SPL targets for speaker design

[ShinOBIWAN]

90 dB drivers, driven by 32 Watts gets you to 115dB, no?

A 25dB increase in SPL requires 323w.

Ie. 2^(25/3) = 323

 

[tade]

Speaking of the TTS response; I have read that the native people of in the arctic are suffering from massive hearing loss. Interestingly, the cause turns out to be that the actic regions are exceptionally free of noise. Especially when hunting, a party will travel for hours, barely speaking. The change in modern times is attributed to the use of rifles for hunting. The extremely loud and short sound gives the ear no chance to protect itself.
Earplugs maybe?

 

[gedlee]

Environmental noise is by far and away the largest cause for hearing loss. Some musicians suffer from this because of long exposure times, but for a home music listener hearing damage is simply not going to be an issue unless you listen very ofetn very loud. Farmers are well know for hearing loss from daily work.

A car with the windows open at 60 MPH will be > 120 dB(C). Ever open a window at these speeds? I seriously hope nobody drives at these speeds with the windows open. And a sun roof vent or cracked window for air generates well in excess of 110 dB(C). We are regularly exposed to very loud sounds in our daily life - table saws, routers, etc. These are the things to worry about, NOT your stereo. (Notably I am excluding some car stereos which are known to exceed 120 dB regularly and have been know to cause damage in young people - the most suceptable group.)

 

[x. onasis]

A 25dB increase in SPL requires 323w.

My mistake, thanks for the correction. 32 Watts only gets you to 105 dB. I still maintain that's plenty loud for living room listening.

 

[graaf]

Re: Re: Re: SPL targets for speaker design

You are correct about the book, but SPL doesn't know the size of the room. SPL is SPL big room or small, it doesn't matter.

SPL is SPL - but certainly NOT the loudness
and there lies the main difference between PA and music reproduction at home

I think that what we have here is a misunderstanding
I think that the title and topic of this thread misses the real issue

the loudspeaker are not for SPL but for LOUDNESS, because what auditory system perceives is loudness

the real issue is loudness
and from loudness perspective big and small rooms differ significantly

loudness depends on SPL - BUT not only
it depends also on other factors

in particular early reflections (<50-60 ms) in reverberant spaces add to the perceived loudness
the SPL stays the same but the ear becomes more sensitive to the sound in presence of early reflections, in a reverberant space
this mechanism is taken into account in architectural acoustics design:

There are two ways to make sound "louder". The ordinary way is to make it physically louder. Cup your hands and yell, makes it louder. The other way is to add as many reflections as you can. Cup your hand over your ear makes it louder, or talk in an office. Not all reflections help hearing but those that do help people understand what they are hearing are called "early reflections". Adding early reflections raises the apparent loudness of the direct sound in a comfortable, natural way, much more agreeable than turning up the volume

see: http://www.church-acoustics.com/aa102.htm

SPL stays the same BUT loudness goes up

and the smaller the room the more early reflections we have (unless we treat the room with sound absorbers to make it acoustically "appear to be bigger")

therefore I am not convinced that the given SPL is subjectively perceived as the same loudness in a smaller room and in a bigger room, in a listening room and in a big auditorium or a stadium, that "SPL is SPL" explains all

the SPL is the same but NOT THE LOUDNESS

this is the first big difference between "PA sound systems" and music reproduction at home

but not the only

there is also an issue of ambient noise which is significantly different in a quiet listening room than in "PA application"
the "PA sound systems" have to play above significantly louder ambient noise then domestic equipment
human hearing is sensitive to RELATIVE SPL
therefore what is important in music reproduction is not absolute SPL but rather the DYNAMIC RANGE

so what about those "115 - 124 dB peak" "High Level Music"?
typical example is a rock concert, peaking at even 130 dB but :

In reality, most concerts have 70 dB or less dynamic range

see: http://www.yamaha.com/yamahavgn/CDA...,,CNTID%253D48553%2526CTID%253D227500,00.html

It means that there is 130 dB of music in peaks and 60 dB ambient noise AT BEST

In comparison in a typical indoor PA application field, that is in church acoustics, it is assumed that

normal background noise on average is 50-60dB

see: http://www.allchurchsound.com/ACS/glossary/s.html

and sometime "PA sound system" has to cope with even higher ambient noise, for example:

