Thoughts on DSP multiamp/attenuator setup?

Every single bit increase on audio bit depth in sum just about double the theoretical reference points of signal resolution.
I honestly think that the potential signal degradation in the doubled amount of interconnects might have a slightly larger impact than a slight decrease in input signal on the adc.
It might be nearly impossible to measure any difference in signal quality between the two, so I think it is a waste.

Want higher quality? Use AES in and skip a conversion stage, live with digital volume control.
 
Every single bit increase on audio bit depth in sum just about double the theoretical reference points of signal resolution.
I honestly think that the potential signal degradation in the doubled amount of interconnects might have a slightly larger impact than a slight decrease in input signal on the adc.
It might be nearly impossible to measure any difference in signal quality between the two, so I think it is a waste.

Want higher quality? Use AES in and skip a conversion stage, live with digital volume control.

For absolute best-quality digital you're absolutely right. If I get rid of the multipurpose functions of the stereo that would work better, however, my "big rig" is multi functional so the analog in was pretty important for general use (without separate A/Ds)

Regarding progress:

I finally got things hooked up yesterday :wrench: Perfect example of "perfect is the enemy of the good", I was overthinking it before I even tried it. The network connectivity and software made it really easy to implement an EQ and crossover solution (right now I'm just using it as a sub processor/high pass filter for the 2 way mains). I wound up with some asymmetrical filters and notches, stuff that would have been very difficult with analog solutions.

We had a brownout while things were running (summer in new england) and everything fired back up cleanly afterwards, no scary pops or any such, which is very good to know both for peace of mind but also for power management solutions.

The 24dB attenuators were a good call, there would have been a lot of extra pre-dsp attenuation needed had those been excluded. Because it's a multifunctional system, sometimes volumes are very low (late night, background while working, etc) and other times more robust. The clipping settings in the DSP control gain structure, so there are a few levers to pull, but the analog attenuation was very helpful to have on the tops.

Key takeaways- unit doesn't have uglies on powerup/down, is working as intended, seems to be as-new from proaudiostar (at a very affordable $419). The software is limited in that some setup functions are front-panel only, but you can make full, high-resolution adjustment to everything that needs repeated tweaking- filters, XOs, EQ, compression, etc. Sound quality, I can't speak to as yet. It seemed to be a very neutral device, but it was a setup with a lot of changes happening so I can't speak to much of the little things.

I'll update as I continue to play with this setup- adjusting crossovers live from a networked laptop is very, very fun, and I can break out the other toys pretty fast and try some other serious stuff.
 
As long as you think it's worth it that's all that matters.
It would be nice if you would try to compare with/without the extra proaudiostar kit, if it is at all possible to try remain semi-neutral during testing.

Wow, somebody answers fast! The attenuators were from Naiant, I'm assuming those are what you mean. Proaudiostar is just the place with the great price on the Venu360 DSP itself.

I'll experiment with and without the attenuators as I get used to it, but we're going to be talking some pretty severe sensitivity differences in some cases. The tops are some 96dB or so, while the subs are probably 86 in their passband, but they're EQ'd to operate below Fc (sealed and boosted). When we get into 3 ways with a 108dB comp tweeter, high 90s efficiency midwoofer, and low efficiency subs, every way to adjust levels will come into play I'm sure.
 
Last edited:
I'm all digits until the DSP output into the amps. At that point, easy to make a little box with ganged pots (4 of them) to control the volume (2 mids, 2 subs) with a single knob. Couldn't be easier to make or work better.

B.

Right now I only use 4 channels, balanced, but likely this weekend I'll start with the serious stuff, using separate driver channels etc. 6 channels isn't really any harder to do, but if they're properly balanced... that's now 12 gangs. It's getting a little squirrely at that point, not to mention pots are the noisiest and least consistent performers out of pretty well all passive components (speakers get a pass).

I spent enough time building stepped ladder attenuators, wiring transformer volume controls, and all that jazz to know the pitfalls- I've done everything from single-stage stepdown with a mil spec pot hung off the back to relay-controlled ladder networks, and just fixed-attenuation for entire systems- couple resistors in-line, swap them out to change level. Knowing what's to be lost makes a compromised volume control a hard sell for a guy like me- even if my ears aren't what they were when I was designing those types of solutions some years back.

What I'm really looking forward to trying out are some of the horn bits- I'd previously used analog XO primarily, whether active, passive, or hybridized. Being able to just throw a curve at a horn is going to be very satisfying indeed, particularly when some things just don't fit right. The best measurements I've ever seen from a top octave were from a modified dome tweeter attached to a deep horn with a foam collar to keep device diameter small but add additional directionality below the natural device cutoff.

I took all the time prototyping the device, and made a solution that worked amazingly well as a supertweeter for a wide-band midrange type system, but never could make the XO work for a supertweeter in a horn system. Now, I just will throw protection caps on that and the main horn, hook 'em up to amps, and away I goes.
 
... Knowing what's to be lost makes a compromised volume control a hard sell for a guy like me.

