Principles: Time of Arrival Delays, vs Phase tuning ?

[Wolf]

I've not used the 830668, so not sure to what you are referring.
In my Purveyor build, I used that PR, with the 5.25" SLS woofer, not the 10".

 

[mdsimon2]

Your posts^^ were claiming how bad the sound is if DAC’s are not synchronized because of delays, which the above example shows you that is non-sense

I believe there are two separate issues being combined into one and causing confusion. One multiple DAC setup has an issue with asynchronous clocking and the other does not.

Case 1
Multiple DACs being fed from a synchronous multiple output AES or SPDIF or TOSLINK or I2S device (such as a miniDSP nanoDIGI or a multiple digital output audio interface or RPi5 I2S output). Downstream DACs are synchronized via their AES / SPDIF / TOSLINK / I2S input. I agree that any local clock drift in downstream DACs (say an internal clock used for ASRC like in many ESS DACs) does not matter and everything will be sync'd when at the DAC output.

Case 2
Multiple asynchronous USB input DACs being fed by a USB host with some means to create an aggregate audio output device (such as VB matrix or ALSA multi device). In this case DAC outputs will NOT be synchronized, resulting in nulling on the order of several dBs which will vary over time. This nulling is measurable both electrically and acoustically.

Michael

 

[john61ct]

Not my claim, but seen pretty universally stated as fact.

I have no plans at this point to use "DACs" as such, but multiport format converters that include ADC and DAC per port.

The problem of cost arises when going past 8 channels, so I'm looking at going to ADAT expansion "breakout boxen" coupled with low cost interface that functions as Master Clock

so hopefully the question becomes moot.

Using dozens of two-port "just DACs" does not solve any of my needs.

Remember here the topic is using DSP for Crossovers, specifically the distinction between phase tuning and delays to handle "flight time" differences.

 

[john61ct]

My bad, it was the 830945

https://www.audiosciencereview.com/...ements-peerless-sls-series.33333/post-1165199

I've not used the 830668, so not sure to what you are referring.
In my Purveyor build, I used that PR, with the 5.25" SLS woofer, not the 10".

I guess I'd need two of those PRs, they might look cool stacked horizontally in my "wannabee Overdrive" 12" cube rather than one 10" Cyclops.

 

[john61ct]

THANK YOU!

In my case, say I have say eight analog streams, same music content just different full-range channels.

My choices for crossover splitting / routing these further (if DSP is required say over 16 endpoint channels) are:

multiple RPi5 units running HLC or CamillaDSP and 8-port HATs

or a single PC + interface w/ ADAT breakout boxen, total say 8 analog ports in and 24 out.

I would love it if the RPi5s could do just as well as the PC, even if it takes more time & trouble.

What say you?

The transportation sync issue may be resolved using SPDIF, and the different convolving times compensated for by inserting flight time delays with the "slower" units.

I believe there are two separate issues being combined into one and causing confusion. One multiple DAC setup has an issue with asynchronous clocking and the other does not.

Case 1
Multiple DACs being fed from a synchronous multiple output AES or SPDIF or TOSLINK or I2S device (such as a miniDSP nanoDIGI or a multiple digital output audio interface or RPi5 I2S output). Downstream DACs are synchronized via their AES / SPDIF / TOSLINK / I2S input. I agree that any local clock drift in downstream DACs (say an internal clock used for ASRC like in many ESS DACs) does not matter and everything will be sync'd when at the DAC output.

Case 2
Multiple asynchronous USB input DACs being fed by a USB host with some means to create an aggregate audio output device (such as VB matrix or ALSA multi device). In this case DAC outputs will NOT be synchronized, resulting in nulling on the order of several dBs which will vary over time. This nulling is measurable both electrically and acoustically.

Michael

 

[Waveform Fidelity]

Two issues as I see it. First the general comment about DAC’s and drift causing audible problems is clearly not true. Second, I can see how if a design using Async USB DACS are not locked to their high-stab crystal reference that cancellation might occur, but that is more about a design flaw in the design, not about any fundamental limitation of all DAC’s

 

[john61ct]

I don't see it as settled for phase issues, but will just withhold my judgment pending further input, documented testing.

Maybe even by my self at some point.