It's difficult to imagine a more idyllic setting for an open-air concert by vocal master Tony Bennett than a starlit gold-sand beach on Mexico's Pacific Coast. Bennett recently entertained a private audience of about 175 at the Four Seasons Resort at Punta Mita near Puerto Vallarta, Jalisco, Mexico, performing against a backdrop of sand and sea. But along with the scenic surroundings came the sounds of the surf, creating a background noise level of up to 75 dB SPL. To make Bennett and his band heard without the sound being too loud, Bennett's long-time production manager and sound engineer Tom Young built his sound system for the event around self-powered loudspeakers from Meyer Sound.

see: http://www.meyersound.com/news/2005/tony_bennett/

that is 75 dB of background noise!
this is a situation when one needs "High Level Music" "sound system" capable of clean 130 dB in peak

ambient noise levels in a typical listening room are MUCH lower
I have about 45 dB (unweighted) of background noise IN PEAKS in my listening room and well below 40 dB(A)

to recreate full dynamic range of a rock concert I need a sound system capable of replicating its 70 dB dynamic range that is to produce 110 dB SPL at the listening space - that is 20 dB less than "PA sound system" used at the concert
but let us assume safely that it is 15 dB less
to this we can add equally safely about 10 dB (I believe that more could be added) increase in the ear sensitivity in presence of early reflections

Now – substracting this 25 dB (15 dB less ambient noise and 10 dB increase in the ear sensitivity) from McCarthys "PA" "125 dB in peak" we arrive at around 100 dB in peaks as reasonable SPL target for loudspeaker dedicated for music reproduction at home
ANY MUSIC - indluding the wildest rock concerts

and this is completely in agreement with National Research Council of Canada statement that 90 dB (unweighted) average is:

very loud and considered far beyond normal listening levels

see: www.soundstagemagazine.com/measurements/test_loudspeakers.htm
and it is also completely in agreement with my own experience and measurements (and of other people like x. onasis in this thread) that with recordings of the highest peak/average ratio i.e. well recorded symphonic music the "average" is in the 70’s (70<80 dB), loud is in the eighties (80<90 dB) and occasional peaks hit 95-100 dB

this is the SPL-as-a-factor-of-loudness target for design of loudspeakers dedicated for music reproduction at home

please correct me where I am wrong 🙂

best,
graaf

 

[gedlee]

Your analysis is a good atempt and gets at some issues, but you are mistating some things.

Loudness CAN increase in a small room because of reflections etc. and this is called the Reveberant field. One must be well into the reverberant field however for this effect to occur. At best it might subtract 6 dB from the numbers listed. In most cases in home environment, the lsitener is more in the direct field and only a few dB are added to the perception from the reverberany field. The relationship between loudness and SPL is not as simple as you state. The complexity is why I tend to ignore this factor when I do designs because it is very hard to quantify.

Your use of background noise to scale SPL is not clear to me and I would never use background noise in a home environment as a factor in the design, but if I did it would only push it up.

Hence, I have already agreed that one might want to take 10 dB off of McCarthy data for home listening, but I think that any more than this is unwarranted. And you MUST consider the musical dynamics in a design, and not simply the long term SPL.

 

[graaf]

Loudness CAN increase in a small room because of reflections etc. and this is called the Reveberant field. One must be well into the reverberant field however for this effect to occur. At best it might subtract 6 dB from the numbers listed.

well, I am not an expert, I have only my experience, some measurements of average and peak in my room but the people I quoted from seem to suggest that it is not a matter of added dB but rather of an increase in hearing sensitivity:

There are two ways to make sound "louder". The ordinary way is to make it physically louder. Cup your hands and yell, makes it louder. The other way is to add as many reflections as you can. Cup your hand over your ear makes it louder, or talk in an office.

I imply from the above that adding of early reflections is not about making the sound "physically louder" (what I read as "more sound pressure" like in that example "cup your hand an yell") – it is like having bigger pinnae, like a bat or something ("Cup your hand over your ear") – it seems to be rather about raising hearing sensitivity rather than raising the sound pressure level by some dBs
These people are professionals in PA and I assume that they are not telling fairytales
than perhaps they are right and only my conclusion is unfounded? what am I missing?

In most cases in home environment, the listener is more in the direct field and only a few dB are added to the perception from the reverberant field.

