Download the jbl speakershop program. JBL Speakershop

JBL Speakershop includes two independent programs: Enclosure Module and Crossover Module.

The Enclosure Module is designed to determine the required volume and dimensions of low-frequency loudspeaker enclosures. The sound quality of the design is assessed at normal listening level (small signal analysis, including group delay, phase and amplitude-frequency response, voice coil resistance) and at maximum volume (large signal analysis, taking into account the thermal acoustic power index at mid frequencies and maximum power at various deviations).

The Enclosure Module utility allows you to independently choose two directions for designing enclosures: taking into account specific speakers or by selecting suitable speakers for an existing enclosure (limited space).

The program module under consideration offers modeling of enclosures with a bass reflex of a custom, optimal and design designed for a unique frequency band, enclosures with a passive radiator, as well as closed systems of an optimal or custom type. Simultaneous demonstration of all types of designs facilitates their comparative analysis.

The program describes the structure and main parameters of each type of housing, and contains lists of their advantages and disadvantages. For beginners, there is a help file to make the job easier, and examples are included with relevant notes and instructions.

The set of minimum parameters required for designing the housing includes the name of the manufacturer and model number, as well as the value of the resonant frequency of the speaker, the volume of air with an elasticity equal to the elasticity of the speaker suspension, and the quality factor of the device, taking into account all losses. The full list of parameters includes a long series of mechanical, electrical and combined values ​​of the designed device. Among other things, the JBL Speakershop Enclosure Module plots graphs of maximum sound power, amplitude-frequency response (normalized and when a 2.83 V test signal is applied), voice coil resistance, group and phase delays.

The second part of the JBL Speakershop program - Crossover Module - is designed to determine the parameters of crossover filters that separate the signal into low and high frequencies. The utility calculates two- and three-way passive separation systems of the first, second, third and fourth orders using a number of standard filters: Chebyshev, Bessel, Butterworth, Gauss, Legendre, Linkwitz-Riley and some others. The result of the work is the construction of a detailed electrical circuit diagram of a unique crossover system with a detailed description of each element.

In Russia, the JBL Speakershop program has become widespread among radio amateurs who are developing their own car speaker systems. However, the amplitude-frequency characteristics of a car sound reproduction system calculated and plotted in this utility are very inaccurate and strongly depend on the design features of a particular car. For proper operation, additional data must be entered into the program, for example, the transfer function of the car interior.

The JBL Speakershop program was created in 1995 by specialists from the American company JBL. The company is part of the Harman International Industries association, specializing in the production of high-end speaker systems and related electronics. JBL products became the basis for the development of the THX standard, and the company's dynamic heads are used in cars from the world's leading manufacturers.

The JBL Speakershop interface language is English only. However, on the Internet there is a detailed description of the work in Russian.

System requirements for the utility are minimal. JBL Speakershop runs on the Microsoft Windows operating system, including its latest versions: Vista and 7. The only exception is the lack of support for 64-bit operating systems.

Program distribution: free

Software for calculating subwoofers and speaker systems.

JBL Speakershop includes two independent programs: Enclosure Module and Crossover Module.

The Enclosure Module is designed to determine the required volume and dimensions of low-frequency loudspeaker enclosures. The sound quality of the design is assessed at normal listening level (small signal analysis, including group delay, phase and amplitude-frequency response, voice coil resistance) and at maximum volume (large signal analysis, taking into account the thermal acoustic power index at mid frequencies and maximum power at various deviations).

The Enclosure Module utility allows you to independently choose two directions for designing enclosures: taking into account specific speakers or by selecting suitable speakers for an existing enclosure (limited space). The program module under consideration offers modeling of enclosures with a bass reflex of a custom, optimal and design designed for a unique frequency band, enclosures with a passive radiator, as well as closed systems of an optimal or custom type. Simultaneous demonstration of all types of designs facilitates their comparative analysis. The program describes the structure and main parameters of each type of housing, and contains lists of their advantages and disadvantages. For beginners, there is a help file to make the job easier, and examples are included with relevant notes and instructions.

