Exposure of fake Chinese tube amplifier (2018)
stereonet.com
stereonet.com
I had a friend in sound production tell me that he had the experience of ramping up sliders to show an effect to a client and watching their faces brighten at the improvement. Then, he noticed something was misconfigured and the sliders were doing nothing.
That said, there are people who can hear the difference between 320kbps mp3s and uncompressed, because they can sense the faint temporal errors that can make a sound a touch before the sound they should follow.
This depends a lot on your setup. In perfect room with 5 digit sound equipment and encoded with a bad encoder? Sure.
Good encoder on something like phone speakers, especially with examples not specifically crafted to show weaknesses? Never.
Of course there is a lot of in-between, but for most people and setups, mp3 is not the limitation.
The half of this ability lies in the audio system being listened to. As we like the say, the performance of something is capped by the weakest link in the system.
If you have a good sound source feeding a semi-decent amplifier paired with a good set of speakers, you can feel the sound stage and depth difference between 320kbps MP3 and a CD.
However, modern mastering techniques generally maximize loudness and erode some detail, so everything becomes disputable.
I have some properly mastered CDs, and listening them blows my ears away every time, others don't do the same because they're compressed for the sake of loudness and they don't sound as spectacular.
You got a link to some independent blind listening test results showing someone has this ability? I’m extremely skeptical of the above claim.
Moreover, there's a file which I just made [0] which shows the sound difference between a FLAC and a MP3 file.
How I made this?
1. Rip the CD to FLAC.
2. Convert FLAC to MP3 with lame (320CBR, 44.1KHz, Compression level 5).
3. Load both to Audacity, subtract MP3 from FLAC.
4. Amplify remainder to 0.0dB.
5. Render to FLAC.
P.S.: To verify I tried this process with two copies of the same FLAC file, and what remains from subtraction is a clear 0. So, the process is working.
[0]: https://share.getcloudapp.com/5zurmeLz
[0]: http://0x0.st/oBAz.flac (alternative link if the first one dies).
But that doesn't prove anything, and more specifically it doesn't prove that there is an audible difference. What matters is the perception of the sound. In the end, the only reliable differentiator is a properly set up double-blind listening test.
The surprise part is not the difference itself, but the amount and legibility of difference even with a state of the art compressor at maxed-out settings.
> But that doesn't prove anything, and more specifically it doesn't prove that there is an audible difference.
On the contrary, it proves that some audible information is stripped from an MP3 file, even at highest settings, and this stripped detail inevitably adds more information to the sound stage. It wouldn't be there if there was no contribution.
Therefore, audibility of this information depends on the audio chain used. As I said before, a semi-decent amp with a good pair of speakers can render this, and the difference is not Oh, I hear a cymbal here, but Oh, this recording is really enjoyable and immersive.
The equipment I cite here is not a 5-figure system either. A good vintage amp (Akai, Sony, Marantz, Technics, etc.) with an "affordable" high end speaker (Heco Celan GT/Revolution 302, which are smallest members of their respective families) is more than enough for hearing the difference.
Investment in audio equipment has decent return in terms of quality up to a certain point. Past that point, it becomes meaningless for most people and scenarios.
Also, music has layers, and one listens a song completely in "multiple passes", this is why ABX tests play the same song 5-10 times and ask for evaluation.
Moreover, as per the study published in Audio Engineering Society in 2016 [0], MP3 changes the emotional perception of the songs, again verifying there's audible differences.
[0]: (Open Access) https://www.aes.org/tmpFiles/elib/20220528/18523.pdf
It doesn't prove that at all. Our perception of sound is quite complex. Sounds that would be very noticeable in isolation can become inaudible in combination with other simultaneous sounds, or even sounds that came shortly before. MP3 and other lossy compression schemes use a psychoacoustic model that takes all kinds of things like that into account.
A - B being audible does not mean that A and B can be audibly differentiated. Perception of sound is not maths.
> Moreover, as per the study published in Audio Engineering Society in 2016 [0], MP3 changes the emotional perception of the songs, again verifying there's audible differences
Yes, as I said, listening tests are what we should use as datapoints. Not mathematical differences between the waveforms.