I cannot find the equations for the critical distance but I have found this info:

How far from a sound source is the critical distance? In the average living room, the critical distance for a loudspeaker is about 86cm! That means we listen to most music well into the reverberant field.

see: http://www.postaudio.co.uk/education/acoustics/room_acoustics.html

and here is "critical distance" calculator:
http://www.mcsquared.com/ssdesign.htm

Are they wrong?

The relationship between loudness and SPL is not as simple as you state

I’m sure that it must be complicated
Therefore I have only stated that SPL is a factor of loudness but not the only factor

Your use of background noise to scale SPL is not clear to me

I think it is an example of psychoacoustical phenomenon called partial simultaneous masking, when a signal is made less audible in a presence of a masker, a noise or unwanted sound

consequently, to achieve the same loudness a tone in a presence of a masker needs to have a higher sound pressure level (SPL)
ambient noise is such a masker – the lower it is the less masking and the less dB SPL we need to achieve the same loudness

everyone can do a simple experiment – turn a TV or a radio on and adjust the volume to a level at which it is easy to hear what is going on
than turn on a noisy device – a vacuum cleaner, an electric kettle, a blender etc. whatever
what happens? It is obvious – we can’t hear the TV/radio as well as before – "loudness of the signal is decreased by the masker" but its SPL of course stays the same
and again adjust the volume to a level at which it is easy to hear what is going on
a comfortable level
after a few moments turn the noisy device off
and now You hear – not that obvious but it is the same mechanism working:
suddenly the radio/TV starts to play uncomfortably, way too loud! Even though the SPL stays the same

there can be 60 dB or more of that masker ie. background noise at a rock concert
and only 40 dB or less in a quiet listening room

can we draw a conclusion? I suppose yes we can
but perhaps I am wrong?

Hence, I have already agreed that one might want to take 10 dB off of McCarthy data for home listening, but I think that any more than this is unwarranted

why have You agreed on that 10 dB? on account of what phenomenons?
if those phenomenons are real then why only 10 dB? why not more?

best,
graaf

 

[Salas]

and it is also completely in agreement with my own experience and measurements (and of other people like x. onasis in this thread) that with recordings of the highest peak/average ratio i.e. well recorded symphonic music the "average" is in the 70’s (70<80 dB), loud is in the eighties (80<90 dB) and occasional peaks hit 95-100 dB

this is the SPL-as-a-factor-of-loudness target for design of loudspeakers dedicated for music reproduction at home

please correct me where I am wrong 🙂

best,
graaf

You are right about how most people listen to their Hi-Fi at home. But to listen to 100db peaks without compression or perceivable distortion (i.e. true Hi-Fi) you need 10dB headroom. So your speakers need to can go 110dB max at listening position so to get 100 dB clean. Now an average home listening seat is 3m away from the centerline of a stereo pair. If one speaker can do 115dB max @ 1m, two speakers will do 118dB centerline at 1m, when 2m horizontally separated. Lose 5dB at 2m to listening position, lose another 2.5dB at 3m (1.5dB better in total at 3m than inverse square law, benefiting from small room reverberation gains). So you are down to 110.5dB max headroom at 3m. Dr. Earl Geddes stated that we must be aiming for 125dB - 130dB max SPL capable speakers for real dynamic range, but given the circumstances he agreed with me that we can deduct roughly 10dB from live or club situations, if we are talking home listening. Ok then, 125dB is his base, 115dB is our base, -10dB is his concession for home Hi-Fi. So 125-10=115. Where is the disagreement?

 

[gedlee]

well, I am not an expert, I have only my experience, some measurements of average and peak in my room but the people I quoted from seem to suggest that it is not a matter of added dB but rather of an increase in hearing sensitivity:

I imply from the above that adding of early reflections is not about making the sound "physically louder" (what I read as "more sound pressure" like in that example "cup your hand an yell") – it is like having bigger pinnae, like a bat or something ("Cup your hand over your ear") – it seems to be rather about raising hearing sensitivity rather than raising the sound pressure level by some dBs These people are professionals in PA and I assume that they are not telling fairytales
than perhaps they are right and only my conclusion is unfounded? what am I missing?

There is no such things as "increased hearing sensitivity". There is a relationship between the integrated SPL over the first 10 - 20 ms or so and loudness. This is why early reflections can increase the loudness in a small room. But if one tries to minimize these early reflections, like I do, then this effect is also minimized.

