Integrated Cross Delay

[gdougherty]

http://www.prosoundweb.com/article/innovason_takes_the_lid_off_pandora/

Pretty cool idea.

[Jeff Scott]

OK...so they seem to be saying that no matter where you are sitting in the audience , you will perceive true stereo placement?

Did i read that right? Sounds like marketing hype to me. How can that be physically possible?

[gdougherty]

No, it's cross delay panning. The signal is simply repeated on the opposite side but at a delay. You get big stereo in the middle and mono still at the extremes, but you're not missing what's been panned to the other side.

[badgerman]

Someone mentioned this before. Is there a way to do it in SAC now?

[MikeDee]

Perhaps I'm misunderstanding the big picture here, but I thought Bob has been doing this for quite some time with his native Echo/Delay plugin + panning (which would mean that, yes, it can be done in SAC and SAWStudio). :confused:

My apologies if my observation is inaccurate and/or incorrect.

[KUI]

You are correct, Bob was there first by about 15 years or so!
He is almost never rewarded in the press for his ground breaking innovations.

KUI

[Paul Henry]

Anybody have a post link or step-by-step? What sort of delay times are we talking, like 1ms head sized delays or distance between the stacks 50' delays?
Phase reversal?

[Jeff Scott]

I've done this (and continue to do this..) with Bob's Delay. 18 to 25ms is required.

[Paul Henry]

I've done this (and continue to do this..) with Bob's Delay. 18 to 25ms is required.

So just put on a 25ms delay and swap the channels, that's it?

[soundchicken]

yep, pre-fader the effect and hit button in the lower right from normal to cross.

[soundchicken]

http://www.prosoundweb.com/article/innovason_takes_the_lid_off_pandora/

Pretty cool idea.

...and they don't even have the comments turned on for the article.

[gdougherty]

He may have been, but I'm not sure Bob was the first to be doing this.

[Bob L]

You have to pst patch my echo/delay and then use the cross delay option... pan the chan itself left or right and raise that side of the plugin... balance out the source and delayed signal... anywhere from 15 to 25 ms ... many times I use partial panning and set each delay side slightly different.

You can see a demo of this in the SAWStudio demo videos of the band recording, editing and mixing project.

Bob L

[TomyN]

well, I talked to Carsten on the PLS. It's not only delay (he states that delay is the smaller part) but delay and filter.
Because he used SATlive for development I get some more information and I showed him SAC, and he seems to like it.

So, there might be a plug.

Tomy

[ckuemmel]

Hi, as Tomy said. The delay is only one part, because it would not work in big relations. What happens if I would put 20ms on it. Besides the fact that people in the middle would already hear an echo in high frequencies I would create a combfilter starting from the very lowest frequency. It is not as simple :-) Sorry :):):)

[Yogi]

Yes it is as "simple". Comb filtering is actually worse with two mono sources spread across a venue. By delaying one side by a specific amount the actual comb is spread by frequency across the span of the speakers. ie. the filtering points that would take effect at say 1K would be totally different than those for 2K and not just harmonics. I've done the math and cross delays actually make any comb filtering disappear (from a perception point due to psychoacoustics). The net result is that people in the middle swear they are hearing stereo (and full rich stereo at that). The people not in the sweet spot hear varying degrees of stereo but always hear full signal from a volume perspective of all channels. The sweet spot is REALLY wide with this. In our church perceived stereo runs almost 75% of the chairs. Try getting that with REAL stereo. If you really want the full effect put on a set of headphones that are monitoring the mains. It is truly uncanny.

[gdougherty]

Hi, as Tomy said. The delay is only one part, because it would not work in big relations. What happens if I would put 20ms on it. Besides the fact that people in the middle would already hear an echo in high frequencies I would create a combfilter starting from the very lowest frequency. It is not as simple :-) Sorry :):):)

As Bob shows in the video, the blend of the crossed delay seems a crucial component. I know I've been unhappy with the results (especially on vocals) just dropping the delay on a channel, adjusting times and running with it. I need to spend some time playing with the SAW demo and the delay to see what really works well. The intuitive answer to me is that you want near zero of the delayed signal on the side you're panning towards and a pretty strong delayed signal on the opposite side. It also seems like it'd work best with hard panned signals.

