Can someone explain the 21ms delay in the mixer presets?

[Jason T. Shannon]

Hello there! In the Synchron Player mixer (all instruments, presets), there is a 21ms delay programmed into all of the Decca Tree and Surround microphones. The close/mid don't have a delay. I'd love to understand what the 21ms is supposed to represent. At first I though maybe it was a pre-delay, but in that case I'd think that the number should vary for each instrument grouping (a larger number for the instruments in the back of the hall?). I understand that these mixer presets are just starting points, but some clarification on that would be really helpful, thank you! 

Regards,
Jason Todd Shannon

[Pixelpoet1985]

I also asked about the mixer presets several times and never got an answer. Don't understand. What's the problem in answering a simple question? I find it very difficult to understand what's going with the presets, they are very different to other libraries on the market.

For me this delay is more like a pre-delay of a reverb. They treat the ambient mics like a reverb and focus on the close mics. I spent a long time with the delay settings and still have no clue what setting should be right. I'm not that happy with the mixer presets.

I also don't know if the close mics have a delay built in or not. If not: the close mics should have a delay. And even close and mids should have slightly different settings, because they are not at the same position. I experimented with this a lot and found out that this also reduces the phasing issue with the vibrato crossfades in the strings (which is annoying and doesn't get fixed, I mean this a no-go considering the price tag, no other library has such severe phasing issues).

In general they use three numbers: 15ms, 21ms, 42ms. I first thought that these are the distances of the close microphones, but this would mean only three positions from the decca tree. Every close mic needs a number on its own. And only VSL can measure and tell us these numbers.

Would be really nice if someone from VSL could finally tell us a bit about this. In my opinion there is a huge difference in sound when using various delay settings. I know the saying is "If it sounds good, it sounds good", but it does not sound good, for my part. 

[PaoloT]

I too think that some more words on the mix presets in the manuals would be very useful. I’m trying to reverse-engineer the presets to try to understand how they work.

As far as I understand, all the mics are time-aligned. So, there is no need of delay to align them. The delay values seems to be offsets, and used to increase the perceived distance of the instruments from the Main and Ambient mics. The numbers can be the same for the violins and the brass, since they are just an offset, and not the actual value.

The values they used are in milliseconds, so 15, 21 and 42 ms are equivalent to 5, 7 and 14 meters of added distance.

I've not yet tried, but I guess one can delay the violins so much as to move them behind the brass — if these latter are not delayed. This would open incredible possibilities. I don't know, however, if it would still sound "natural".

Paolo

[Air]

Hi,

Yeah Paolo got it right! The mics are all time aligned if the Delay is turned off or set to 0. It is a runtime offset like you would have it in a real recording. You can play around with it. 

As also said before the presets are just starting points. If you wanna have a time aligned sound turn them off. If want to move the tree or the room mics away from your close mics or mids use the delay (3ms = 1 meter). If you want the room playing a 16th note or what ever after your close mics, go for it.  

Hope that helps

Best

Bernd

  

[Pixelpoet1985]

@Air said:

Hi,

Yeah Paolo got it right! The mics are all time aligned if the Delay is turned off or set to 0. It is a runtime offset like you would have it in a real recording. You can play around with it. 

As also said before the presets are just starting points. If you wanna have a time aligned sound turn them off. If want to move the tree or the room mics away from your close mics or mids use the delay (3ms = 1 meter). If you want the room playing a 16th note or what ever after your close mics, go for it.  

Hope that helps

Best

Bernd

  

Many thanks, Bernd, for clarifying.

I have two further questions:

• In which situation would one delay the main mics? You gave an example, but I never heard of this approach. I didn't know that this is a common thing in recording.
• All of the strings presets have the room mix or the decca tree panned by about 40 to the left or right. Never heard of panning the tree... What's the point here? Because of lacking outriggers? Because of the mics you chose for the tree? I mean if I want a "real" recording situation I don't have different pannings, I have one tree setup for the entire orchestra (unless they are recorded separately like the samples).

[Air]

Hi,

The main mics would be delayed by itself in a real recording as the distance to the players is larger then from the close or mid mics. So the signal is reaching the room mics later.

 

Which preset and library are you talking about than I can have a look. In the power the pan itself should be left - right but maybe the balance was a little changed so you could hear more from the violins.

 

Best

Bernd

[Pixelpoet1985]

@Another User said:

Which preset and library are you talking about than I can have a look. In the power the pan itself should be left - right but maybe the balance was a little changed so you could hear more from the violins.

For example: Elite Strings. Every preset has the balance changed. So you answered the question already. Was just wondering why. So, in a "normal" recording situation the balance should be in the center.

[Air]

HI,

For time alignment it can be mad in both directions depending of what is important. If you want  be really precise in the sound it is better to have close and mids before the rooms. 

For samples everything is time aligned to zero so you do not have any delay when playing with the keyboard. So to go back to real recordings you have to move the tree back again.

