Headphone Cue Mix in a digital environment

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Rochey

Well-known member
White Market Member
Joined
Jul 2, 2004
Messages
2,697
Location
Dallas, TX - Formerly UK
Dumb Q for you boys 'n girls.

I have up to 24 channels of outputs available to me - 8 of while are lower quality.
I was thinking of mapping these to dedicated headphone mixes (Cue Mixes) for the talent.
DAW's such as cubase should allow me to do aux mixes of this kind.

However, there is delay (latency) involved in the foldback of the instrument to headphones in that case. (e.g. the mic has to go through AD - Processing - DA before making it back to the musician).

Obviously, this is a common problem. My initial thought to resolve this would be to mix a little of the microphone input directly with the DA output, giving you live monitoring on top of DAW mix output. As far as I know, this is already quite popular.

What happens when you have voice effects? if you mixed the DAW output and Live-Mix, then you could face strange phase issues, right?

What do folks do in the real world?

Cheers

Rochey
 
Pro Tools HD/TDM provides HW level monitoring on their DSP cards. The plugins and protools mixer run on the DSP cards so the buffer settings are not part of that equation.

I have used MOTU hardware with their CueMix software which provides low-latency hardware monitoring (mixing any input to any output). Some of the MFG's provide HW based plugins. UAD's Apollo apparently does this with UA plugins.

At work I use a PT HD system with 16 channels of outputs which feed an aviom-mixing / headphone-mixer platform. As long as I disable Delay Compensation in PT it's fine for multitrack recording. I also use plugins with the lowest reported latency on channelstrips in these situations.

For a native solution you really need to run your computer at the lowest buffer settings possible. That can work well in many cases if you have a screaming machine.

This is the worst part of DAW recording IMHO. I prefer tape sends on the left side of the console, and tape returns on the right. AUX buses to HP/Cue system fed from the tape returns so that with the tape machine in input/sel-rep they hear whatever they are supposed to hear. DAW world is a little different.

Hope that helps!

Cheers,
jb
 
While I haven't personally investigated this, it is an issue for bands using in ear monitors. 

If you mix dry signal with same delayed signal you can get flanging (comb filtering) if delay is significant wrt wavelengths of the audio content.

While I didn't participate myself, some friends who were checking out a new digital console, used it to mix in ear monitors for a band at a gig, and reportedly did not get complaints. The console (Behringer X32) specifies propagation delay of less than 1mSec for console input to output, and less than 2mSec for stage box to console to stage box...

http://www.lsbaudio.com/publications/AES_Latency.pdf

This AES paper suggests that percentage of happy campers on in-ear monitors listening to vocals  is already deteriorated 50% by 2 mSec. So shorter is better. 

Other than vocals (listening to yourself) can tolerate longer delays.

JR
 
Agreed and agreed to the above. 

Rochey said:
What happens when you have voice effects? if you mixed the DAW output and Live-Mix, then you could face strange phase issues, right?

If, when you say voice effects, you mean time based effects then you may very well be fine.  You can look at it as a little extra predelay on reverb, extra delay on delays, etc. 

The only part of JR's comments that I don't fully agree with is
JohnRoberts said:
Other than vocals (listening to yourself) can tolerate longer delays.

Great musicians will and DO feel the un-resonsiveness of their instruments with miniscule delays. 

All in all we learn and adapt.

Michael
 
yup, I met one of them once.. a couple decades ago. he was a (good) bass player and the smpte to midi locked drum machine we were using made him crazy with random midi timing errors. (Problem wasn't the drum machine but that's another story I won't bore you with).

In ear monitors are a relatively new thing, while headphones are not, but monitor wedges have always introduced a few mSec of delay due to the air path.

That AES paper I linked to describes delay thresholds for other instruments.

If Behringer can do a digital console with <1 mSec latency that seems a fair benchmark.

JR
 
Rochey said:
My initial thought to resolve this would be to mix a little of the microphone input directly with the DA output, giving you live monitoring on top of DAW mix output. As far as I know, this is already quite popular.
I've tried that; the comb filtering resulting from the mix of direct and delayed signal) is more disturbing than the simple basic delay. The only solution is either to minimize delay by decreasing the buffer size, or using hybrid monitoring, i.e. live tracks going directly analog into the cue system and pre-recorded tracks going through the DAW. This can be very tricky when doing punch-ins.
What happens when you have voice effects? if you mixed the DAW output and Live-Mix, then you could face strange phase issues, right?
Having effects does not change the fundamental delay problem; it may add a couple milliseconds.
What do folks do in the real world?
I use "tape-type" monitoring with low buffers (224 at 88.2 k SR, about 2.5ms).
 
