Janet Jackson had the power to crash laptop computers (2022)(devblogs.microsoft.com) |
Janet Jackson had the power to crash laptop computers (2022)(devblogs.microsoft.com) |
Some dude hears somebody tell a story about sth 20 years ago, puts it in a blog, and here we are on HN, nobody questioning whether it's actually accurate. Of course Raymond Chen isn't just any random person, but the more important it would be to actually check? I mean, who hasn't heard people tell stories from decades ago, including colleagues reminiscing about the good old times "before y'all were born" only to realize later that it was vastly exaggerated or even outright made up.
Anybody around here with some actual first-hand info or at least another source besides this blog entry? I'd love to hear!
Thank dog for SSDs
In fact, if a magnetic HDD crashes, you may still recover some or all of the data by doing something hardcore, such as letting it sit for some hours in the freezer of your refrigerator, or immersing it in a bowl of rice overnight.
However, SSDs (and other flash storage devices) need to be switched on once in few months, otherwise there's a chance that some data stored in them may be permanently lost, as some cells may loose their power.
"As a reminder, an SSD's endurance rating is calculated based on how long it can store data if left unplugged after a certain amount of data has been written": https://www.tomshardware.com/pc-components/storage/unpowered...
The experiment Tom's is reporting on found twelve instances of data corruption on a low-end drive that had been subjected to over two thousand full drive writes, four times its rated write endurance, then left on a shelf for two years. This is a demonstration of a bottom of the barrel SSD wildly exceeding expectations.
It's really important in conversations like this to accurately convey not just the existence of the failure mode, but also the realistic chances of running into this problem, and the extent of the problem when it does manifest. If a deliberate torture test can only produce a few kilobytes of data corruption after twice the duration and four times the abuse the drive is supposed to be able to handle, this problem should be described as extremely minor.
CD drives however, can store data indefinitely without needing refreshing.
But if you're stuck with hardware that old, an SSD isn't an option.
Modern ones use more exotic materials.
Why the weasel words? Does Raymond Chen not know which models? Or is it actually apocryphal.
From the follow-up post: https://devblogs.microsoft.com/oldnewthing/20220920-00/?p=10...
Neither us nor the OEM ever figured out why. They suspected that it was a weird combination of different bin combinations from different parts, but ultimately we had to change the method of delivering video to stop it happening.
Also seems not unreasonable for an employee like him not to specifically name and shame hardware partners. Maybe it'd all be fine, but I wouldn't blame him at all for not wanting to risk it.
I heard enough of stories of a bottom of the barrel SSDs wildly exceeding expectations by actually crashing with a partial or a full data loss waaay below their expected write endurance and while still powered on. Sure, these are the real bottom of the barrel, like Netac or KingSpec - but I won't expect any non-server grade SSD to retain data at all for any meaningful time.
Many of us have old cheap flash drives, which may have some backups (family photos, videos, career files, etc.) we may not want to lose - so they may qualify for such periodic basic maintenance (just plugging them into the PC once in 6 months or so).
I think most home users don't know this can be a potential problem for flash drive storage.
You mean “vocabulary”, “terminology”, possibly “nomenclature”.
In my experience, flash drives tend to get problems suddenly leading to data corruption, and it may not be immediately apparent to the user till it is too late. I haven't such problems in recent years though, so maybe flash drives have become smarter too.
> When the German engineer Karl Benz invented the first petroleum-powered automobile, he did not just create an engine with wheels; he set in motion an industry that revolutionized the way society was structured.
Example (for both functions):
/* Emits a 7-Hz tone for 10 seconds.
True story: 7 Hz is the resonant
frequency of a chicken's skull cavity.
This was determined empirically in
Australia, where a new factory
generating 7-Hz tones was located too
close to a chicken ranch: When the
factory started up, all the chickens
died.
Your PC may not be able to emit a 7-Hz tone. */
#include
int main(void)
{
sound(7);
delay(10000);
nosound();
return 0;
}
And most speakers can play infrasound for many non-sinusoidal waveforms [0]. They'll drop the fundamental and some lower-end harmonics but can still give a sense of what it sounds like
For resonance the external driving force must match the resonance frequency of the system, but wind is rarely/never purely sinusoidal.