And don't believe everything you read on the internet

Are they wrong?

I'm not going to do the calulations, but it seems to me that the critical distance would be much further back than quoted in your reference. By "critcal distance" do they mean the point at which the direct field and the reverb field are equal? This would imply that there was no direct field in a small room, that everything that we hear is reveberation. And at any rate there is the time aspect to consider and the direct/reverberant ratio is a steady state reference.

I think that you are over simplifying the problem. Lets stick with SPL and leave out "loudness" as this is much easier to talk about.

consequently, to achieve the same loudness a tone in a presence of a masker needs to have a higher sound pressure level (SPL)
ambient noise is such a masker – the lower it is the less masking and the less dB SPL we need to achieve the same loudness

can we draw a conclusion? I suppose yes we can
but perhaps I am wrong?

For what you say to be true the noise has to be close to or greater than the signal. In a home hi-f as well as a concert this is not the case. The noise level is so far below the signal that no masking occurs.

why have You agreed on that 10 dB? on account of what phenomenons?
if those phenomenons are real then why only 10 dB? why not more?

I would agree on 10 dB due to the reverberant field increase in SPL. This is a bit higher than it would actually be however, since 6 dB is about all the reverberant field can add to the level.

You are right about how most people listen to their Hi-Fi at home. But to listen to 100db peaks without compression or perceivable distortion (i.e. true Hi-Fi) you need 10dB headroom. So your speakers need to can go 110dB max at listening position so to get 100 dB clean. Now an average home listening seat is 3m away from the centerline of a stereo pair. If one speaker can do 115dB max @ 1m, two speakers will do 118dB centerline at 1m, when 2m horizontally separated. Lose 5dB at 2m to listening position, lose another 2.5dB at 3m (1.5dB better in total at 3m than inverse square law, benefiting from small room reverberation gains). So you are down to 110.5dB max headroom at 3m. Dr. Earl Geddes stated that we must be aiming for 125dB - 130dB max SPL capable speakers for real dynamic range, but given the circumstances he agreed with me that we can deduct roughly 10dB from live or club situations, if we are talking home listening. Ok then, 125dB is his base, 115dB is our base, -10dB is his concession for home Hi-Fi. So 125-10=115. Where is the disagreement?

You are right that there is not a big difference in our positions. But I would still contend that the levels that I listen to - ON OCCASION - are much higher than what is being quoted here. I know for a fact that I have on numerous occasions listened at 110 dB(c) slow average which would have peaks at 125 dB or so. A home theater setup creating real world "loudness" of a concert video, this is not hard to do.

If others listen lower fine, but my stated targets are not unreasonable for a premium system.

 

[Salas]

You are right that there is not a big difference in our positions. But I would still contend that the levels that I listen to - ON OCCASION - are much higher than what is being quoted here. I know for a fact that I have on numerous occasions listened at 110 dB(c) slow average which would have peaks at 125 dB or so. A home theater setup creating real world "loudness" of a concert video, this is not hard to do.

If others listen lower fine, but my stated targets are not unreasonable for a premium system.

I just put down the 100dB clean peaks example discussed. Give the same system to a youngster, or get carried away in an HT room, and here is your 130dB upper limit - 10 concession. Our 115dB max @ 1m single home speaker example, climbs to 120dB. Spot on again. Regards.

 

[pdan]

The demand for ever increasing SPL's is a crude and vulgar demand: a rough taste that betray infantile desires and expectations. However, since the "human" is a rapidly attenuating species, these demands are quite in order, "reasonable" even.


Cilla

 

[gedlee]

The demand for ever increasing SPL's is a crude and vulgar demand: a rough taste that betray infantile desires and expectations.

WOW! Thats a bit over the top don't you think?

 

[Salas]

The demand for ever increasing SPL's is a crude and vulgar demand: a rough taste that betray infantile desires and expectations. However, since the "human" is a rapidly attenuating species, these demands are quite in order, "reasonable" even.