OK, no need to go into the arithmetic of noise from pots at the far end of the chain. Just think about this: compare the natural ease and precision of twisting a knob which is perhaps at arm's length to your chair (like mine is) to fooling with a graphic interface.

(Not sure where you get 12 pot count from?)

B.
 
OK, no need to go into the arithmetic of noise from pots at the far end of the chain. Just think about this: compare the natural ease and precision of twisting a knob which is perhaps at arm's length to your chair (like mine is) to fooling with a graphic interface.

(Not sure where you get 12 pot count from?)

B.

12 gangs (not necc 12 devices) total for truly balanced operation- 2 per channel. Regarding end of chain, most of my gain happens in my amp, so anything immediately preceding that will have a significant impact. If I had lower gain in my amps (which would be preferable but I'm not screwing up my NCore with my giant mitts that can't solder SMD to change gain structure) I could do things somewhat differently.

My current front end setup runs into my pre-processor (McIntosh Mx121 for source selection, HT decoding, RIAA EQ) into the Venu, so from an ergonomics standpoint, I get a knob or a remote (When I'm streaming I can also control volume from the network client, digitally, with a phone or remote).

If I'm so inclined, I can always set up the system for AES/EBU stereo input and forego HT and Vinyl functionality for that time.
 
12 gangs (not necc 12 devices) total for truly balanced operation- 2 per channel....If I'm so inclined, I can always set up the system for AES/EBU stereo input and forego HT and Vinyl functionality for that time.
I went through similar contemplations so you have my complete sympathy and without me trying to second-guess your choices* except to suggest have a look at my signature below.

I presently have my Macbook Air as the centre piece. Analog sources like my golden-age Sony FM tuner go to an ADC, then into laptop via USB (although a sound-card input (if Macs had such things) or the stock Mac analog input measures perfectly good - surprise).

My Behringer DSP is fed by an inexpensive USB to AES converter, so digital right to the DSP output stage.... which makes it logical to use 4 ganged VCs.

You may vomit at the thought of having the output of a fine moving coil cartridge reduced to mere dots and dashes. Until you try it.

B.
* a bit odd you think the big gain takes place in your power amp
 
I went through similar contemplations so you have my complete sympathy and without me trying to second-guess your choices* except to suggest have a look at my signature below.

You may vomit at the thought of having the output of a fine moving coil cartridge reduced to mere dots and dashes. Until you try it.

B.
* a bit odd you think the big gain takes place in your power amp
26 dB of gain from the amplifier. There’s dac output stage gain in the prepro of course, but I view that as the source level. I have run arrayed 1541s with no output stage, just transformer loaded, but really, isn’t it best to consider fixed source gain as source level for the purposes of a setup like tuhis? Similar situation for the hdmi, lp, and all other fixed level devices (without modification).

So when I say gain, I mean that I’m running both the prepro main level at less than unity taking the dac and phono as source level before the mc attenuates/buffers, the dsp at a little more than unity in the bass, and more than unity but attenuated at amp input for the 80hz+, and then 26dB from the amp for both the bass and full range.
 
I'll experiment with and without the attenuators as I get used to it, but we're going to be talking some pretty severe sensitivity differences in some cases. The tops are some 96dB or so, while the subs are probably 86 in their passband, but they're EQ'd to operate below Fc (sealed and boosted). When we get into 3 ways with a 108dB comp tweeter, high 90s efficiency midwoofer, and low efficiency subs, every way to adjust levels will come into play I'm sure.

It should be very easy to match sensitivities using the DSP software.

I assume you have already measured each driver in isolation ... would you mind post the measures here? :)

How are you approaching the time alignment between the several drivers?
 
It should be very easy to match sensitivities using the DSP software.

I assume you have already measured each driver in isolation ... would you mind post the measures here? :)

How are you approaching the time alignment between the several drivers?

The DSP is certainly capable of matching things for me, I am experimenting with solutions still. It's a slow process for me- working stiff with a family. I listen over a 50dB window or so (from quietest to loudest). I don't want to start the chain off with an A/D based off a signal at -80dB- if I attenuate output instead of input, I can feed it some 24dB more.

I've measured the driver/horn/cabinet combos in question before but don't have a good working in-room measurement, or the level of resolution I want for most measurements. That's one of the next steps as I start doing a little more with the setup.

Time alignment, for now, is not key as the only DSP crossover is at 80hz.

The mid/top are a 2 way constant directivity design, on top of dual sealed 12"s per side. They're based on JBL 2206h (with a thin damping/sealing coating), and DDS ENG90 waveguides with JBL 2426h.

I'll begin doing some more robust measures as I start doing more detailed builds that leverage the DSP for more crossovers.

With respect to time alignment, there are a few ways to do it- I can measure impulse timing of individual drivers, which may be the easiest.
 
Thank you very much for the detailed response.

