SkoomaDentist
The Greater Finnish Empire
No bio...
User ID: 84
Whoosh... (see also and from 1:10 onwards)
Do you use commercial models/implementations or do you have a recommendation where to start if I want to set it up myself?
For photo noise reduction I've used Lightroom and (free but specific to Olympus / OM cameras) OM Workspace. I don't think there are good open source models as the models are camera specific due to different sensor and bayer filter characteristics (and why they can produce really impressive results - check the OM-1 original vs denoised comparison here where the original is far from what I'd consider usable while the Lightroom one is perfectly fine).
For audio stem separation Ultimate Vocal Remove is my tool of choice (you can download the models from the settings page). I start by removing vocals with MDX-Net / Kim Vocal 2, take the residual and remove drums with Demucs and then possibly remove the bass from the residual of that. Be aware that if you just split the stems from the original they will not sum to 100% and thus you want to go the recursive route. I'm sure there are some newer models that you can install manually but I haven't used those as the existing ones work well enough for my purpose (removing distracting vocals or emphasizing instrumental part to better hear what's happening).
Also extremely impressive considering it predates any non-research AI models by at least half a decade or more.
LLMs and other generation AIs are much more HW intensive than domain specific neural networks. 500M - 1B parameters is plenty when the model doesn't have to understand instructions or global context (hell, there are some task specific audio models in wide use that can fit in just a couple of megabytes while performing well). Sure, a more powerful gpu will run the models faster but when you're doing noise reduction on a dozen culled and selected photos it doesn't really matter if it takes 15 seconds or a minute to run the process when on cpu it would take half an hour. Likewise stem separation taking a few minutes is a non-issue when you only need to run it once or at most a few times for a song (such as when remixing or isolating instruments for practising the lines).
There are apps that can run in the cloud but they're (expensive) subscription or credit based and having to pay $30-$50 per month per app gets really expensive. Not to mention they tend to have ridiculous censorship (think photos of people on a beach where you want to remove some distractions and the cloud version complains that your content breaks the terms of use).
I'm a firm believer that nothing important has happened in the PC space for the last 15 years
There is one exception to this: A gpu that can run local AI models. Not LLMs or image generators but the sorts of custom neural networks that have become quite common for eg. photo noise reduction, audio stem separation etc. In those cases even a fairly low end "good gpu" will do fine (such as this ancient NVidia Quadro P2000 Mobile aka Gefore GTX 1050) but the difference between "have" and "have not" is massive.
Office documents and PDFs, along with even many browser tabs, are definitely a solid set of real-world work computer requirements, but they're also not that demanding.
A better way to think about it is that web pages and apps are ridiculously ineffective while Office and PDF viewers are if not optimized, at least not 1000x pessimized by design simply due to having decades long history, no time for rewrite from the ground up in "modern" languages and generally the core is written by people who have at least some idea what they're doing.
A general rule of thumb I've found is that it's flat out impossible to have too much disk space if you use the computer for any sort of entertainment or hobby stuff beyond web browsing and the right amount of RAM is to take your initial estimate and double it. Hence me using this 2018 model laptop for C++ development, photo editing, running AI audio tools etc without problems because I made sure to upgrade to 64 GB ram and many TBs of SSD storage.
'Man wakes up in 2040, tells his friend "I want to fuck a toaster". The friend replies "You know you can just order a sexbot, right? I really don't want to know about the details."'
- Prev
- Next

Agreed. What most frustrates me about Covid talk on The Motte is the insistence that there were only ever two situations in the western countries: a full lockdown or the Swedish "let's do nothing"-approach. As if my country (you know, right next door to Sweden) with zero legally mandated "lockdowns" but a bunch of voluntary recommendations and public health response changes didn't exist.
I kept track of restrictions during the Covid era and the only government mandated ones were restrictions to large events, bars, restaurants and gyms. Everything else was voluntary (including bar / restaurant closures when the pandemic started) or just recommendations with no penalties. The officials outright recommended that "going out in the nature is a very good idea now".
More options
Context Copy link