That being said, in this case I see the study uses MP3s with bitrates of 32, 56 and 112 kbps. Is that a serious study? Those are very low bitrates for MP3! The authors of the study picked the lowest bitrate nearly indistinguishable from the original sound (i.e. already not perfect by their own admission), and then picked two bitrates even lower. No wonder the test subjects are going to hear a difference. Nobody is going to argue that those MP3 bitrates should be considered good quality. Any result obtained at these low bitrates is completely irrelevant for compression with even moderately decent bitrates (let's say starting from 160 with a decent encoder such as LAME).
I don't really care one way or another whether people can hear a difference between lossy compressed audio and the original (for me, with my ears and on my equipment, 192 kpbs MP3s are already more than good enough). What I do care about is the use of proper methodology.
Also this wasn’t some regular idiot, but a client that had overseen a marketing agency for 15+ years, so really she should know better.
As a bonus the paper I used for the experiment was the original stock that she was making a fuss over. (While paper is available in tints via different whitening approaches, those tints are noticeably different from one another.)
So it seems sometimes people have moments of perfectionism, it’s good to keep this in mind if we notice ourselves being fussy about a decision that probably doesn’t matter. It’s almost always better to just get on with it, than be paralysed by decisions.
I can *easily* hear the difference between MP3 at any bitrate and uncompressed on carefully-chosen programme material. In particular stuff that uses TR808 or TR606 hihats is absolutely destroyed by MP3.
I do not claim to be a golden-eared purist, and would roundly debunk the claims of the gold-plated SATA cable brigade. MP3 genuinely sounds really bad.
I'm not surprised. Most of the cheap devices containing tubes I see use a starved plate design, where the tubes are operating well below their ideal range, use a LED for "lightning" (because there's often no AC power to light up the tube filaments), all coupled with $1 reject tubes. Sure, it distorts very easily, so they can be used on music making, but so would a pair of diodes. This is just the next step in the evolution.
Seriously, though, they're called "modeling amps". And many of them sound quite nice.
https://www.amazon.com/Nobsound-MS-10D-Tube-Power-Amplifier/...
Arent't the tubes just the preamp stage?
That's not something necessarily bad.
The tubes are instead only for decorative purposes, and apparently to "turn the circuit on", albeit in a way that is difficult to detect. The article writer used capacitors to simulate the presence of tubes and the amp kept working as normal.
And my apologies if other reviews and discussion are wrong about how the tubes are being used in this amplifier, they could be just repeating incorrect facts they hope are right as well.
The article writer suggest that they're used as a timer, probably to simulate how tube amps take a while to turn on. He later "replaced" them with capacitors and the amplifier kept working as intended, which probably means they're being used as diodes to turn the circuit on, for purposes of faking operation.
This is all in the forum post.
If the schematic is correct, they're not preamps nor buffers.
Naturally that is a custom transformer, just not the high-voltage range that a normal tube amp requires.
Looks like the physical layout might have been originally designed to be fully equipped for high-voltage operation, which would mean two additional audio transformers would be needed in addition to a high-voltage version of the present power transformer. Transformers are probably the most costly components to the BOM.
The PCBs would all be different too, and a hobbyist having the proper transformers could discard the present ones and completely rewire for conventional high voltage operation, probably easiest on a point-to-point basis without bothering to design an alternative HV PCB for the proper passive components needed.
If HV transformers were ever originally involved I could imagine the supply of the HV units ran out before the supply of tubes & sockets. Therefore it could have been redesigned as a solid state amp more cost-effectively using the tubes mainly for decoration and to maintain a premium pricing cache. Especially if full tube amps were originally issued and gained a decent reputation to begin with.
It would also need 4 output power tubes for Stereo Class-AB operation, plus a couple of triodes for the inverter, so I don't think there's enough room there for it. If it were Class-A it would probably need much bigger tubes to deliver the wattage it promises, and there doesn't seem to be much space in the box.
I'm considering buying this fake Chinese tube amp amplifier to figure out what's up, they used their smarts to fake the tube function, but those are still smarts. Might help with IP protection, ironically.
