[sdiy] Niche ACM-1 how's this thing work?

brianw brianw at audiobanshee.com
Wed Mar 22 07:14:44 CET 2023


Also, yes, the stepped analog gain for the preamps can be a challenge. The circuits can be designed for ultimate fidelity, but end up with horrible stepping (meaning: don't ever change the preamp gain after sound check), or they can be designed to minimize zipper sounds with some sacrifice in quality.

THAT Corp. actually came to Mackie to see the preamp circuit design because Mackie managed to minimize the zipper noise in their DL32R digital mixer enough to impress the manufacturer of the chips.

But the pure digital gain stages within the mixer (after the mic pre stage) can be completely smooth.

Brian


On Mar 21, 2023, at 11:10 PM, brianw <brianw at audiobanshee.com> wrote:
> Digital mixers can go beyond 24-bit stepping, and smooth the steps out. There's a lot of math going on in the bus mixers of digital mixers. A first-order approach is to just low-pass filter the "gain" value (like filtering analog CV). It's not even necessary to interpolate because the low-pass filter is continually doing that.
> 
> But, yes, if the controls for the digital mixer are stepped, then that is a source of a problem.
> 
> You have me curious about trying to pit a $5,000 Mark Levinson preamp against a modern digital mixer.
> 
> Brian
> 
> On Mar 21, 2023, at 10:59 PM, Mike Bryant <mbryant at futurehorizons.com> wrote:
>> Based on what you are saying, most digital mixers should also be unusable with that type of music as they also step the gain in increments.  There's a little interpolation but not much.  And if you have to change the input gain control which is stepped analogue gain followed by a digital correction, there's no hope.
>>> From: brian
>>> Sent: 22 March 2023 05:43
>>> 
>>> Perceptible noise is about more than just dB differences - it's also important to consider frequency.
>>> 
>>> For electronic music with strong, low-freqeuncy bass notes, the zipper noise will be in a different octave. Psychoacoustically, that means the zipper noise is in a different critical band (see Bark scale).
>>> 
>>> I've heard some audiophile products in the $5,000 range with DPOTs and rotary encoders that sounded fine playing The Beatles, but once Aphex Twin hits the speakers there is zipper noise all over the place when adjusting the volume encoder. Those even have the zero-crossing circuits, but it doesn't help because the gain is still changing in steps, and that amounts to FM (i.e the step contains infinite frequencies, and the step multiplies the audio signal).
>>> 
>>> Brian
>>> 
>>> 
>>> On Mar 21, 2023, at 8:03 AM, Mike Bryant wrote:
>>>> With a 10 bit MDAC or DPOT, any zipper noise should be 60dB down on full house.  Now of course we don't operate at full house so you have to take your headroom off that but chosen carefully it shouldn't be hugely annoying.  But if it is there are some DPOTS with zero crossing switching included so they only change gain as you pass through zero, eliminating most annoyances.
>>>>> From: Tom Wiltshire
>>>>> Sent: 21 March 2023 14:56
>>>>> 
>>>>> This is a good point. MDACs or even digital pots work best in situations where you set a value and then leave it. That's ok for quite a lot of programmable stuff, but it doesn't suit things where there's an element of performance control likely. For that, a DAC controlling a VCA is a better bet, since the VCA's CV can be heavily smoothed to avoid any zippering. A mixing desk is a bit of both depending on how you use it, so it's a tricky one.
>>>>> 
>>>>> The old Sequential Pro-FX 500 system I was looking into some months ago used plenty of MDACs ( the 7524, 8-bit, still available) for audio level control. It was probably a top-end solution at the time.
>>>>> 
>>>>> On 21 Mar 2023, at 14:38, Roman Sowa wrote:
>>>>>> This may not going to work in every application though. Imagine Buzz Lightyear talking to Woody: "glitches and zipper noise, glitches and zipper noise everywhere"
>> 




More information about the Synth-diy mailing list