If you like to move the close and mids back, try it.

Best

Bernd 

[Pixelpoet1985]

@Air said:

For samples everything is time aligned to zero so you do not have any delay when playing with the keyboard. So to go back to real recordings you have to move the tree back again.

I'm not really fond of that keyboard delay thing. At least this explains why the sampled legato is really short (even compared to the silent stage). Wouldn't it be "better" if there was some delay instead of zero? In my opinion, there has to be some tiny bit while playing with the keyboard. The silent stage samples have some delay. But that's another topic. 😄 But to be honest: That's false advertising. If VSL claims to be time-aligned, than it should be "correct" and not zero. This is mentioned nowhere.

Sorry for asking again: I'm not quite sure if I understood correctly. Time-alignment with Synchron means: Everything is at zero and NOT to the "correct" position?

If everything is at zero and the tree is put back again to match reality, why is the mid and the main center microphone not delayed? 

My thinking with the new information would be:
– Close: 0ms
– Mid: e.g. 3ms ?
– Tree Center: 21ms ?
– Tree Stereo: More than 21ms?
– Heighs and Surrounds: Even more ?

In my opinion it would be nice if there was a mixer preset which would really resemble reality. Maybe it's too much asked, but, as I said earlier, there is a huge difference in sound with the delays we now have in the presets. I spent a lot of time comparing to other libraries. Why not make a mixer preset with the "correct" delays? I think this would also be a huge thing for other users (who also complain about the sound). And as I also said before: this reduces the phasing issue with the vibrato crossfade in the strings.

[PaoloT]

@Pixelpoet1985 said:

That's false advertising. If VSL claims to be time-aligned, than it should be "correct" and not zero. This is mentioned nowhere.

In my humble opinion, it is always a good idea to learn about something before accusing someone of giving false information.

Time-aligned means that all the channels start to play in sync. This wouldn't be possible without time-alignment, for the hard reality that sound takes more time to reach distant positions. Having to do it live, one would do this by adding delay to the closer microphones, so that they are all heard at the same time. In recording, you print this delay in the recorded track.

If you want the same situation as in the original recording, just add delay to the mics, increasing it in order of distance. This is hardly what is intended to be ideal in actual use, for the bare reason that we humans hear from a single position. Our brain can add focus on 'spots', and it is on this illusion that a sound engineer plays when mixing mics, delays, reverbs.

Paolo

[PaoloT]

@Air said:

If you like to move the close and mids back, try it.

Exploring the mixer is a lot of fun! What I discovered, by delaying the closer mics, is that the attack gets smoothed out in a way that can't be done with the Attack MIDI parameter. Mixing can be a powerful sound design tool!

Paolo

[Air]

Pixelpoet1985: You are mixing up a few things. The legato timing has nothing to do with time alignment or microphone alignment.

Making presets with the correct delays of the recordings is an option we can think about but that would not be time alignment. Time alignment means "all micrphones start at the same time". This happen when you turn off the delays or set them to 0 as I mentioned before. So please this no false advertising!

The more different delays you have set to the mics the more phasing you will get. That is way I made presets in which the close and mid mics are aligned to get a compact close sound. After around 15 to 20 ms the ear can hear 2 signals which are quite the same as delay and not so much of phasing. Thats why it is set to 21ms and it is very close to the real distant of the recordings.

Is the vibrato phasing also happening when you have opened just 1 mic?

Best

Bernd

[Jason T. Shannon]

BERND,

As the original poster. I wanted to thank you for all the info. Through this thread I've really started to get a firm grasp on how to use the delay function, and already implemented some major improvements in my mix template. 

An important follow up question that I do have is regarding the Decca Tree channels/microphones. Were these channels time aligned individually or as a pair? In other words, do I need to delay/offset the Main-LR and Main-C differently to restore the equilateral triangle required for the Decca Tree? I'd think that the spatial relationship between L/C/R channels is important to maintain. 

I'm planning on creating my own presets (to the best of my ability) that match something close to the original recording configuration, it doesn't seem like it would be difficult to do since I understand the general layout of the Synchron Hall. For those that are interested: perhaps we could work together to dial this in. 

Regards,
Jason Todd Shannon 

[Air]

Hi Jason,

You are welcome! 

The LCR is a recorded. So L-R are around 3 ms later than the C depending on where the section was placed in the room!

Best

Bernd

[Pixelpoet1985]

First: I never accused someone. If it sounded harsh, I'm sorry. It's very easy with written words to interpret something. It was not my intention!

@Another User said:

If you want the same situation as in the original recording, just add delay to the mics, increasing it in order of distance. This is hardly what is intended to be ideal in actual use, for the bare reason that we humans hear from a single position. Our brain can add focus on 'spots', and it is on this illusion that a sound engineer plays when mixing mics, delays, reverbs.