[Yogi]

Hard panned is the only way I do it with zero on one side and the delay on the other. And no two channels set the same for panning or delay(panned either right or left). We tend to have 4 to 6 vocalists at any one time and it really sounds full and rich with that BUT even solos sound great. I've had people listen to our tapeit sessions of our worship sets and swear it was taken from a CD. If anyone would like to hear them send me a request.

[gdougherty]

Hard panned is the only way I do it with zero on one side and the delay on the other. And no two channels set the same for panning or delay(panned either right or left). We tend to have 4 to 6 vocalists at any one time and it really sounds full and rich with that BUT even solos sound great. I've had people listen to our tapeit sessions of our worship sets and swear it was taken from a CD. If anyone would like to hear them send me a request.

If you hard pan only and don't set any the same, then do you vary the apparent panning with the delay times?

[Paul Henry]

Hi, as Tomy said. The delay is only one part, because it would not work in big relations. What happens if I would put 20ms on it. Besides the fact that people in the middle would already hear an echo in high frequencies I would create a combfilter starting from the very lowest frequency. It is not as simple :-) Sorry :):):)

Hi, thanks for joining the conversation, and welcome to the Forum!

[Yogi]

Here's the way I do it, YMMV. Say I have six channels of vocals. Chan 1 pan hard left, chan 2 hard right, chan 3 hard left, ... Each channel is one or the other. At the same time I set the cross on the delay on both the right and the left of the delay to the same number BUT that number is not the same as any other channel's delay, varying from about 18 to 30 ms. What you get in the left side is no delay for those panned left, and all of the panned right signals delayed by varying amounts. In the right side you get no delay for any signals panned to the right but varying delays for the signals panned to the left. With 6 vocalists they are never dead on in timing. If you listen in headphones you would swear you're listening to stereo "panning".

In fact what you are hearing is stereo delay that fools your brain into thinking the location source of the sound is somewhere that it's not. When you think about how you locate something by sound it is the delay difference of that sound striking both ears. Plug up an ear and try to figure out where sound is coming from. That's what normal panning actually does. This just happens to accentuate the delay between your ears and the sound stage gets very large. I wouldn't use anything else for vocals. I also use it on guitars, but never on bass or drums since it tends to smear both of those. If you aren't using the stereo outs of your keyboards you can use it there too. Most keyboards are already doing this to a very small degree (delays less than 3 ms with a much more complex delay structure) in their stereo outputs to widen their sound.

[JeremyJo]

I guess here is where a per channel delay built in to SAC would be nice. I will have to try and find the time try this technique out.

I still think proper LCR panning would be more useful, but perhaps they could be used in conjunction with one another.

JJ

[gdougherty]

I guess here is where a per channel delay built in to SAC would be nice. I will have to try and find the time try this technique out.

I still think proper LCR panning would be more useful, but perhaps they could be used in conjunction with one another.

JJ

This is not at all what's talked about when we're looking for per channel delay. In that case it would be a built-in time delay control that would simply be for alignment, not effects. One control, 0-20ms or something like that in .1ms steps or so.

What would be cool is a panning option that automatically does all this behind the scenes like the Innovason does.

[Jeff Scott]

I set up the cross delay on a 7 piece vocal ensemble at the church I teched for several years ago. Each mic panned as if to represent their position onstage but each had the Cross delay on it. Sounded HUGE!. A really outstanding. Choral leader didn't like it. Sounded too different from their usual Mono. They could actually hear the width from the stage.

[ckuemmel]

[quote=Yogi;180157]Yes it is as "simple". Comb filtering is actually worse with two mono sources spread across a venue. By delaying one side by a specific amount the actual comb is spread by frequency across the span ....]

