Is there a pattern between blacker background and less resolution?

Just for speculative fun though, a standardized test for plankton performance. Rough draft ideas:

1. Taxonomize different "species" of plankton by getting 30 or 40 examples of different types of audible plankton, from people qualified to identify it.

2. Analyze the sound wave properties of these different types of plankton for their mathematical features.

3. Categorize, reduce, simplify into a few basic types based on commonality of different quantitative features.

4. Make a short test track that runs the gamut of plankton types. Basically an almost black recording that lives in the less significant bits. For hard analysis this only needs to be a few seconds, but for fun, make it longer and segmented so that humans can listen to the different segments and subjectively perceive and experience the test track.

4a. A louder version of this track exists that's basically just left-shifted-bits so that people know what plankton they're listening for and what it's "supposed" to sound like if it hadn't been aritificially blackened.

5. Here's where my own experience could help with some ideas. Make some basic algorithms using RMSD, post- null testing data, a few other formulas all glued together, to create the "objective" plankton performance score. This score is what it is and isn't what it isn't. We don't want it enshrined on a temple altar similar to what some do with SINAD and such. It should definitely be paired with subjective appraisal of how different chains pull out the plankton and qualitatively express it. Indeed these subjective impressions would be vital in the early iterations of tuning the algorithm to produce an objective score that tracks as well as possible with the subjective experience of the enjoyment of micro-details within music.

Once we finish stage 5, we have the weaponry to affect global change. Because once audio manufacturers see their equipment exposed to meaningful tests and meaningful scores and measures, they have an "Oh, Schiit!" moment, caught with their pants down. Engineering goals and methodologies start changing. Optimistically, there is even a trickle effect that eventually influences the recording/production/mastering process. A renaissance in recording quality ensues.

(Wishful thinking isn't always bad, it's a key part of every major breakthrough and human innovation.)

 

Can not reference the tv analoges as I do not have a tv set. With dacs I find the noise that is eliminated by either in the recording or the dac takes the life out of the music. This noise allows the full sound to be heard, warts and all. I have had the LIM and Holo dac( our level 3 loaner) in my system and found both lacking as they lacked grit, density, noise(not sure how to say this) and the result was they were boring to me. The Burl 2 bomber and the Gungnir a2 MB have tone density, noise, grit and balls, they are alive. So I agree with Whitney that the blackness used in many dacs and music recording covers up the details.

 

I noted the part in bold in my LIM vs Holo Spring review, though I found the Holo more "offensive" in this regard to plankton. The LIM had more plankton but still less than the either BF2 or A1-GS. And far less than A2 (in relative DAC terms). Gungnir Multibit A2 I found was a very very plankton and tonally rich DAC that paired well with the warm ZDSE. But the Holo stuff still has a certain hi-fi excitement that may synergize better with certain gear/preferences.

 

I very much love this, specifically. So much so that a heart rating does not satisfy.

 

DACs all add noise from their line stages. Some of them are just terrible about playing back depth cues from phase shift, poor clocking, poor built in anti-aliasing filters with inadequate bandpass rejection, crosstalk, poor analog line stages. The easiest way to hear this is with recordings that abused older, noisier digital reverbs.

The mixing or mastering can kill the depth cues with poor gating and digital noise reduction. Denoise Led Zeppelin or My Bloody Valentine. Now imagine that to a lesser extent on every f'ing track of some overproduced, self-recorded pop monstrosity with hundreds of tracks where the only recorded things were some direct input guitars or bass and layered and edited vocals.

 

Noise is only noise if you can hear it. I just did a master of a pretty clean recording where the mix had a -65 dbfs noisefloor and the master had a -57 db noisefloor.

The only noises on Billie Eilish recordings are her mouth noises because of her total lack of mic technique using her crappy Neumann TLM 103 exacerbated by her brother's total lack of production skill. Yes pump up the weird mouth noises and breaths. The only plankton is whatever is swimming inside her mouth.

 

How is this for a simple test?

1. Record a low level signal (maybe a plankton exemplar) through the chain
2. Re-record shifted 6 bits to the left (or whatever)
3. Subtract 64 lots of (1) from (2) and see what's left.

This is strictly a time domain thing like Marv's burst tests. Probably need to average over a bunch of recordings.
Maybe it tells us something.

 

Yeah that's actually great. You'd either get what you want right away, or you'd suss out some starting features to hone in on. Noise tends to be entropic/chaotic whereas your plankton features will have recognizable quantitative patterns.

• "The trick would be how to tell what is noise and what is the music."

So #1, #2, #3, helps you suss out the real plankton pattern, or "DNA" so to speak. Garns also implied doing the same signal over multiple recordings. The chaos (noise) will vary on each recording, whereas the order (signal) will not. This gives you datasets where some simple averages, or hybridized sum squared error algorithms, or what-not, help finalize the capture of the plankton out of the noise.

THE N WHAT?

Once you've captured your plankton as outlined above, you're good to go. I could even imagine running SINAD on a DAC's faithful reproduction of the plankton test signal, as the ultimate irony that mocks the cult of SINAD. Remember, SINAD is not a SIN. The "sin" in SINAD is the use of SINAD on static-amplitude pure sine waves, as a holy proxy for musicality and performance. Putting the "AD" back in SINAD fixes it, where AD now stands for "acute details".

 

I mean I guess I'm a fan of measurements that convey time domain information. My thought is that this is like a ladder linearity measurement but with time domain information thrown in, maybe capturing presence or absence of time dependent nonlinearities. Even with something like the burst measurements it's not wholly clear how to interpret. This would probably be worse. But it seems interesting.

 

Hmmm. My thought for the plankton test was to keep it simple and have an inarguable objective measure showing artificial blackness at the expense of plankton. The measure captures that in a simple inarguable way but in so doing obviously does not capture other valuable things that could be argued for. So maybe what I was thinking of is not really a plankton test but a "false noise rejection test."

Because I think you're right, plankton is undoubtedly extremely time sensitive. Those little brushy scratches on the drumhead have to exist at JUST the right place in the peaks and valleys of the signal, for example, to convey their qualities in the analog out into the transducer.

Maybe plankton testing becomes a battery of several tests. I'm a strong believer each individual test has to be simple enough and have an inarguable enough logic to it, to become understood and adopted, used, and thus starting to create movement change in the audio and recording industries.