koolbass

The monitored signal does bypass the A/D/A converters. That's the whole point.


Most interfaces provide direct monitoring, either by software-firmware switching transistors or by a hardware button making a direct hard-wire connection between the inputs to the outputs (possibly also operating through transistor switching). Either way, it's a direct analog path from input to output with a parallel path to the A/D converters so the host can record it.


The purpose of ADM is to allow you to do that switching from the host. The standard also allows for passing pan and level settings from the host to the interface so the direct-monitored mix will match the playback mix (less any plugin FX of course).

Most interfaces provide direct monitoring, either by software-firmware switching transistors or by a hardware button making a direct hard-wire connection between the inputs to the outputs (possibly also operating through transistor switching). Either way, it's a direct analog path from input to output with a parallel path to the A/D converters so the host can record it. The purpose of ADM is to allow you to do that switching from the host. The standard also allows for passing pan and level settings from the host to the interface so the direct-monitored mix will match the playback mix (less any plugin FX of course).

What you're describing is NOT ASIO DIRECT MONITORING but just "mor-me" monitoring of the analog audio signal.  RME uses, what they call, TRUE ASIO DIRECT MONITORING with what they call "TotalMix" and MOTU has something similar called "Cue Mix".  It actually takes the signal, ONLY if you're using ASIO mode, and converts the audio signal to digital.  At that point it copies the digital information, and sends the exact same info TWO different places.  Packet one goes to the DAW and packet two, which is identical, goes to the digital to analog OUT converters.  BUT, with RME, you have the ability to mix and blend ALL of your audio in and outs any way you want.  So, in the case of my particular interface, you can have 5 different musicians with 5 different stereo mixes, hearing oneself and the other musicians, and all have ZERO latency monitoring, mixed along with whatever they want being sent from the DAW, be it a click or pre-recorded material.  When you're also monitoring pre-recorded material, you can use the 4 different sends along with the master out to have five DIFFERENT mixes/levels for each musician of the pre-recorded music along with the live music.

What I've just described is not simple to set up, but RME and Sonar both allowing the saving of templates, so, after certain set-ups, it's easy to save and recall.

If you don't understand what I'm talking about, I suggest going to the RME site and look up Asio Direct monitoring.  Once you actually understand and use it, it's not so complicated.  And one could NEVER have true zero latency monitoring from a DAW, because you have the time it takes to travel to interface, convert from analog to digital, travel time to cpu, numbers crunching, sending back to interface, converted from digital to analog, and then sent through to speakers.  True Asio Direct Monitoring is into interface, convert analog to digital, software mix within interface hardware and immediately back out through the digital to analog converters.

I'm more inclined to believe the explanation of brundlefly than yours, because every sound-chip, like emuk1 or ac97 have an analog mixer section, without the need to pass trough the AD/DA section, its based on  IBM mixer schematics or something like that, there is an analog codec, and a digital controller , so, if what we want its to monitor the analog input(so it can pass directly to the analog output) there is no need for a AD/DA conversion.

RME are not layers, and what they are saying its correct, its just you that misinterpreted their words...
Of course if you active the input analog signal (lets say your vocals)and already you are playing some audio from your DAW, you are summing or mixing the playback signal with the input signal(analog input + DA output), not a big deal...

I'm more inclined to believe the explanation of brundlefly than yours, because every sound-chip, like emuk1 or ac97 have an analog mixer section, without the need to pass trough the AD/DA section, its based on IBM mixer schematics or something like that, there is an analog codec, and a digital controller , so, if what we want its to monitor the analog input(so it can pass directly to the analog output) there is no need for a AD/DA conversion.

 
Carl, it appears I have led you astray, just as the ASIO "Direct Monitoring" literature led me astray with phrases like "True Direct Monitoring" and "Zero-latency". I just did an emipirical test of my E-MU 1820m, and it's clear that the monitored signal passes through the A/D/A converters, because there is a delay in the monitored signal of about 2ms, which is typical for a round of A/D/A.
 
So it's not direct, and it's not zero-latency. At best it might be called "less-indirect monitoring", and "lower-latency" (about 1/3 to 1/2 of what you can get making the full round-trip through the host with 1-2ms buffers).
 
Koolbass, sorry for making you struggle to explain it, and thanks for inspiring me to look deeper. None of the references I found, including my own PatchMix manual make it really clear that the monitoring is after the converters. I could have guessed that since for some of the features to work without it would require having a pile of analog mixing circuits in the interface. The 1820m dock is big enough that I could imagine those circuits being there, but I should have known the "direct" monitoring goes through the converters, because the 1820m also has hardware DSP that necessitates that.
 
So, I was wrong, but anyone calling this "direct" or "zero-latency" monitoring is fibbing slightly.
 
 
 

Yes, unless your soundcard has an analog mixer built in, then ALL sound has to pass through AD/DA converters. 

As far as I'm aware that's pretty much all soundcards, because soundcards employ digital mixing.   If you don't want to go through converters, then get a little small analog mixer to go with it.

brundlefly
 

just did an emipirical test of my E-MU 1820m, and it's clear that the monitored signal passes through the A/D/A converters, because there is a delay in the monitored signal of about 2ms, which is typical for a round of A/D/A.


This post its about ADM, and not about the method that  particular manufacturer employs to achieve monitoring outside the ASIO host, neither the effective latency that their audio device may give, because each manufacturer can be a different world, as you can see, the  E-MUs interfaces does have DSP, so obviously passes trough the converters like any other digital effects processor.

Kroneborge


Yes, unless your soundcard has an analog mixer built in, then ALL sound has to pass through AD/DA converters. 