No it isn***180;t. You are right, when you say that you will have combfilters anyway. But you speak about 1k or 2k. But a delay of 20ms has the first cancellation at 25Hz, then at 75Hz at 125 Hz and so on. We cannot speak about a good sounding signal anymore. Yes, people outside will hear it good, but for the people in the middle it is absolut unnatural. Besides the fact that you will hear Echos in the middle.

Thanks for beeing welcome

[Yogi]

The one big plus in doing this is it doesn't matter how close you are to a speaker you hear ALL of the mix. With LCR your proximity to any particular speaker affects the volume of certain aspects of the mix. For instance if a vocalist is more right than left if you are located closest to the left speaker you don't hear that vocalist. Regular panning drives the sweet spot to a very narrow position withing the seating area. With the delays you would always hear everything in the mix some just delayed (very minimally) no matter where you sit. The sweet spot winds up being huge. The further you are from a speaker the more pronounced the delayed sources coming from that speaker.
One other thing, if you're dealing with an installed system and have comb filtering effects try this and I'd bet you'd be amazed at how those effects go away. We also found the EQ curve changed slightly and actually got smoother. As for the examples of 1K and 2K, those were just example frequencies. The varying delays means the comb filtering for any given frequency shifts between channels. IOW, filtering points between channel 1 and 2 will not be the same as filtering points between 2 and 3 and so forth. This has the effect of reducing overall comb filtering throughout the venue. With a truly mono system the frequency points do not shift around and the effect is more prevalent. Pssst. I have physics on my side in this discussion. You can discuss it all you want but until you try it and MEASURE the results you can't understand how it works. If you're running a mono system find a point that comb filtering is especially bad (this is always easy to do with a mono system). Then split out all of the vocals in the delayed fashion described (you can do this before hand with scenes). Listen to the results. Tell me what you hear with regards to comb filtering.

[TomyN]

Hi,

well, I think the best way would be to have the C-Panning as a plugin so you could listen and compare.


Tomy

[ckuemmel]

Hi,

well, I think the best way would be to have the C-Panning as a plugin so you could listen and compare.


Tomy

www.tonmeister-online.de/video.html (http://www.tonmeister-online.de/video.html)

[ckuemmel]

[quote=Yogi;180213]The one big plus in doing this is it doesn't matter how close you are to a speaker you hear ALL of the mix. With LCR your proximity to any particular speaker affects the volume of certain aspects of the mix. For instance if a vocalist is more right than left if you are located closest to the left speaker you don't hear that vocalist. Regular panning drives the sweet spot to a very narrow position withing the seating area. With the delays you would always hear everything in the mix some just delayed (very minimally) no matter where you sit. The sweet spot winds up being huge. The further you are from a speaker the more pronounced the delayed sources coming from that speaker.quote]

That***180;s why there is Pandora.

@Yogi: Nothing against your method.
I only say that Pandora is working different and has some big advantages e.g. in a typical Mainhang (left signal) - Outfill (right signal) Situation where Pandora works as well. A system with only time difference will have strange things going on in the crossover of these.
Physics as well :)

[gdougherty]

The one big plus in doing this is it doesn't matter how close you are to a speaker you hear ALL of the mix. With LCR your proximity to any particular speaker affects the volume of certain aspects of the mix. For instance if a vocalist is more right than left if you are located closest to the left speaker you don't hear that vocalist. Regular panning drives the sweet spot to a very narrow position withing the seating area. With the delays you would always hear everything in the mix some just delayed (very minimally) no matter where you sit. The sweet spot winds up being huge. The further you are from a speaker the more pronounced the delayed sources coming from that speaker.
One other thing, if you're dealing with an installed system and have comb filtering effects try this and I'd bet you'd be amazed at how those effects go away. We also found the EQ curve changed slightly and actually got smoother. As for the examples of 1K and 2K, those were just example frequencies. The varying delays means the comb filtering for any given frequency shifts between channels. IOW, filtering points between channel 1 and 2 will not be the same as filtering points between 2 and 3 and so forth. This has the effect of reducing overall comb filtering throughout the venue. With a truly mono system the frequency points do not shift around and the effect is more prevalent. Pssst. I have physics on my side in this discussion. You can discuss it all you want but until you try it and MEASURE the results you can't understand how it works. If you're running a mono system find a point that comb filtering is especially bad (this is always easy to do with a mono system). Then split out all of the vocals in the delayed fashion described (you can do this before hand with scenes). Listen to the results. Tell me what you hear with regards to comb filtering.
Just in case you missed it, like I initially did, you're chatting with one of the gentleman who is the subject of the article. I do think he knows a little about which he speaks. ;)

