Eythavon (@Eythavon) profile picture

Game controller reviewer

Joined April 2015
Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)Photo by Eythavon (@Eythavon)
Padlat has been updated with results from our latest high-precision latency tester. The website is still under development, so please excuse any incomplete data. We’ll be accelerating data updates going forward. English: padlat.pages.dev/en-US Japanese: padlat.pages.dev/ja-JP
4
6
1
34
13,642
That said, someone has brought this up, and it’s something I was already planning to look into. So I’ll be looking for professional ADC evaluation kits for ENOB testing. I’ll do my best to tell players roughly how many effective bits they can expect from their selected sampling setting, and how close that comes to fully utilizing the resolution that mainstream stick modules can deliver. However, all of these figures will still be approximate, because actual operating conditions vary too much to give an exact value for everyone.
1
1
6
1,054
This is also why we switched to a different MCU for the production version of HS2. Our goal is to achieve a higher effective number of bits, and we’ve placed particular emphasis on reworking and optimizing the circuit design. We also offer an oversampling option, allowing players to trade speed for greater precision. In practice, all of these improvements are only just enough to fully capture the 13–14 bits of effective resolution offered by today’s mainstream thumbsticks. I believe every one of these changes is necessary.
1
16
1,192
Native 16-bit sampling does not guarantee that the complete controller achieves 16 bits of effective resolution, nor does it automatically translate into a “16-bit experience.” Native bit depth describes the ADC’s quantization scale, while actual effective performance must be evaluated in the context of noise, distortion, signal range, and test conditions. In addition, if the thumbstick’s mechanical travel limits mean that its actual output voltage covers only part of the ADC’s input range, fewer raw ADC codes are available across that usable travel. This is a separate issue from the reduction in effective performance caused by noise, and the two should not be confused. This leads to two conclusions: A design can accurately be described as using “native 12-bit ADC sampling” or “native 16-bit ADC sampling,” provided that the hardware and actual sampling configuration support that description. However, this does not mean the complete system achieves the same number of bits of effective resolution. When the thumbstick’s actual output uses only part of the ADC’s input range, the ADC needs sufficient resolution and noise margin to prevent the conversion stage from becoming a bottleneck. What matters is the quantization step size and noise performance over the range actually used. If a particular thumbstick is measured to provide around 13–14 bits of effective resolution over a defined mechanical travel range, within a specified measurement bandwidth, and under clearly stated evaluation criteria, choosing an ADC with sufficiently fine quantization and sufficiently low noise can help preserve that detail. This is where a high-performance 16-bit ADC can provide real value.
1
2
17
1,762
I believe there are still a lot of misconceptions in discussions about controllers and thumbsticks. Some people present images and so-called “evidence” to support their conclusions, but if they have not properly understood the test conditions, what the data means, or the underlying concepts, those images alone do not prove their conclusions are correct. Before discussing the technical details, you first need to understand what you have actually measured. When discussing thumbstick precision, we first need to distinguish between two concepts: an ADC’s native bit depth and its measured effective number of bits, or ENOB. For conventional thumbsticks with analog outputs, the ADC is a key factor in determining sampling resolution. If we compare native bit depth to the markings on a ruler, it determines how finely the input range can be divided in a single conversion. ENOB, on the other hand, reflects actual conversion performance under specific test conditions, accounting for noise and distortion. How finely a ruler is marked and how precisely you can actually measure with it are related, but they should not be confused. A nominally 12-bit ADC therefore does not necessarily deliver a full 12 effective bits. Likewise, native 16-bit sampling does not automatically mean you achieve a full 16 effective bits in practice. The amount of useful information you ultimately obtain also depends on sensor noise, the actual output voltage range, circuit design, and subsequent processing. For example, under the same test conditions and evaluation criteria, if a 12-bit ADC implementation achieves around 10 effective bits, while a 16-bit implementation achieves 12 or 13, or even 14 or 15, that is a meaningful improvement. The fact that the latter falls short of an ideal 16 bits does not negate the value of that improvement. “An ENOB below 16 bits” does not mean “the sampling is not natively 16-bit,” and it certainly does not mean “a 16-bit ADC has no value.” You cannot simply point to noise in actual measurements and use it to dismiss the finer markings on the ruler itself. Evaluating a design requires looking at both its quantization resolution and its actual performance. Oversampling combined with filtering can also improve effective resolution, but that improvement does not come without a cost. It requires more sampling resources and involves trade-offs between bandwidth, filtering time, and latency. How much improvement is possible, and what it costs to achieve it, depend on the specific implementation. So, when the sensor and circuit design can support it, there is absolutely value in using an ADC with higher native resolution and better noise performance. It can improve existing designs while providing headroom for more capable thumbsticks in the future. Exactly how many effective bits a design achieves, and how much improvement it delivers, should be established through measurements under the same conditions. You cannot apply one blanket conclusion to every product. As for what comes next, I can reveal just one thing for now: a completely new thumbstick is expected to launch in about two months. I believe it has the potential to reshape the thumbstick market. When it arrives, its performance will speak for itself and put today’s flawed conclusions to rest once and for all.