The set of minimum parameters required for designing the housing includes the name of the manufacturer and model number, as well as the value of the resonant frequency of the speaker, the volume of air with an elasticity equal to the elasticity of the speaker suspension, and the quality factor of the device, taking into account all losses. The full list of parameters includes a long series of mechanical, electrical and combined values ​​of the designed device. Among other things, the JBL Speakershop Enclosure Module plots graphs of maximum sound power, amplitude-frequency response (normalized and when a 2.83 V test signal is applied), voice coil resistance, group and phase delays.

The second part of the JBL Speakershop program - Crossover Module - is designed to determine the parameters of crossover filters that separate the signal into low and high frequencies. The utility calculates two- and three-way passive separation systems of the first, second, third and fourth orders using a number of standard filters: Chebyshev, Bessel, Butterworth, Gauss, Legendre, Linkwitz-Riley and some others. The result of the work is the construction of a detailed electrical circuit diagram of a unique crossover system with a detailed description of each element.

In Russia, the JBL Speakershop program has become widespread among radio amateurs who are developing their own car speaker systems. However, the amplitude-frequency characteristics of a car sound reproduction system calculated and plotted in this utility are very inaccurate and strongly depend on the design features of a particular car. For proper operation, additional data must be entered into the program, for example, the transfer function of the car interior.

The JBL Speakershop program was created in 1995 by specialists from the American company JBL. The company is part of the Harman International Industries association, specializing in the production of high-end speaker systems and related electronics. JBL products became the basis for the development of the THX standard, and the company's dynamic heads are used in cars from the world's leading manufacturers.

The JBL Speakershop interface language is English only. However, on the Internet there is a detailed description of the work in Russian.

Hello, today I’ll tell you how to install JBL Speakershop on Windows x64 bit. Since without additional work the program will not work with it. Perhaps I will disappoint you by saying that simply downloading JBL Speakershop for 64 bit will not work, because such a version of the program simply does not exist.

Therefore, you will need to install a virtual machine, then install Windows XP on it and only then install our Speakershop. The process may seem complicated, but don't worry. We will do everything quickly and simply.

Despite the fact that there are analogue programs that work with modern versions of Windows x64, many were upset by the lack of such an opportunity in Speakershop, since they were used to making calculations there.

It should be noted that JBL Speakershop should run not only on Windows XP, the main thing is that the system is 32-bit. It’s just that I ran it myself on the XP emulator, which is why I’m using this as an example.

Installing VirtualBox

If you have any problems with the Internet on a virtual machine and you have downloaded the program distribution, you will have to use second way - create a shared folder for Virtualbox and the main operating system.

Creating a Virtualbox shared folder

In order to create a shared Virtualbox folder, go to the top menu in “Settings“

Then go to “Shared Folders” and add the one you need.

If you use this shell for anything other than calculations in JBL Speakershop, then create a special folder for exchanging files between systems. You can also put the downloaded JBL Speakershop in it. But in this example, we are doing this only for Speakershop, so I add (share) only the folder with the program distribution. Find the desired folder on your computer and specify the path.

For convenience, the next time you start, check the “Auto-connect” and “Create a permanent folder” checkboxes.

Don't forget to confirm everything by clicking OK. A warning may appear during the process, skip it.

Now go to the device tab and select “Mount guest OS disk image...”.

We install the appropriate software.

During installation, select a location, leave the rest as default. After installation, the virtual computer will reboot. There is very little left.

Go to My Computer and launch Guest Additions.

We perform standard manipulations. When finished, it will reboot again.

After rebooting the virtual computer, select “Run” from the Start menu. On the command line enter the following:

net use x:\\vboxsvr\jblspkrshp

The drive letter can be anything (x), but don’t worry and leave x (unless, of course, you already have one on your system). But the last word you need to enter is the name of the folder that you added to the system. In our case it is jblspkrshp. But if you added another one, then enter its name. Please note - not the path, but only the name of the folder.

Click OK. We restart the machine.