Yes, but the problem is: I don't know the numbers. 21m or 42ms seems very vague. Every instrument and each mic would need a number on its own, isn't it?

[Pixelpoet1985]

@Another User said:

Is the vibrato phasing also happening when you have opened just 1 mic?

The curious thing is that it happens after some time of playing and not instantly – no matter which preset (senza, regular, molto). I don't know at the moment, but if I recall correctly it also happens with just the room mix. I haven't tested all possible scenarios (I only have the standard microphones), but figured out that some delay eliminates the phasing.

[PaoloT]

@Another User said:

I don't know the numbers. 21m or 42ms seems very vague.

On the contrary, that's very precise! Just convert the time to distance, and you get how much the delay moves the instrument to the back of the stage (or the mics toward the listener). As an added information, Bernd wrote above that 21ms (=7m) is also a safe amount, since it doesn't risk to become phasing.

Paolo

[Pixelpoet1985]

@PaoloT said:

On the contrary, that's very precise! Just convert the time to distance, and you get how much the delay moves the instrument to the back of the stage (or the mics toward the listener). As an added information, Bernd wrote above that 21ms (=7m) is also a safe amount, since it doesn't risk to become phasing.

Yes, but using this value for everything would mean that each instrument was recorded at the same position / distance. 7m for the strings? Come on. 

I mean if the original delay of the main mic was e.g. 7ms for the strings and it was negatively delayed by 7ms to match the close mics of the strings. Why do you add 21ms which is way more delay? I don't understand that. This is what I mean with "vague".

I'm asking for the exact numbers and only VSL can give us this information. I don't know at what position the close and mid microphones (and all the other relations) are. Maybe it would have been better if there was no delay built in and all the different delays were made in the Synchron player.

[Jason T. Shannon]

Just posting this follow-up because I thought it might help others...

There are stage plots of the Synchron recording session published on the website, which show where the sections and microphones are placed. Knowing that the Synchron stage is 30m wide and 17-19m deep, it is pretty easy to derive the original distances and spacing between the microphones and the sections. 

It looks to me that the 21ms delay in the string presets (as an example) are right on, as Bernd has already suggested, because the distance between the MAIN L/R and the string sections are between 5-7 meters (so a good delay variable to use as a starting point would be 15-21 milliseconds). 

Additionally, I did some tests in my DAW, and confirmed everything that was suggested in this thread. The microphones are clearly time aligned, meaning that when you disable the delay, or use 0 as a variable, the transient attack of the samples are all aligned to the same starting point. The one minor exception to this is that the relation between the MAIN L/R and the CENTER of the Decca tree were time aligned as a group, for the purposes of maintaining the spatial relationship of L/C/R. 

QUOTE: "And I think it is a very wrong decision by VSL to avoid delay while playing on the keyboard."

I was of a similar mind until I understood what was going on. Now I have to respectfully disagree with this. Time alignment is the right approach because I want minimal latency as a starting point, and it is easy enough to add the delays back in to achieve whatever room layout you'd prefer, including a preset that approximates the original recording layout/timing. I wouldn't want to be in the opposite situation where I was trying to compensate for baked in delays and I was trying to remove them. 

And lastly, a note on phasing, because this is something that I've struggled with in the past. To avoid phasing, you have to play with the channel delays and the phase button in the mixer just like you would in a real recording session. These are real audio signals of real microphones and phasing is an avoidable physics problem. This is an absolute requirement unless you are only using a single microphone. Additionally, you have to be careful of how you are routing signals in your DAW. For example, if you are routing a dry signal then using a send for a reverb with a less than 100% wet mix, this can cause phase issues. Make sure to use 100% wet reverbs on FX sends and be careful of any FX that might be splitting and recombining signals. Finally, make sure your MIDI setup isn't double triggering notes. I use a ROLI Seaboard and it double triggers MIDI if it isn't configured for the VSL Synchron player correctly. These things can also cause phase issues! 

 Regards,
Jason Todd Shannon

[Pixelpoet1985]

@Another User said:

And lastly, a note on phasing, because this is something that I've struggled with in the past. To avoid phasing, you have to play with the channel delays and the phase button in the mixer just like you would in a real recording session. These are real audio signals of real microphones and phasing is an avoidable physics problem. This is an absolute requirement unless you are only using a single microphone. Additionally, you have to be careful of how you are routing signals in your DAW. For example, if you are routing a dry signal then using a send for a reverb with a 50/50 wet/dry mix, this can cause phase issues. Make sure to use 100% wet reverbs on FX sends and be careful of any FX that might be splitting and recombining signals. Finally, make sure your MIDI setup isn't double triggering notes. I use a ROLI Seaboard and it double triggers MIDI if it isn't configured for the VSL Synchron player correctly. These things can also cause phase issues! 

Thanks for the input! Very appreciated! But I don't use any processing. The phasing just happens. And it's also not my job figuring this out. In no other library I have there is such a severe phasing issue. I think it's because of the delays.