4
7
1
38
5,521
Heliumstrike2 Launches Open Beta for Pro Players 📢 The next generation is finally here. 🟣 0.07 ms end-to-end latency—speed beyond your wildest imagination. 🟣 16-bit resolution chip—taking sampling to a whole new level. 🟣 6400 Hz gyroscope—master motion aiming with complete freedom. Fill out the form at the link below to sign up for an early access trial of Heliumstrike2: forms.gle/N9DjLKE2oTVgjNop8 All professional players: Welcome to a new era. #LeadJoy #Heliumstrike
6
9
4
66
8,958
Eythavon retweeted
Heliumstrike2、プロ選手オープンテストを開始📢 ついに、次の世代が来ました。 🟣0.07msのエンドツーエンド遅延、速さは想像を超えられる。 🟣16ビット解像チップ、サンプリングを新たな次元に。 🟣6400Hzジャイロスコープ、自由自在にモーションエイムを使いこなせる。 以下のリンクでフォームを入力して、Heliumstrike2の先行体験を申し込みできます: forms.gle/N9DjLKE2oTVgjNop8 プロ選手の皆様、新時代へようこそ。 #LeadJoy #Heliumstrike
5
28
11
154
20,861
I built a custom web-based driver for the Razer Wolverine V4 Pro to unlock even more of its potential.
9
3
3
41
4,971
Making a truly great product is incredibly difficult. We've switched MCUs five times and gone through more than a dozen hardware revisions, constantly refining and experimenting. I just hope all the time and effort we've put into this will be worth it in the end.
3
2
32
3,116
No fake or non-functional features. Only features genuinely supported by the firmware. Unofficial, independently developed third-party software. Compatible with GameSir G7 Pro 8K and Tarantula 8K. gs.evua.cc/
8
9
1
66
8,053
A few tips for the new Apex update The aim assist nerf in July was the most severe one and had a pretty noticeable impact on how the game felt for controller players. Some of those changes were rolled back in August, which improved the overall stick feel. Also, the higher your frame rate, the less noticeable the impact seems to be. So if your system can handle higher FPS, I wouldn’t recommend capping your frame rate anymore. Lastly, 8K polling-rate controllers do have a slight advantage over 4K, 1K, and other lower polling-rate controllers.
6
5
1
56
13,714
Scoring Criteria: A controller must excel significantly in the area it specializes in.
28
17
4
331
106,679
HeliumStrike 2 took a lot longer to release than I originally expected. But I’m pretty happy with how it turned out. I basically pushed everything as far as I could with what I have right now — even the cost ended up going a little over budget, haha. It took longer than planned, but I think the final result was worth it.😉
4
6
1
66
15,389
XBOX配列、開発中
5
28
10
204
32,174
The latency testing for the GameSir G7 Pro 8K has been completed. Its latency is very close to GameSir’s own Tarantula 8K, although it is slightly higher. Even so, it remains one of the strongest options among competitive controllers. The G7 Pro 8K also offers wireless functionality, and its wireless performance is quite impressive as well.
6
9
2
114
24,316
The latency testing for the Flydigi Apex 6 and Apex 6 Pro has been completed. Overall, both controllers deliver excellent latency performance at a 1000Hz polling rate, and their wireless performance is also very impressive. Overall, they rank among the better-performing 1000Hz controllers currently available.
🤖 Made with AI
1
7
2,609
From now on, please stop buying the black version of Genesis. A new white version is coming soon, featuring the new angle-sensing TMR JS13Pro, which fixes its biggest weakness.
11
4
77
19,502
Considering that similar situations exist with other devices—for example, Logitech mice have macro functionality, yet Logitech mice themselves are not banned from competitive tournaments. Instead, what matters is whether the macro function is actually enabled or being used. I think this sets a very good precedent. A feature itself is neither inherently good nor bad; what matters is how you use it. Of course, I completely understand how complicated and time-consuming it can be for tournament organizers to inspect every device. So requiring controllers to use a dedicated tournament firmware in which these functions are completely unavailable before they can be used in competition also seems like a perfectly reasonable approach to me. I can also share some of my own experience leading product development. Originally, I wanted to build a controller purely as a competitive FPS peripheral. However, many players who play FPS games also play other types of games, and they do not want to buy another controller just for those games—especially when a single controller is already relatively expensive. Because of that, I was ultimately pushed to include features I originally did not want to add, such as vibration and gyroscope support. At the same time, I also knew that these features are unnecessary for FPS games, especially in competitive tournaments. That is why I decided to provide two separate firmware options: an FPS-focused firmware and a general-purpose gaming firmware.
3
6
23
3,309
Honestly, there’s a huge difference between making the product I personally want to make and making the product the broader market actually wants. I’ve had to make compromises because of that.
2
1,062