Installation:
1) Unzip the archive, open the folder DISK1 and run the file SETUP.exe
2) Select the path where you want to install the program and click OK
3) During installation, the program will ask you to insert Disc 2

C:\USERS\C50A~1\DESKTOP\JBL_SS\ DISK1\ , DISK1 change to DISK2, click OK.

SPEAKERSHOP consists of two independent and complementary parts:
Enclosure Module- for calculating acoustic design
Crossover Module- to calculate the parameters of separation filters.

Enclosure Module
This software helps you determine the volume and dimensions of the cabinet and evaluate the sound quality. The design is analyzed in two stages. The first step is to determine how it will perform at normal listening levels. Secondly, the maximum volume mode is simulated for the structure. This step is called large-signal analysis and includes thermal acoustic power standards in the mid-frequency range and maximum power characteristics at various excursions.

Two ways to use the program


There are two ways to design enclosures using the SPEAKERSHOP Enclosure Module program. One of them involves designing an enclosure for certain selected speakers. At the same time, the characteristics of the case vary. Another way is to find suitable speakers for your existing cabinet: you select speaker models. The design method can be selected using the Variable command on the Options menu.

When the SPEAKERSHOP Enclosure Module program is launched for the first time, the default mode is set in which the parameters to be changed are the characteristics of the acoustic design.

The spreadsheet contains columns for designing six cases. The first three are intended for calculating enclosures with a bass reflex - for optimal, custom (i.e., designed by the master himself) designs and for enclosures designed for a specific frequency band. The next column is for a custom passive radiator cabinet design. The last two columns are for optimal and custom design for closed-type cases. Because the spreadsheet shows different types of designs at the same time, you can easily compare them. The speaker settings are shown in the lower left area of ​​the spreadsheet. The graph below is the same for both methods.

The mode when the variable value is the speaker itself is set using the Variable-Loudspeaker command in the Options menu. This is in case you select suitable speakers for an existing case. The mode is very convenient for calculating the sound reproduction systems of cars, when it is necessary to select a speaker for a strictly specified volume, as it allows you to quickly check the operation of several different acoustic systems in a specific housing or in a certain limited space.

Variable-Loudspeaker mode uses a different type of spreadsheet menu. Instead of showing six different cabinet designs, as is done in Variable-Box mode, six different speakers are shown simultaneously. This makes it possible to quickly compare up to six different models.

Speaker options


If you are still new to designing acoustic cabinets or are in a hurry and want to enter only the minimum parameters required to design the cabinet, select the Parameters-minimum option in the Loudspeaker menu. A window will appear in which you can enter the minimum parameters, including the name of the manufacturer (Manufacturer), name of the model (Model), Fs, Vas and Qts. Nominal efficiency or sensitivity need to be entered only when designing enclosures with a bass reflex.
To enter complete parameters (mechanical, electrical, combined), select the appropriate command. Below we provide a brief explanation of the parameter designations.

Mechanical parameters
Fs- Natural resonant frequency of the speaker (Hz).
Qms- The quality factor of the speaker at frequency Fs, when its mechanical (not electromagnetic) losses or attenuation are taken into account.
Vas- A volume of air having an elasticity equivalent to that of the speaker mount (cubic feet or inches, or liters).
Cms- Mechanical compliance coefficient of the suspension (inches per pound or millimeters per Newton).
mms- Mechanical mass of the diffuser taking into account the aerodynamic load (ounces or grams).
Rms- Mechanical resistance in the speaker suspension (pounds per second or kilograms per second).
Xmas- The maximum or peak linear amplitude of the speaker's voice coil (inches, centimeters or millimeters). Typically defined as the distance a coil can travel in one direction while still being able to maintain a constant number of oscillations in the magnet gap. This parameter determines the maximum amplitude of oscillations at which distortion does not appear.
Sd- "Piston/cone area" of the speaker (square inches or square centimeters). Represents the area of ​​the moving part of the speaker.
Dia- "Piston diameter" (inches or centimeters).