JohnRoberts said:
This AES paper suggests that percentage of happy campers on in-ear monitors listening to vocals  is already deteriorated 50% by 2 mSec.
Funny thing is the same people who complain about 2 msec delay in their IEM's don't say a word about their ears standing 5ft away from their floor monitors, which accounts for 5 msec.
Other than vocals (listening to yourself) can tolerate longer delays.
For different reasons, drummers and percussionists are very sensitive to delay. They hear it as an echo, they notice it, but most of them can live with it, probably because they're used to discriminating so many rhythm elements. Singers have the problem of dual-path; the electrical path and the bone conduction path. This leads to uncomfortable phasing effects, with the apparent sound source shifting from outside to inside.
 
Yes, percussive or transient sounds have very specific timing information that we can lock onto.. (insert Hass et all phenomenon here).

I have been paying attention to this recently in the context of digital consoles in combination with in-ear monitors for live use. To determine whether 1-2 mSec latency was short enough.  Early reports suggest it is , but some musos may be more sensitive than others.

My concern was the added 2 mSec to a floor wedge's 4-6 mSec would be unnoticed, while 2 mSec vs, 0 mSec on IEM would be more apparent. 

Yes longer delays are obviously problematic for perhaps a different reason than combing (actual perception of lag or delay).

JR

PS: I don't know if the internal bone conduction path is more of an issue when working with in-ears or cans. It seems SPL at the ears could be run significantly lower than when using wedges on a loud stage, so lower level from the cans, means bone conduction path seems louder......  I guess it has always been that way tracking in the studio on cans. 
 
Tracking in the studio is often done without cans. This is nothing new and has been done thois way for a long time. Especially for live to tape/disk basic tracks or whatever... There may also be a PA or monitors in that situation as well as headphones.

One could say that live off the floor headphone-less recording/tracking sessions also incur latencies since the talent are obviously not going to be spaced 0ms from each-other.... Add in some headphones and close micing and you've got closer to 0ms mixed with latent signals of perhaps 10-20ms depending on the size of the studio... ie; one ear off, one ear on <the headphone>.

Also, from each person's perspective the acoustic delays are different because they hear themselves sooner than others.

Stage situation may be very similar.

I remember thinking about this when doing mixing/tracking in all analog studios with significant outboard patched in and tweaked... The feeling that everything was absolutely in sync and resonating along with the music and musicians... A definite unique moment in time never to be revisited... The feeling that all the signals and related performances were very near instantaneous was quite exhilirating. There was no latency.

We have moved on from there.

Yes. Latency matters. The less of it the better IMO.

I used a TC2290 delay once in a test to prove to myself the importance of sub-millisecond delays. That unit could be used to set delays on the sub millisecond levels.

Mult and pan the signal hard L/R. Delay one side or both sides such that they are equal. Place yourself in the center/sweet spot and start tweaking the delay in the smallest increment that you are able to on one side. Observe the signal move across the stereo field and make note of how small the delay is. 10th's of an ms? 100th's of an ms? I don't honestly remember now but it was certainly significantly less than a millisecond.

A millisecond is so far from instantaneous that Stevie Wonder could drive a truck through that hole.

Cheers,
j
 
I have some concern that utilizing digital consoles, and then using plug-in processing within the console is causing even more delay. I work with a band quite often, and we recently did a show at a venue with a very nice system. However, the in-ear monitor sound was causing problems for the lead vocalist.

I know for a fact that plugins were used to process audio in the monitor console. So, now we have the console delay, but now we also have a plugin delay.

Hmm...

Add to the fact that the lead singer prefers to also have a stage wedge present, now we have a multi-path zoo!
 
IMO there are two different issues... #1 pure delay, and #2 relative path delay between two paths that get combined.

I am mainly concerned about people listening to themselves with 2 paths present.

Short delays altering localization is also well studied in AES journal papers. Of course panning with delay instead of level has different consequences for mono compatibility.