I fully understand the slow process, i have all the gear for my L/R 2-way stored away for more than 6 months, composed by:
- AE TD12M @ 4ohms
- Beyma TPL-150H
- DSP Monocor DSM-48LAN
- Crown Xli 800(for the TD12M)
- Parasound V3 (for the TPL-150H)

Work, family, etc, have been in the way, i'm hopping to complete this in the next two months.

I will follow this build with great interest, i'm sure it will help me learning and achieve a better result for my own project.

Please keep this thread updated :)
 
The DSP is certainly capable...I don't want to start the chain off with an A/D based off a signal at -80dB- if I attenuate output instead of input, I can feed it some 24dB more.

Getting warmer. Easy to gang Bourns pots (and a big knob) to control the amps. Easy to buy a resistance taper that feels just right to your fingers and ears.

Bourns Potentiometers | Mouser Canada


For those with a Behringer DCX2496, you might consider using a USB to AES-coax converter - cheap, invisible, and maintenance free. Skips over the input analog stage and eliminates two AD/DC conversions.
B.
 
Last edited:
Sooooo long overdue update as I've gotten a little more experience with the Venu.

First- I'm happy as a clam that I bought the attenuators. When running digital in, even though I know that truncation is a non-issue from a tech standpoint, I still do want to be in the top half or quarter of the digital volume control. The attenuators, combined with the output clip level controls, let me either run digital into the Venu, or up the output and run analog (the AES digital in is louder than the analog from the McIntosh Mx121 I use.)

Doing this requires a quick crossover reconfiguration, including unplugging the conversion transformer from the crossover input, plugging the analog preamp out into that input, reconfiguring inputs to analog, and upping the output clip level adjustments. Takes me 2-3 min max.

Basically, the system is currently configured thus:
Diagram.png

The better solution would be to reconfigure the power amp gain but not really into SMD work. Ideally I'd like to figure out how to do some other things but for now, having options for optimizing signal path into the crossover in either analog or digital cases is pretty great. I wish I had a good way to convert from either HDMI or Optical to AES/EBU , with a level control, but haven't found a decent solution there.

The key challenge with this rig is the desire to maximize the quality of multiple sources with differing connection types.
 
Last edited:
]The better solution would be to reconfigure the power amp gain but not really into SMD work.

I've gone to use a decent PA amp with one volume control pot per channel. Many modern recordings have very "bright" top end because of compression, nice to turn down the tops sometimes. On properly recorded stuff I have them all set to the same level. Normal use is usually around 9 o'clock absolute max. On very dynamic recordings I can sometimes turn them to 3 o'clock.
No extra boxes or connectors here, just:
Soundcard with foolproof big volume knob -> DSP -> 4 channel amp -> speakers.
 
I've gone to use a decent PA amp with one volume control pot per channel. Many modern recordings have very "bright" top end because of compression, nice to turn down the tops sometimes. On properly recorded stuff I have them all set to the same level. Normal use is usually around 9 o'clock absolute max. On very dynamic recordings I can sometimes turn them to 3 o'clock.
No extra boxes or connectors here, just:
Soundcard with foolproof big volume knob -> DSP -> 4 channel amp -> speakers.

Computer source solves a LOT of my woes, but unfortunately doesn't play nice with some of my sources that I just don't want to give up yet. Right now I can switch system configs from analog/ht to digits with a few buttons, and no meaningful latency.

That's why I need so many durned boxes :) Qobuz streaming 96/24 directly into the DSP is pretty sweet.
 
Hi badman.
Pano posted, on the first page, a great guidline. I was ,if I understand your dilema, in a similar spot many years ago. I set up the power amp (s) to yield the max db output I wanted when the xo channel outputs were maxed as well as the xo main input. The way this setup works then is to just adjust the input to the xo for proper level, the channel outputs are the used to adjust overall db requirements. No need to touch the power amps because I have them already maxed out db for my room. All adjustment happens at xo input. One knob. Hotter cds, other equipment, doesn't matter.
Hope this helps.
 
Last edited:
Hi badman.
Pano posted, on the first page, a great guidline. I was ,if I understand your dilema, in a similar spot many years ago. I set up the power amp (s) to yield the max db output I wanted when the xo channel outputs were maxed as well as the xo main input. The way this setup works then is to just adjust the input to the xo for proper level, the channel outputs are the used to adjust overall db requirements. No need to touch the power amps because I have them already maxed out db for my room. All adjustment happens at xo input. One knob. Hotter cds, other equipment, doesn't matter.
Hope this helps.

From an operational perspective, I'm at one knob- For analog/video, it's an analog domain control in the McIntosh, for digital, it's operating digitally in 24 bits on the squeezebox touch. There should be sufficient headroom, though I can change the output level settings as needed.

With respect to Kaffi- wax isn't the only issue (and is best served in a digital system by converting and de-noising). I do have integrated video including a bit of gaming where latency is a huge no-no, but more and more my move seems to be long term towards PC sourcing/XO into a good external A/D+multichannel D/A. Solves so many problems, though good hardware still is pretty pricey.