 

[KSTR]

Case 2
Multiple asynchronous USB input DACs being fed by a USB host with some means to create an aggregate audio output device (such as VB matrix or ALSA multi device). In this case DAC outputs will NOT be synchronized, resulting in nulling on the order of several dBs which will vary over time. This nulling is measurable both electrically and acoustically.

Correct.

I would think it should be clear to anyone that multiple independent DACs can only "work as one" if the master clocks are synced, or ideally, coming from one single clock source.
Otherwise they'll drift apart, it is like the futile attempt trying to run two CD or record players in full sync.

Therefore, either the signal is the clock (AES, SPDIF) or the local clocks are physically the same. With DIY this can be done, letting one DAC being the clock master and "copying over" the clock output(s) of the main clock(s) to the other boards with a cable connection, disabling their local clocks.
Some USB DACs have an option to sync to an external clock, either a 10Mhz master clock input or an SPDIF input. You'll still have a sample offset with these of a few samples but that offset stays constant which is the all-important detail.

 

[KSTR]

Multiple rPIs (gen5) can be synced by using I2S in slave mode, with either one rPI being the I2S master or, better, letting one DAC chip be the I2S master (if the chip supports master mode, that is). rPi5 supports 4 I2S data lanes (8 channels). With the usual 8 channel HAT's one would need to patch the driver to set up I2S slave mode, though. Sample offset will likely need to be addressed in some way, and in general the syncing of the logical data inputs to the I2S engines.

 

[Bland_username42069]

I am using the SYN for that, not trying to "make" anything related to that.

AVR means AC powered amps right? Yes I don't want that.

I can always slot an AVP in instead if I want later.

You can just use the pre outs from a Dolby logic processor into whatever amps you please

 

[john61ct]

Thanks so much for your guidance, but my budget is so low, and desired number of ports likely so high

trying to get what I need at below say $10 per analog port.

Target host OS is anything except Windows unless impossible to avoid.

Happy to futz around with firewire and PCI (pre PCIe) and even Cardbus/PCMCIA if needed, so even XP or Win2000 are on the table, rather than newer versions.

But really, IF the clock sync issue is a myth, RPi5 at 8-ports is my ideal way forward.

8-ports per unit, might need 3-4 of them

Anything else with USB seems out of reach, compare Clarett+ 8Pre or Scarlett 18i20 vs Saffire Pro 40 on eBay

Also trying to get units without preamps, and with external Word Clock, in case that clock sync issue turns out NOT to be a myth

And all this is only relevant, to the extent I find I even need DSP at all.

ADAT interface hosts, high SQ but cheaper than RME Digiface USB?

I am looking for an interface "hub" based on multiple ADAT, clean and transparent. I'm not recording / mixing, just for recreational home DSP - crossovers timing delays / phase alignment tuning Happy to buy secondhand to save money, vintage discontinued no problem. 24-bit, 44.1 / 48kHz is...

www.audiosciencereview.com

As good value for PC I/O as Echo Audiofire 12 ?

Total interface noob here, please ELI5 I don't need to record or mix music in the creative sense, just looking to get multi-channel analog in and out of HTPC type hardware. Primary use case is DIY convolving, loudness contour, bass management, room correction PEQ. The Echo Audiofire 12 unit...

www.audiosciencereview.com

ISO ECHO Darla, Mona, Gina or Layla-24 host adapter cards & cables

From the interface vendor ECHO, specifically the PCI bus (or Cardbus!) cards shipped with the above 24-bit interfaces, also the DB-9 cables Anyone have one or more of these? not those compatible with the OG previous 20-bit versions, nor the later "3G" models. The breakout boxen are commonly...

www.audiosciencereview.com

Yamaha AD824 Converter AD input only

Looking for, with the MY8-TD ADAT version 8-port Mini-YGDAI Expansion Card what I've gathered not an interface, just "breakout box" converter, in my case I'll need an interface with ADAT inputs, e.g. RME Digiface USB zero digital outputs on its own Usually the included Mini-YGDAI Expansion...

www.audiosciencereview.com

ISO cable for OG RME Digiface HDSP CardBus/PCMCIA connection

My goal is not recording or mixing / producing music, but a high count of multichannel DSP and crossovers for recreational reproduction, cheap. If members have other similar gear sitting in closets, or alternative suggestions feel free to speak up. In this case I like the RME OG Hammerfall...

