Game!」 S1, just this Monday. I like it. It's cute and oddly relatable.
This season I'm only I'm watching Pop Team Epic and Last Encore, with the latter being disappointing so far. Seems like I just can't get into any of the "serious" Fate entrees.

Well, considering I'm on a weird half vacation / half part-time (?) job thingy on my hometown with 3 Mbps internet, might as well funpost on HF.

Spoiler: Hopefully my poor R70x doesn't get lost somewhere on the Pacific Ocean.

よろしくねぇー。

Pretty good man. Just watching seasonal animu and loving every moment of it

 

[214324]

Spoiler: Violet Evergarden 06

What. They give us last week's cliff hanger then give us something completely unrelated?

This legit felt like a filler episode. We get introduced to somebody who gets development but we'll probably never see again after this. Meanwhile the story is nowhere to be found.

 

[Muinarc]

Just did my taxes. 31 minutes total. I think I can do them faster, oh well next year I go for a new record!

 

[castleofargh]

https://www.hear-it.org/Your-ears-differ-

I was about to RMA my headphones but I think I might ask my mom to RMA my ears. I hope the first paragraph of that website is true, I was going nuts for an hour trying to determine if my headphones really do have a channel imbalance.

Also, Kamidori Alchemy Meister is a good game. I started playing it a few weeks ago and I clocked 100 hours on the first route, not bad.

very likely that both your ears and the headphone have some degree of imbalance somewhere.
pick a few tones(because you don't know where the imbalance might be) in mono, check if they feel centered. if not, turn the headphone around and check if the movement goes to the other side. diagnostic complete!

 

[U-3C]

Depending on what he was doing, it could have been very CPU un-intensive (relatively) allowing him to not feel a difference.

Yes. It's more of a common digital painting workflow thing. You eventually get into the habit of writing with your limitations and upgrading might not feel like an upgrade at all. In this case, it was literally a few words commenting on how the cpu change can't be tested, but subjectively, despite the generation jumped, he can't feel a difference. After the the videos testing all the other hardware components, you can have a general idea of how different components impact different things in Photoshop. Buying a new cpu, which requires a new motherboard, and as a result new ram, and depending on what windows you have, you might need a new windows licence. All this doesn't make sense compared to a simple RAM upgrade if that's what is choking you.

Quad core desktop CPUs from Intel haven't been improving much in raw power, but it's already past the point of diminishing returns for most digital painters. People could paint decades ago and the many workflows have been optimised to not use too much cpu power, especially with all the people who have limited hardware. Feeling smooth and consistent is a lot more important than great but choppy performance with high end components. If you are rendering things in 3D, better components may make a difference, but for painting with your hands, the biggest limiting factor is still a human one, assuming your hardware is past a certain point.


The biggest problems have always been with mobile devices. I think today, we have gotten to the point where for digital painting, moderately small high end laptops have reached that point of diminishing return for 2D and basic 3D with. But until a few years ago, it's hard to get very portable devices with quad core CPUs that can stay consistently above 3 GHz on all four cores, has more than 16 gb of ram, and is light enough to be held with one hand. That point of diminishing returns haven't been met yet so every generation was a decent leap.

Ivy Bridge to Haswell was a big deal for people interested in mobile painting setups due to power efficiency. Skylake also focused on efficiency, and brought a lot of powerful processors into small devices. Kaby Lake...meh. Nice to have over Skylake but honestly didn't do much. The interesting thing is that quad core processors are making it to 13 inch 2-in-1's, and they might be able to maintain above 2 GHz on all four cores. Aside from issues Hybrid mentioned the processing power is still not enough to keep up with the human hands sadly. This will make the experience feel choppy, as every click will have some lag, affecting your workwhen working large (say for professional prints). I think I explained lag issues for digital painting before. A short turbo can help immensely though, but if course, it's not a real solution. Still, seeing high core count processors potentially becoming mainstream really gave me a bit of hope for what I do specifically, which is a pretty niche thing that nobody really focuses on designing products for so any kind of progress can bring tears to my eyes.

There's always been some compromise, like weight, heat, performance, battery life, driver issues or just poor/immature design in general, and it never quite felt like there will be a full desktop replacement. Many people bought the early Surface Pro and Cintiq tablets but ended up not using them unfortunately. Android tablets with a digitiser can barely scrape by with light work. Now I finally feel that the general tends are finally, for the first time, leading towards something I personally can be excited by. Holding a 13 inch xps is very comfortable. More and more products are moving in that direction too so there will be some diversity. Digitisers on these devices are still crap but they are getting there, and higher core counts (as in 4) are finally coming to even the smallest of these windows computers. Maybe some device can finally come out within the next 5 to 10 years that can replace my sketchbook and a desktop solution at the same time when I'm working outside for my own specific workflow. That plus recent lack of sleep for two weeks is what got me posting my recent mood swings online.

 

[214324]

No they just won't maintain above base clock for the most part. Turbo is only for up to 100 seconds of load and even then, the amount of power draw over TDP is dictated by how good the cooling is. You will never run above 2GHz at best in the long term (or 1.6 GHz since that's the actual baseclock speed for most of the U-series quad cores right now). Higher core count doesn't make things smoother either when it comes to drawing since it's a single-core load task. Multicore loads require much more parallelized workloads; pen strokes are a single process workload.

Jitter isn't a result of processing power either. Whatever delay there is (the kind of stuff you can see in slowed down footage) really isn't an issue. Most people don't notice it; unless you just like to make really fast pen strokes (like mouse-flick high dpi level kind of flicks) it's not going to matter. Jitter is a result of digitizer/pen interactions and digitizer polling.

I wouldn't bother talking about things you don't have an inkling of understanding about (basically every enthusiast community with 99% of its members).

 

[deadlylover]

You will never run above 2GHz at best in the long term (or 1.6 GHz since that's the actual baseclock speed for most of the U-series quad cores right now).

Manufacturers are sort of getting better at incorporating better cooling and getting the most out of the new 15W quad core parts (for the more premium laptops anyway), the new XPS 13 for example runs at a ~26W TDP and can sustain a phenomenal 4 x 2.7GHz. With some undervolting I'd be surprised if it couldn't fit squeeze out another couple hundred megahertz. To put that in perspective, the old Skylake XPS 15 ran at 4 x 2.8GHz! (both comparing base level chips on a repeated Cinebench load from notebookcheck, i5-8250U vs i5-6300hq, though the XPS 15 was infamous for being real poopy back then with throttling)

Intel are really starting to push the cTDP angle with these new chips, they even list the cTDP specs in the ark where the previous generations used to be a bit more hush hush.

I wonder what they're going to do in the ~28W TDP segment now which is dominated by the 13" touchbar Macbooks, Apple tends to want a beefier iGPU so perhaps they're going to use that new Intel+AMD chippie instead. I have no idea if they've given up on the Iris Pro lineup.