[Yogi]

I still say i have physics on my side. I've been doing acoustics AND sound for over 4o years. For simplicity the cross delay works just fine without being overly complicated. The resultant sound for the majority of the audience in a left and right speaker setup WILL result in greatly reduced comb filtering. I have the math, the physics AND the acoustical testing to prove it. Bob came up with this method 15 or 20 years ago and I personally feel it's one of the best things I've ever heard. And with SAC IT'S FREE! I'd say it's the best "trick" per se for sound that I've ever seen, and I've seen a LOT of tricks.

[ckuemmel]

I do not want to make you angry and as I said, nothing against your method. It***180;s cool when people try to imrpove things. I just think that pure delay, which needs big times, has some weak points. Maybe I am wrong, maybe not, I am convinced that I am not wrong :). Other people have to think about that. But I want to ask you something. Do you mean that you reduce overall combfilters by averaging severall channels?

@gdougherty: You caught me :) Yes, I measured and tested it a lot and for me, everybody can have a different opinion, pure delay does not work.

I am new here and maybe you can answer me one question. I do not know where to look. Can I use SAC with other DAWs like Sequoia? I would kick out my console in the studio if it is possible.
Thanks in advance

[TomyN]

Hi,

:-) The DAW to use with SAC is SAW.


Tomy

[Guitarkeys.com]

Not wanting to get off topic - but if you compare SAWs audio quality with Sequoia you will want SAW.

[Craig Allen]

Not wanting to get off topic - but if you compare SAWs audio quality with Sequoia you will want SAW.

But you did. Sequoia is a fine product as is SAW.

To actually answer the question, yes, you can use SAC with other DAWs. It's more about the interface you use - you need one with multi-client ASIO drivers in order to do this. A number of folks here use Reaper to record. I use SAW because while you can use other DAWs, the integration between between SAW and SAC makes it a lot easier. I hope this helps.

[gdougherty]

I do not want to make you angry and as I said, nothing against your method. It´s cool when people try to imrpove things. I just think that pure delay, which needs big times, has some weak points. Maybe I am wrong, maybe not, I am convinced that I am not wrong :). Other people have to think about that. But I want to ask you something. Do you mean that you reduce overall combfilters by averaging severall channels?

@gdougherty: You caught me :) Yes, I measured and tested it a lot and for me, everybody can have a different opinion, pure delay does not work.

I am new here and maybe you can answer me one question. I do not know where to look. Can I use SAC with other DAWs like Sequoia? I would kick out my console in the studio if it is possible.
Thanks in advance
Ditto what Craig already said. RME an MOTU are the only confirmed interfaces with multi client drivers. If you've got something different just fire up two ASIO apps and see if they'll work at the same time

BTW, welcome and kudos for your work with Innovason. It's a nice board and that's a sweet upgrade.

[ckuemmel]

Thank you for your hints. I have RME and will check it out the next days.

[Wurst Werner]

@Carsten: Any chance on a C-Panning VST Plugin? Lawo has some connection to brainworx...so that would be nice to have:cool:

[TomyN]

What about a SAC plug-in? :-)

Tomy

[Wurst Werner]

What about a SAC plug-in? :-)

Tomy

Even better!

[gdougherty]

Frankly, until Bob bothers to publish an SDK for developers with the remote control included, I'd rather not see a native plugin. VST's at least have some level of remote control. Native plugins have none with the exception of the 4 Bob has updated.