Only as long as the oscillator is really connected. Which, after seeing this article, has a potential of not being the case ;)
I like your analogy to copy protection btw
I want to give you $100,000 for doing illegal work for me. I buy a $100k watch, hop on a plane, meet you for drinks on a yacht, and say "my friend, you are such a good friend, have my watch!" No taxes, duties, or paperwork.
Rinse, wash, repeat.
When daddy gives his daughter a several thousand dollar watch, he's not giving her a watch. He's passing along several thousand dollars of family wealth, to be used for both communicating the family's status but also convertible into currency if need be. A bonus is that it likely appreciates in value over time.
When daddy gives his mistress some jewelry, he's paying her for services in a way his wife (or her attorneys) can't trace.
Before the age of near credit card / digital payment/transfer systems and worldwide near-instant communication, rich travelers would wear a rolex because it served as a sort of compact money storage unit. If you're wealthy and traveling to places wealthy people liked to go and needed help, your rolex serves as proof of resources, or can serve as a payment itself.
The only advantage to a genuine mechanical watch is surviving uh...solar flares and EMPs. That's cool, hey maybe someday that'll matter. Especially solar flares, because that fucks with the light from the sky, like it did in the 19th century. Anyway, it'll matter, but not yet.
Also, there's expensive oscillators with much lower time variation, they're like $60 for a 100x lower Tau (I think it's called Tau, there's a whole language I don't understand). Same size, basically. Just two wires. That's the whole luxury element of a watch in it's essence, the idea you are getting time from it, the luxury of time. Jules Verne talks about this, in Around the World in 80 Days, there's a watch that loses a second per year. You could do that for real with that oscillator. To say nothing of the CSAC 45 chip on the market, kind of big, but somebody made a clock that loses a second per...million years or something, out of that. It just looks ugly, is the problem. Too big. Hey maybe it'll be cool to wear!
So far I'm not aware of any changes to the fundamental drivers that enable / motivate this.
I would expect a fake to be something like:
- it mimics a particular product but doesn't perform as well (e.g. SD cards that claim to be Kingston 128GB but are actually discard-quality 16GB)
- or it looks much like a real brand and product but isn't on close inspection (e.g. COBY vs SONY for quite a while)
- or it claims to be a particular thing and isn't (e.g. a tube amplifier that never needs replacement tubes because the tubes aren't connected at all to the LM379 IC actually delivering 6 watts.
All of those things are fakes.
This is a solid state amplifier with a tube preamp stage, which is... well, if you know what you're buying, I guess you could buy that, but I personally wouldn't. I suppose it looks like a tube power amp, superficially, but the nature of a tube pre-amp and a tube power-amp is to look pretty similar anyway.
The bigger general problems, I think, are in Amazon's total lack of care in co-mingling shipments sold on behalf of third-parties, and the general OEM problem where the same equipment has eight different faceplates or cases put on the same board and sold for a range of prices up to two magnitudes different.
No it isn't, if you check the schematics in the article you'll see that the tubes aren't biased properly and are mostly disconnected. They can't amplify audio. It's not even a starved plate design (which wouldn't need the 6.3V filament voltage).
The circuit is clever enough to stop amplifying as soon as one of the tubes is disconnected, so they're a glorified switch. Which IMO is worse than if the tubes were there only for asthetic purposes.
I'm specifically thinking of:
- Non-compliant USB / HDMI cables sold through Amazon and NewEgg.
- Some products from China being containing quantities of heavy metals or other toxins that violate U.S. law.
- Wall-wart power adapters whose product descriptions lie about the output voltage. Enough to cause problems for some but not all devices.
In retrospect, it seems unscrupulous Chinese manufacturers/sellers are just one part of the problem. They couldn't get away with it if Amazon, the FTC, Congress, etc. were more effective.
An online acquaintance informs me this was a "thing" 10-15 years ago.
I'm pretty convinced that as long as whatever-it-is looks the part, we don't care. And it could be tech like this, white goods (which are just a branding exercise), real estate even, and then even more ephemeral things like things we are presented with on screens. For most of us, even if we are holding something in our hands, how can we tell if is fake or not?