For those of us who’ve had to deal with distinctions between audio and control signals, this is actually pretty major. However, it’s already become something in modular synthesis. People who get started in Eurorack, for instance, may not need to worry nearly as much about different types of signals as those who used Csound or, more importantly for this marketing copy, Cycling ’74 Max.

This is when “the other shoe dropped”. They planned this all along (since 2012). Hidden in plain sight was a more subtle strategy than people might have imagined.

The continuity between these new modular features and the rest of the DAW’s workflow probably has unexpected consequences. Before getting information about BWS3, one might have thought that the “Native Modular System” promised since the first version might still be an add-on. What the marketing copy around this “killer feature” makes clear, it’s the result of a very deliberate process from the start and it’ll make for a qualitatively different workflow.

这是一个比较初步的简单实验。

图像结论其实并不意外：数据量越多当然表现越好；迁移学习在极小量数据上表现良好；Prototypical 模型可能因结构的特异性会表现出一定程度上的优势；数据量越小，过拟合问题越严重。。。

If the SM58 noise floor is calculated at room temperature, the voltage output is 0.00000032 volts.

This is equal to -130 dBV. The SM58 Vocal Microphone Specification Sheet says:

Sensitivity (at 1,000 Hz Open Circuit Voltage) –54.5 dBV/Pa (1.85 mV)

At this sensitivity, the self-noise would be 94 - (-54.5 - -130) = 19 dB SPL, which is pretty typical and certainly not "lower than can be typically measured".

Why should I add myself to "audio" group?

I'm reminded of a JavaScript tool written by Aaron Parecki that automatically adds a start fragment to the URL of his page when the audio on the page is paused. He's documented it here: https://indieweb.org/media_fragment

So, the umpteenth attack on FM, still it goes on.

Neil Gaiman reads "A Christmas Carol" by Charles Dickens. (A comment says the actual reading starts at about 11:50.)

That's an awful lot of models. Are we to understand that these are all limited to WMA, a proprietary format, and this is hard-coded so that it cannot be overcome with a software/firmware update? Seriously?

https://github.com/hypothesis/vision/issues/85