[ckuemmel]

I will speak with Lawo about that. Right in the moment there are no plans. It is too new

[Cary B. Cornett]

For years now, I have wished that somebody would make a panning control that controls both level balance and inter-channel delay times. My original thought about this was that center panning would have exactly the same result as now, equal level to both sides and no delay. As you pan further left, level drops on the right AND the feed to the right is increasingly delayed.

It may be that part of the Innovason panning setup is as I described, I don't know.

This discussion, though, has suggested a potential weakness in my idea: Delay times that are ideal for localization with headphones might not work well for speakers, and the optimal delay times for the speakers might be dependent on the distance between speakers, so a range of inter-channel delay times that works in a studio control room might be all wrong for a wide concert stage.

I also get the sense that the cross-delay panning trick may work differently from what I originally wished for, and, to make things more confusing, not everyone may have the same idea of how it is supposed to be done, so that a difference of opinion about how well it works may spring from a difference in the way it is applied.

To give you an idea of how easily this confusion can happen, when I first heard about applying this trick in a live sound setting, I assumed that it was a simple matter of inserting the SAW delay in the master FOH output channel pair, set for cross delay, with source and delay at equal levels, and a delay time about 5 ms or so greater than the acoustical delay between speaker stacks.

My next thought was that this was a bad idea for center panned sources (slap on everything), so there should be a submix with the cross delay in it for all "off center" sources, and everything that is dead center kept from going through the cross delay at all, just fed dry.

Now, in this discussion, it sounds as though it is necessary for each input channel to have its own cross delay inserted, with delay times and dry/wet balance trimmed for each source. The complication involved in doing that makes me wish that it could be integrated "under the hood" in the pan control, with a selectable option of not using the delay. Then again, maybe that gets us back to the "one size fits all" delay problem. :rolleyes: :mad:

Or are all of these possible methods potentially valid, depending on the situation? :confused:

[RBIngraham]

One thing I've been somewhat reluctant to add (yeah I know... me.. reluctant? ) is that it seems to me if you're dealing with a good system installation where you might have left, right, a front fill, underbalc delays, maybe some upper balc delays, maybe you have a center cluster don't know, plus subs or maybe royalty box delays, etc, etc, etc.... Usually that system would be all time aligned to the system is aligned to each other as much as is possible.

Throwing something like a cross delay panning into that mix I think would be a recipe for disaster and could easily cause more harm than good. I think this technique is definitely for those that are only doing a simple L-R house system.

I use time delays all the time, but it's usually to allign the system with the live source (ie. the actor or musician) on stage in the upstage-downstage plane, not the left to right.

Not trying to take anything away from this technique, just pointing out that in a complex sound system, this may not work all that well. Or at least in order to make it work well, you would have to spend A LOT of time tinkering around with multiple delays times to each system in order for it to work, more than likely.

[ckuemmel]

You are right. It would bring problems. That is one point why the delay is the minor part in Innovason´s Pandora

[Yogi]

This technique is only meant for L-R, but in that environment it works VERY well. Much better than stereo panning. The sweet spot is huge and every channel is mixed into BOTH speakers so the mix is exactly the same for both sides. However if you are in the middle (and middle in this scenario is quite large...probably 65 - 75% of the audience) then it sounds really really good.

[RBIngraham]

This technique is only meant for L-R, but in that environment it works VERY well. Much better than stereo panning. The sweet spot is huge and every channel is mixed into BOTH speakers so the mix is exactly the same for both sides. However if you are in the middle (and middle in this scenario is quite large...probably 65 - 75% of the audience) then it sounds really really good.

Thanks for the info. I so rarely work with just stereo this just isn't for me I guess. Which is what I expected. If I was going to play those kind of games I would probably have to go with something like track the actors. Otherwise I ussually have a mono vocal reinforcement system with a seperate stereo system just for orchesta and the fills used as needed depending on many variables.

[Yogi]

Yep Richard, I couldn't see it being of much use in your environments. When we have someone that is just speaking (announcements and the sermon) they are strictly mono with almost no processing except light compression. This technique is primarily for groups who normally use two stacks left and right but run mono (or stereo with minimal panning) to both stacks. In that environment the difference is astounding.