3mo ago

Swedish amplifier enables transmission of 10x more data per second

interestingengineering.com

13 comments

I totally get how this would be useful in imaging systems, but I'm not understanding how it applies to communications.
The only thing I can think is perhaps carrying more modes through a multimode fiber? I never understood amplifier bandwidth to be a limiting factor, though.
What communications systems use a wide bandwidth of light (300nm is a LOT) into a single amplifier?
- That's a great question. My guess is the bandwidth comes from bonding those extra modes and from the lower signal-noise ratio. That lower SNR means they could modulate with more sensitive but faster modes.
  
  I wanna know if I can plug my guitar into it
- In terms of industrial applications, the abstract states
  We have realized all-optical wavelength conversion for a more than 200-nm-wide wavelength span at 100 Gbit s−1 without amplifying the signal and idler waves. As the 32-GBd 16-QAM is the dominant modulation format of current optical-fibre communication systems connecting the continents on Earth, the Si3N4-chip high-efficiency wavelength conversion demonstrated has a bright future in the all-optical reconfiguration of global WDM optical networks by unlocking transmission beyond the C and L bands of optical fibres and increasing the capacity of optical neuromorphic computing for artificial intelligence.
  https://www.nature.com/articles/s41586-025-08824-3
Is 300 nm the diameter of the optical cable? This terminology breaks my brain, 300 nm is 1000 terahertz, which is unreasonably large for a signal bandwidth, it's like one milllion Ethernet cables.
- I'm only making assumptions, but I'd guess that 300nm is the range of frequencies it can amplify. AFAIK fibre cables are used with multiple "channels" by sending data with different frequencies at once. Say your signal range is centered around 850nm, this amplifier could amplify in the range of 700-1000nm.
  But I might be totally off, just guessing.
- From the abstract: "we obtained a continuous-wave gain bandwidth of 330 nm in the near-infrared regime. [...] Furthermore, we realized wide all-optical wavelength conversion of single-wavelength signals beyond 100 Gbit s−1 without amplifying the signal and idler wave."
  Here is the paper: https://www.nature.com/articles/s41586-025-08824-3
  I think figure 4 from the PDF shows it the best. Their amplifier covers 1400 nm to 1700 nm infrared lasers.
But does it go to 11?
- 10 and ^15/16ths sadly. It's being marketed as AI boosted to 12 though!
Audiophiles will go nuts over this
- Honestly I have no idea how audio encoding works but I just imagined going from 80 decibels to 81 decibels (10x) and an audiophile losing his shit
- When I was sort of an audiophile, my first steps were to switch from ALSA to OSS under Linux. I still think I heard the difference for music, and probably yes - because that was OSS without mixing.
  And I still think I hear the difference between FreeBSD newpcm and Linux sound stack. newpcm's mixing is simpler (firefox starting, opening a sound device and music volume sharply dropping in half is not nice), but it seems to spoil the sounds less.
- I can already hear them

13 comments