JR
 
JohnRoberts said:
  It seems SPL at the ears could be run significantly lower than when using wedges on a loud stage, so lower level from the cans, means bone conduction path seems louder......  I guess it has always been that way tracking in the studio on cans.
I'm not sure...just the other day I read athread on another forum where the guy, complained about having not enough level in his cans. A quick calculation showed that his headphone system ran out of steam at 117dBspl ! He wanted more...
He was a percussionist, not a singer, but I see most of the singers dialing insane levels in their cans. I think at least, with monitor speakers, they are alerted by the physical impact of sound and the build-up of distortion. With cans and IEM's, they have no warning.
 
abbey road d enfer said:
JohnRoberts said:
  It seems SPL at the ears could be run significantly lower than when using wedges on a loud stage, so lower level from the cans, means bone conduction path seems louder......  I guess it has always been that way tracking in the studio on cans.
I'm not sure...just the other day I read athread on another forum where the guy, complained about having not enough level in his cans. A quick calculation showed that his headphone system ran out of steam at 117dBspl ! He wanted more...
He was a percussionist, not a singer, but I see most of the singers dialing insane levels in their cans. I think at least, with monitor speakers, they are alerted by the physical impact of sound and the build-up of distortion. With cans and IEM's, they have no warning.

It may just be a drummer thing... I recall one time a few decades ago I was in a studio and between takes decided to listen to the drummer's headphone mix, I got the cans about a foot from my head and thought better of it...  That actually inspired me to design a little accessory headphone amp (HB-1) for AMR where I drove the two legs opposite polarity,, (took 3 amplifiers to make differential stereo).  This roughly 2x signal swing made 4x the power of similar voltage single legged amps.

Many drummers are probably half deaf from years of cymbal wash and whatever.

JR
 
I've recorded a lot of cathedral choirs & organists.  The organist usually sees the conductor through a mirror or CCTV.

But in some places, he might have to play 0.2 second ahead so his contribution arrives at the singers at the right time.[1]  And he may hear the choir 0.5 sec after what he plays and what he plays at a different time too!  Incredible.

Cathedral organists are some of the most remarkable musicians I've ever known.  They are among the very few classical trained musicians who can improvise on demand and many do this on a daily basis.

[1] Singers are completely thrown if they hear themselves (or their accompaniment) even slightly out of time.
 
I had an interesting time figuring out how to get reverb in the headphones without monitoring the dry signal thru logic using my Symphony.  What I had to do was use Maestro's zero-latency mixer for the dry signal.  In Logic, turn on Software Monitoring, then input monitor the audio track so signal shows up in Logic.  Then, on the audio track, set up a send to my reverb of choice, and set the audio track's output to a Bus.  Mute the bus.  So, logic is still listening to the audio input and sending it to the reverb as normal, but the dry signal that's showing up in logic isn't being sent to the mains.  My headphones still have the dry signal with no latency from Apogee's Maestro interface, and also the reverb from logic.  This also lets me run the buffer settings pretty high so I don't get clicks and pops (256, 512samples...)
 
Yes, low-latency monitoring, as provided by some soundcards is the next best thing to analog hardware monitoring in terms of latency. There is still the conversion latency and a few samples in the DSP execution.
The nice thing with Maestro or similar apps, such as MOTU's CueMix, is that it is well integrated in the DAW, doing automatically the complex switching necessary when doing punch-ins.
 
Interesting topic. Is it the general consensus that latency is more of a problem while using in-ears instead of traditional head phones? Wonder why....................drivers actually closer to the ear canal? Perhaps there is also a difference when using closed-back/isolation headphones as opposed to open back ones? Hmmm.
 
Spiritworks said:
Interesting topic. Is it the general consensus that latency is more of a problem while using in-ears instead of traditional head phones? Wonder why....................drivers actually closer to the ear canal? Perhaps there is also a difference when using closed-back/isolation headphones as opposed to open back ones? Hmmm.

I don't know that it's a new problem, or that in-ears are more problematic than cans, only that in-ears and digital consoles are only recently becoming more prevalent in live sound mixing, where I pay attention.. I have heard stories of groups carrying a separate analog mixer for IEM, used in combination with a digital console for FOH. While it seems possible to execute a digital console with less than 1mSec latency (ignoring plugins).

JR
 
mulletchuck said:
The nice thing with Maestro or similar apps, such as MOTU's CueMix, is that it is well integrated in the DAW, doing automatically the complex switching necessary when doing punch-ins.

huh?
When doing punch-ins, you need to switch the monitoring source from PB to direct. A tape machine does it by switching from the sync amp to the input, which is latency-free. A DAW switches from virtual PB to input AFTER aggregation in packets (record buffers), a process which is done in the computer and submitted to latency. Some hardware permits almost latency-free input monitoring by switching from virtual PB to input right after conversion. This is done in DSP in the souncard. The resulting latency is just the addition of conversion times in the A/D and D/A.
Real latency-free monitoring can be achieved by using an analog mixer, which will provide switching between the PB from the DAW and the analog input signal. If you want to do punch-ins on 8 or 10 tracks of drums, you must find a way to action 8 or 10 switches simultaneously. Some mixers offer this possibility, most do not.
 
Back
Top