Cilla

Lynn Olson (our favorite audio thinker) mature in age, and very well put philosophically wise, certainly far away from coming across as 'vulgar' looks to develop a 98-102dB sensitive OB in his Beyond The Ariel thread. His homemade PP DHT does 20W per channel. So he is shooting for 102+13=115dB per channel at 1m for the upper system. 118dB stereo. And he is going to use solid state for lows. If he will use 200W at 50Hz he will get 120-122dB spl @t 1m there. 123-125 in stereo. If he will use his DHT over the 200Hz modal region the 115dB will keep up, since music declines in power more or less like pink noise i.e. -3dB per rising octave. This gets him square in to the 'loud' club.
There is true benefit in headroom, not necessarily abused for compressed average loudness as we know from pimped up cars etc.

A comment for the philosophical part also: Perspective is everything.

As in:

'I am the son, and the heir, of a shyness that is criminally vulgar, son and heir of nothing in particular.'

From:
How Soon Is Now?
Morrissey & The Smiths - Hatful Of Hollow

 

[graaf]

You are right about how most people listen to their Hi-Fi at home. But to listen to 100db peaks without compression or perceivable distortion (i.e. true Hi-Fi) you need 10dB headroom. So your speakers need to can go 110dB max at listening position so to get 100 dB clean.

10 dB means 10:1 power ratio and 3:1 excursion ratio

does it mean that a loudspeaker has to take 1000W max if it has to take 100W without audible compression?
and that it has to go 15 mm linearly if it has to go 5 mm without audible distortion?
😕

is that right?

I am sorry if my qestion are not very wise
I would just like to know therefore I ask 🙂

best,
graaf

 

[graaf]

There is no such things as "increased hearing sensitivity".

perhaps it is not precise expression
but as an example it is a fact that speech is more intelligible in the presence of early reflections and less intelligible in the presence of late reflections
and also what is the point of "cupping your hand over your ear"?
it works
see attached picture 😉 (that is not me 😉 )
by saing "increased hearing sensitivity" I just mean that sounds can be more or less audible in spite of SPL being the same
am I wrong? why?

And don't believe everything you read on the internet

Do I seem to believe? 🙂
I don't - therefore I ask and seek to verify or falsify 🙂

I'm not going to do the calulations, but it seems to me that the critical distance would be much further back than quoted in your reference. By "critcal distance" do they mean the point at which the direct field and the reverb field are equal? This would imply that there was no direct field in a small room, that everything that we hear is reveberation.

yes, this is what they mean and what they imply
I will investigate it further

I think that you are over simplifying the problem.

not intentionally 🙂
I am only trying to understand 🙂

For what you say to be true the noise has to be close to or greater than the signal. In a home hi-f as well as a concert this is not the case. The noise level is so far below the signal that no masking occurs.

but this is only partial masking - the signal remains audible - only its loudness is decreased
secondly, how much close? and what masker? I read (not "on the internet", in an academic handbook of psychoacoustics) that noise of a given SPL is twice as loud as a tone of a given SPL
so it is perceived as if it was 10 dB higher in SPL, that is only an example, an illustration of this complicated issue
thirdly, what signal? I am talking about the soft and the softest part of what we want to be able to hear
I am talking about the music close to the noisefloor. Say 10 dB above the noisefloor

If masking is not a problem - "if it is not the case at the concert" as You have put it - then why people at "Meyer sound" see that PA at Tony Bennet concert in the presence of 75 dB background noise is a difficult problem?
(see the example I referred to and a quotation from their website I cited above)
and why people at Yamaha speak of dynamic range as limited by background noise? In what sense does the background noise limit the dynamic range if there is no masking?
(see another quotation I cited above)

I just can’t understand 😕

If others listen lower fine, but my stated targets are not unreasonable for a premium system.

certainly they are not unreasonable as a target

best,
graaf

 

[Brett]

There has been a lot of discussion bantered about on what the actual SPL requirements for a loudspeaker system design are. I recently read a book "Sound Systems: Design and Optimization" by Bob McCarthy. In it he has a table of SPL values (figure 1.14). Since his numbers are completely consistant with what I have been saying, I thought that I would quote them here. This makes two highly credible references for design goals in terms of SPL.

These are listening levels at the seating position and all numbers are dB SPL with no weighting.

High Level Music: 103 - 112 dB with peaks to 115 - 124 dB
Medium Level Music: 91 - 100 dB with peaks to 103 - 112 dB
Low Level Music: 79 - 88 dB with peaks to 91 - 100 dB

Now if these are at the seating position then typically we would want to add 3-6 dB for the 1 meter levels, but this, of course, depends on the room.