www.audiosciencereview.com

In general, I don't recommend buying a cheap interfaces from small manufacturers. I own 3 interfaces - RME Fireface UC, Focusrite 2i2, and Presonus Audiobox USB. I bought the Presonus first, and it had all sorts of driver issues. The firmware has not been kept up to date, and the drivers no longer work. I could probably try to make it work with Windows compatibility mode, but I can't be bothered. It is now a brick. The Focusrite is 15 years old, and the RME is 11 years old. Both RME and Focusrite have kept their USB drivers up to date, and both work on my Win 11 PC.

If your budget is low, I strongly recommend Focusrite.

 

[john61ct]

You can just use the pre outs from a Dolby logic processor into whatever amps you please

I have no interest in any hardware like that.

 

[john61ct]

Multiple rPIs (gen5) can be synced by using I2S in slave mode, with either one rPI being the I2S master or, better, letting one DAC chip be the I2S master (if the chip supports master mode, that is). rPi5 supports 4 I2S data lanes (8 channels). With the usual 8 channel HAT's one would need to patch the driver to set up I2S slave mode, though. Sample offset will likely need to be addressed in some way, and in general the syncing of the logical data inputs to the I2S engines.

Thanks but since the RPi5 only supports 8 pots each, they will need to be placed at various physical locations seeving different groupings of amps downstream, separate from the stereo "head unit"

I do not want to get into I2S except within a single enclosure, and no HDMI for audio.

Plain normal OTS stuff, S/PDIF or network transports, I'm OK with figuring out PipeWire and JACK, but patching anything software, compiling?

nope, unless there's literally no other way and will save like a grand. Same with soldering

 

[john61ct]

I would think it should be clear to anyone that multiple independent DACs can only "work as one" if the master clocks are synced, or ideally, coming from one single clock source.

Yes I tend to believe you, lots of very tech-masterish members on multiple forums say the same, as opposed to the one guy insisting it does not matter.

Even relying on "the signal is the clock (AES, SPDIF)" I see as uncertain, but promising maybe for a group of distributed Pi.

If I do go for a cheap legacy "one big Kahuna" interface + ADAT-connected breakout box converters (see post above)

> 10Mhz master clock input

that sounds expensive

I see filtering for converters with BNC Word Clock does not add to the price at all

 

[john61ct]

the general comment about DAC’s and drift causing audible problems is clearly not true

Please stop saying that without hard evidence.

At least in my threads

 

[Bland_username42069]

I have no interest in any hardware like that.

lol. It’s literally an upmixer with dsp.

 

[john61ct]

lol. It’s literally an upmixer with dsp.

I'm avoiding proprietary digital upmixing completely for now. Maybe one day if I get into decoding, I hear good things about DTS Neural:X

but that will require some financial windfall.

Also, I want DSP to be "open", try various filter creation software, choice of convolvers. Even miniDSP seems too proprietary.

And are you suggesting 16+ endpoint bass management channels are available cheap?

Finally DC powered, and I don't want big heavy gear with amps I plan not to use.

I appreciate your helpful intentions, but prefer staying within the boundaries I've set so explicitly, even if from your POV they seem irrational

 

[-Matt-]

but that will require some financial windfall.

It really wouldn't. Picking up an old AVR would be by far the lowest cost approach. You could probably get one for $50 that will do the majority of what you have planned. Anyway, it seems you are ideologically opposed to that approach, so I'll try to avoid mentioning it again.

(Although it might help if you could explain your reasoning for this position).

 

[john61ct]

It really wouldn't. Picking up an old AVR

I was talking about DTS Neural:X and a long future hypothetical.

I think I fully explained above, beyond all those rationales just call it an allergy

 

[john61ct]

relevant

First pass at overall system plan, DSP multichannel - "simple version"

context, end goals https://www.audiosciencereview.com/forum/index.php?threads/principles-time-of-arrival-delays-vs-phase-tuning.71596/post-2608572 https://www.audiosciencereview.com/forum/index.php?threads/analog-crossovers-only-using-dsp-to-treat-issues-the-analog-system-cannot-fix.70409...

www.audiosciencereview.com