> Follow-up 2: Yes, I know that the Tacoma Narrows Bridge collapse was not the result of resonance, but I felt I had to drop the reference to forestall the “You forgot to mention the Tacoma Narrows Bridge!” comments.
https://en.wikipedia.org/wiki/Fourth,_fifth,_and_sixth_deriv...
That was such a great machine. We rearchitected our systems around it.
For me, that is less meaningful and more random than a youtube link. I have absolutely no clue what you are talking about.
a known problem in cutting vinyl records are sudden bursts of high volume frequencies around 100 hz, that have the potential to make the needle skip with a normal amount of weight on the tone-arm.
-------
Without the ECU you can easily break it by starting too slow
TFA was lacking details so this is merely a retelling.
Janet Jackson had the power to crash laptop computers (2022) - https://news.ycombinator.com/item?id=41534483 - Sept 2024 (79 comments)
Janet Jackson had the power to crash laptop computers - https://news.ycombinator.com/item?id=32483211 - Aug 2022 (12 comments)
Google's response after looking at the crash dumps: "WAI, your battery is degraded" (IIRC my phone was less than 3 years old).
As for fixes in software, it's either treating it as WAI, or secretly throttling down the phone, like Apple did, for which they got accused of planned obsolescence. Neither choice is good (though actually informing the users would go a long way).
That’s about the time I would have given up on the investigation and called in an exorcist.
> The main event was a brand-new mixing console called the Harrison Series 10, which was the first analog console to feature a digital control surface, with full automation of all parameters. Jimmy Jam and Terry Lewis were the first studio to have it, according to Jam. This meant that they could cut down the time it took to switch songs to about 10 minutes because complex mixes now required little-to-no cross-patching.
https://reverb.com/news/the-making-of-janet-jacksons-rhythm-...
Reality must have been falling apart for someone for a brief moment there.
The issue of music crashing a hard disk drive was a genuine problem. Since I later specialized in hard drives, I can confirm that every manufacturer faced this issue as multimedia laptops became more common and we transitioned to higher areal densities. To state the obvious, we shrank the tracks every time to achieve larger capacities. We were doubling capacities every nine to twelve months when we first introduced MR, GMR, and PMR heads. The hard drive industry employs incredible control theory experts due to the requirement of keeping a head on track. Personal opinion, which I probably should research, but I believe that one of the densest concentrations of Ph.D.s in leading edge control theory could be found within the hard disk drive industry. Amazing things happen when you're trying to fly nanometers off the disc in a track that is maybe 100 nanometers wide at the time.
By 2010ish, as part of our development suite, we actually played music through the speakers to identify these types of issues. The origin of this practice actually came from the ODMs in Taiwan. Therefore, Janet Jackson was not the standard qualification song we used. Instead, it was popular hits from the Chinese pop market. There were also Western songs within the suite, but I remember our team blasting Chinese pop songs at full volume on multimedia laptops.
Laptop development began moving heavily to Taiwan in the mid nineties. By the time the early two thousands arrived, there was a massive amount of competence in Taiwan regarding chassis design engineering. As time progressed, every American PC company continued to outsource development to Taiwan and eventually to China. As development was outsourced, the ODMs would work with suppliers because we wanted to present solutions to the OEMs that were free of issues.
EDIT: By the way, it wasn't necessarily a blue-screen crash. I'm not saying that that couldn't happen, but generally what we did is we had the throughput test that we were run on the hard drives. Then we would go blast the music full bore and there were certain bandwidth requirements that we needed to get out of the hard disk drive. So to add some more details behind this, I would describe it as a performance issue and the blue screen issue was relatively minor. However, this was a number of years ago and I don't remember the exact percentages.
But also the original post was about a 7hz tone somehow resonating with a chicken's skull cavity, which if you know the basic wave equation relating wavelength with frequency is an absurdity. The waves involved are multiple orders of magnitude too big to couple to a volume that small. They'll just diffract around like nothing.