I basically agree, but the system I am designing for music and HT doesn't quite get there. Fronts and sides will all do 120+dB at driver excursion specs above 50Hz or so. Rears will be about 116dB, all at 1m.

 

[Salas]

10 dB means 10:1 power ratio and 3:1 excursion ratio

does it mean that a loudspeaker has to take 1000W max if it has to take 100W without audible compression?
and that it has to go 15 mm linearly if it has to go 5 mm without audible distortion?
😕

is that right?

I am sorry if my qestion are not very wise
I would just like to know therefore I ask 🙂

best,
graaf

Yes, it is right. Human aural perception needs +10dB SPL to describe a sound as twice as loud. Just think of it in this way: My system & speakers must become dirty at twice the peak loudness I listen to music, so to be totally clean and uncompressed for my habits.

Did you ever have an older amp with a nice needle power meter on its front? See it idling at 10W and then momentarily hitting 100W, and then again. 10dB.

You don't need 1kW. A 100dB sensitive speaker needs 100W to hit 120dB. The 87dB average speaker will need the kW and the excursion...to die.

 

[AndrewT]

I have found that 70db average listening level at the listening position is a nice level that reproduces voice at a similar level to what we hear normally.
Up the level to around 80db and most folk would call that loud and would require a conversation to be shouted through that sustained averaged level.
Add another 10 db to get 90db average at the listening position and this I would call very loud. Conversation over/through that is pointless.

Now apply the normal peak (crest) to average ratio of +20db and we end up at 90, 100 and 110db peak levels at the listening position for those three cases.

Go back to the speakers required to produce those levels and for an 8foot (2.4m) listening distance and using one pair of speakers we have to add a further 5db for a 1m test sensitivity value.

We need about 115db @ 1m to reproduce very loud music that has 20db average to peak ratio.

How do we get our 115db?
Try using 160W into 93db/W/m speakers.

That does not seem extreme to me and covers all my listening needs.

If we take the earlier arguments and apply the +10db for the non domestic listening environment we are back to Gedlee's original suggestion of 120db for highest level reproduction (300W into 100db/W/m) or stay at 110db for high quality domestic level reproduction.

What's all the argument about?

 

[Salas]

You are right about how most people listen to their Hi-Fi at home. But to listen to 100db peaks without compression or perceivable distortion (i.e. true Hi-Fi) you need 10dB headroom. So your speakers need to can go 110dB max at listening position so to get 100 dB clean. Now an average home listening seat is 3m away from the centerline of a stereo pair. If one speaker can do 115dB max @ 1m, two speakers will do 118dB centerline at 1m, when 2m horizontally separated. Lose 5dB at 2m to listening position, lose another 2.5dB at 3m (1.5dB better in total at 3m than inverse square law, benefiting from small room reverberation gains). So you are down to 110.5dB max headroom at 3m. Dr. Earl Geddes stated that we must be aiming for 125dB - 130dB max SPL capable speakers for real dynamic range, but given the circumstances he agreed with me that we can deduct roughly 10dB from live or club situations, if we are talking home listening. Ok then, 125dB is his base, 115dB is our base, -10dB is his concession for home Hi-Fi. So 125-10=115. Where is the disagreement?

See my example, based on your and my loudness needs at home. The conclusion was a stereo system that can do 115dB max @ 1m per channel. We just need a 95dB sens speaker and a 100W per channel amp. Irrational? Did you ever spend an evening at the odd pal's house who shopped from a pro store instead of a Hi-Fi shop so to listen to rock better? Didn't you just hate the tone and his listening habits? But the feeling of dynamic freedom was inescapable, wasn't it? He can get a pair of rather nice 12inch 95dB speakers for $500-$1000.
Do it like the old guys. Use hi sens Hi-Fi tech. And you get the tone back too.

 

[gedlee]

Graaf

You pose interesting questions, but very complex ones. Loudness is very dependent on the spectrum while we have no trouble specifying a dB SPL. The Fletcher-Munson curves are a weak attempt at this translation.

In general if the noise masker is 6 dB below the mean SPL then it is not masking and PROBABLY not audible. This needs to be true at all frequencies and depends on the noise spectrum. A coherent noise is audible far below the masker.

I cannot explain the Meyer results unless they were playing back the music at a fairly low level which would be possible with Tony Bennit at a dinner club.