As far as I'm aware that's pretty much all soundcards, because soundcards employ digital mixing.   If you don't want to go through converters, then get a little small analog mixer to go with it.

And yeah, most sound-card have analog mixers built in int, then not ALL the sound have to pass trough the converters , at certain point an be switched to the analog output, and if you are not aware of that, or have any doubt, just see this:

http://www.analog.com/sta...data_sheets/AD1888.pdf



SEE THE DIAGRAM AND LATER FOUND THIS SECTION (PAG 3)

ANALOG MIXER—INPUT GAIN/AMPLIFIERS/ATTENUATORS
Signal-to-Noise Ratio (SNR)
CD to LINE_OUT
90
dB
LINE, AUX, or PHONE to LINE_OUT1
90
dB
MIC1 or MIC2 (Note: MIC Gain of 0 dB) to LINE_OUT1
90
dB
Step Size All Mixer Inputs
1.5
dB
Input Gain/Attenuation Range: All Mixer Inputs
46.5
dB

Thanks to this thread, I now do understand ASIO DM.
My profire 2626 does it but I cannot control it from Sonar.  The software mixer that comes with the 2626 allows for up to 8 seperate headphone mixes, insertion of outboard analog and digital FX, all without latency.

I didn't mean to highjack the thread, and I admit I knew nothing about ADM before this thread.

Thanks too koolbass, I realize there is a difference between DIRECT monitoring thru an interface, and ASIO DM which does convert the signal to digital without running it thru the DAW and the PC's processor (right?)

I would LOVE it if Sonar could access this directly from it's tracks because it would allow (presumably) the use of a control surface as opposed to using the mouse on the profire's software mixer/router. 

I actually like the mixer/router software for the profire  (I have it on it's own monitor), I don't mind it being outside sonar, I just wish that I could control it with my MCU or C4 or SOMETHING.

carlosagm79

BlixYZ



Thanks too koolbass, I realize there is a difference between DIRECT monitoring thru an interface, and ASIO DM which does convert the signal to digital without running it thru the DAW and the PC's processor (right?)





There must be no a difference between your interface DIRECT monitoring, and the monitoring that ADM can give you, because the protocol, from inside... lets say Nuendo, it's just telling the interface that start its own monitoring system from outside the ASIO host, for God sake its so hard to understand?!

Gee, as this thread continues, your fonts get bigger and bigger.  Calm down -- it's a discussion not a life threatening situation LOL.
 
 

Did anyone bother to read my first post #9 where this is all explained very clearly by Line 6?

Line 6 calls it Tone Direct Monitoring, but it's essentially the same thing!

And yes, the signal has to go through the A/D converters, it just doesn't go through the DAW. It is routed from the interface's processing program/ mixer/ whatever the heck you want to call it, to your outputs. That is why there's less latency although not really 0 (Zero, Cero). It's latency that's so low that it is imperceptible to the human ear.

Carlos,

That has been discussed ad nauseam.  What's your point?

ba_midi

pinguinotuerto


Carlos,

That has been discussed ad nauseam.  What's your point?

You realize it might take 2ms for someone to get that

Lol!!!!!!!!!!!!!!!

carlosagm79

pinguinotuerto

ba_midi

pinguinotuerto


Carlos,

That has been discussed ad nauseam.  What's your point?

You realize it might take 2ms for someone to get that

Lol!!!!!!!!!!!!!!!

yeah I have to laugh...

brundlefly


James, you need to do a little research on ASIO Direct Monitoring. It's a standard tha allows an ASIO host to enable hardware monitoring in an interface that supports it. This bypasses The A/D/A converters and is true, zero-latency direct monitoring. SONAR does not support it. SONAR's CTO, Noel Borthwick has said this is becasue the standard is not well-defined.

I guess that is why no other company but Steinberg have managed to implement ADM... Oh wait! I know of at least 7 companies that are not Steinberg that have implemented ADM... I guess Noel is just too shy to send an email to Steinberg to ask clarification for those 5 parameters he can't figure out... Surely he could contact the techs at Edirol that managed to implement ADM in their interfaces, no?

typedef struct ASIOInputMonitor
{
    long input;        // this input was set to monitor (or off), -1: all
    long output;    // suggested output for monitoring the input (if so)
    long gain;        // suggested gain, ranging 0 - 0x7fffffffL (-inf to +12 dB)
    ASIOBool state;    // ASIOTrue => on, ASIOFalse => off
    long pan;        // suggested pan, 0 => all left, 0x7fffffff => right
} ASIOInputMonitor;

Looks very complicated. I'm wondering which part exactly is not well defined...

Some more stuff from the ASIO SDK:

kAsioSetInputMonitor

Purpose:
Set the direct input monitoring state.

Parameter:
params pointer to ASIOInputMonitor structure.

Returns:
ASE_SUCCESS if request is accepted or otherwise ASE_NotPresent

Note:
If the hardware does not support patching and mixing a straight 1 to 1 routing is suggested. The
driver should ignore all the information of ASIOInputMonitor it cannot deal with, usually
these might be either or all of output, gain, pan.
Output is the base channel of a stereo channel pair, i.e. output is always an even channel
(0,2,4...). If an odd input channel should be monitored and no panning or output routing can be
applied, the driver has to use the next higher output (imply a hard right pan).

UnderTow

Samplitude
Sequoia 
Reaper
Cubase
Nuendo
Wavelab
FL Studio
Pro Tools
PreSonus Studio One(just prop hardw)

and maybe other applications support ADM

ADM is the first thing on my wishlist for sonar since i changed from nuendo