I originally posted my own comment with link and prose, but then saw GP's comment which preceded mine by an hour, so I deleted mine. I didn't originally realize it was that same video because it didn't say so, but then I recognized the link hash. I want people to see that video, because it's so great, so I added context as a reply.
What is definitely well documented is Brendan Gregg’s related discovery of performance degradation in servers from vibration of sibling servers / clapping nearby that caused spinning disks to pause their heads.
You're misunderstanding the numbers here. Going from 12 to 7 Hz is most of an octave, nearly doubling wavelengths.
Also SVS's numbers are gonna be the usual marketing stuff, so they're assuming a fat room gain curve, and just looking at their website they have a disclaimer on their graphs that it doesn't represent actual total output capability. Which is a way of hiding that if you actually try to drive it that hard that low with ~3kw electrical in those voice coils are going to torch.
The non lying way to prove that claim is to show large signal Kipple results including the heat soak. They ain't doin' that here.
Basically stuff going this low is really exotic and more in the realm of servos that simulate earthquakes than traditional transducers.
Tom Danley is the world expert on this sort of thing. He used to build stuff like ultrasonic levitation ovens and full scale sonic boom simulators for JPL/NASA.
In the audio world he was first famous as the tech lead behind ServoDrive. This now defunct company made special effects subwoofers using DC rotary servo motors to drive the diaphragm. They were used as special effects subs in that era by big acts like Garth Brooks. But they didn't catch on outside that niche because very little music has significant content below 40 hz as it just turns into a muddy rumble that harms sound quality as a whole. So to use these sorts of things you have to mix for it specifically. Cinema goes lower with the rumbles down to 15hz, but that's basically it.
Getting anything that's like a clean tone at 7hz is not gonna happen without a purpose built device.
FWIW Tom Danley started his own company[1] after Servo Drive failed on the business side, where he focuses on large scale horn speakers using novel topologies. They're among the best in the business at what they. Again, they don't have anything that even remotely tries to go down to 7hz.
[1]: https://www.danleysoundlabs.com/
Tom's a nice guy, I've traded emails with him a few times over the years. He used to be pretty active on the DIY speaker building mailing lists sharing his very in depth knowledge freely.
Without disputing the conclusion, is the wavelength the right measurement, or should that be half the wavelength?
In their younger days, two distinguished engineers, Bryan Cantrill and Brendan Gregg, made this video where they scream at a data storage server nicknamed Thumper. Screaming at it has surprising results, which are observed with a novel software technology called dtrace.
The Sun Fire X4500 was a dense storage sever, 4U with 48 disks and insane IO performance and a newish filesystem called ZFS. The video is not only funny in content, it features technology and technologists that became very impactful, hence the classic tag.
---
I love the lore, so I'll drop more.
While our team previously used AFS (mainly for its great caching) and many storage servers, this hardware combined with its software allowed us to consolidate and manage and access data in new ways, alleviating many of our market data analysis problems.
We switched to NFS, which previously was not performant enough for us on other hw/sw architectures. While using NFS with the Thumpers and then Thors (X4540) was fantastic, eventually the data scales became hard again and we made a distributed immutable filesystem that looked like the Hadoop HDFS and Cassandra file systems, named after our favorite Klingon Worf (Write-Once Read-Frequently).
Interestingly, in 2025 both XTX [1] and HRT [2] open-sourced their distributed file systems which are pretty similar to it, using 2020's tech rather than 2000's. HRT's is based on Meta's Tectonic which is a spiritual successor to Cassandra.
I wrote about our parallel HFT networking journey once upon a time on HN. [3]
[1] https://www.xtxmarkets.com/tech/2025-ternfs/
[2] https://www.hudsonrivertrading.com/hrtbeat/distributed-files...
Edit: I clearly just got derailed by my initial confusion, there was nothing weird about your usage.
If someone does something that has an impact on IOPS, then it's sensible to say they (or what they're doing) are "affecting IOPS."
Screaming at the JBOD wasn't "causing" IOPS, so using that phrase would've been more confusing.