Combined options
Qts- Quality factor of the speaker for the frequency value Fs, taking into account all electromagnetic and mechanical losses.
ho- The nominal efficiency of the speaker at an acoustic load of half the volume (the reflector is located at infinity). Efficiency is entered as a percentage.
SPL- Nominal sensitivity of the speaker at an acoustic load of half the volume (the reflector is located at infinity). Entered in decibels. Sensitivity is taken as measured along the axis at a distance of 1 meter when 1 W of electrical power is applied to the speaker. Since many manufacturers test their speakers at a fixed voltage of 2.83 V instead of 1 W, there is a 2.83 V option in the Full Loudspeaker Parameters window.

Electrical parameters
Qes- Q dynamics for the frequency value Fs. Allows only electromagnetic (not mechanical) losses or vibration damping.
Re- Voice coil DC resistance (Ohm).
Le- Voice coil inductance (millihenry).
Z- Nominal electromagnetic impedance of the speaker (usually 8 or 4 ohms).
B.L.- Speaker motor power (newton/amp, meter/tesla, pound/amp or ft/tesla).
Pe- Thermally limited maximum electrical power (W) that the speaker can handle. Typically represents the maximum electrical power without causing the voice coil to burn out.

Acoustic enclosures and their parameters


1. Bass reflex
The goal of optimizing the design of a cabinet with a bass reflex is to select a volume that provides the most even and smooth amplitude response in the range of tuning frequencies of the bass reflex port. The advantages of this design are greater mid- and low-frequency response, lower distortion due to smaller cone amplitude, higher efficiency, and lower overall cost.

1) A system with a large bass response and a system with a “smoother” bass frequency response; (Top graph)
2) An insufficiently damped system (the volume of the box is small) and an overdamped system (the volume of the box is large) (Bottom graph)

2.Band-Pass design(casing with a bass reflex, designed to allocate a specific frequency band)
Band-Pass- box design that allows you to control the amplitude response in both the lower and upper frequency regions thanks to the use of a double-chamber housing. Moreover, the speakers are located inside the case. (If there is more than one speaker, housings with three chambers, etc. may be used)


Band-Pass design means that you can use speakers that have a higher Q value (smaller magnets) than speakers used with other bass reflex cabinet designs. It provides lower distortion (high-order distortion is filtered out), increased efficiency across the operating frequency band, and requires virtually no low-pass filter.

Disadvantages of Band-Pass include high-order “organ pipe” resonance for the port, which determines the cutoff of the upper frequency values, as well as the complexity of the design.

The Band-Pass design is very sensitive to the Q value of the speaker. 4th order designs work best with speakers that have Qts close to 0.4, and 6th order designs work best with speakers that have Qts close to 0.5. In general, the higher the Qts, the narrower the frequency band. The lower the Qts, the wider it is, but at the same time the unevenness of the characteristics in the operating frequency band also increases. The Vas and Cms coefficients do not have much influence on the design.

3. Acoustic design with a passive radiator (emitter)
A passive radiator (similar to a regular speaker, but without the magnet system and voice coil) acts as the housing port. For this reason, a passive radiator enclosure in many cases behaves similarly to a bass reflex enclosure.

The advantages of a case design with a passive radiator are the same as those of a case with a bass reflex, plus the possibility of using a smaller case, which, however, cannot always accommodate a port of the required size. This ensures that the re-radiation of internal noise from the housing is minimized and the amplitude of the speaker cone is reduced in the area below the system resonance. The latter advantage results from the passive radiator's ability to support the speaker load at very low frequencies.

The disadvantages of the cabinet design with a passive radiator include, as one might expect, the disadvantages of the cabinet with a bass reflex plus poor transient response at the resonant frequency of the passive radiator (Fp). A passive radiator typically requires greater linear movement of the cone than a woofer. The complexity of the design is, of course, also a disadvantage.

4. Closed box
The advantages of the closed enclosure design are its simplicity and usually small size. Deviations in speaker characteristics often have less impact on sound quality. A flatter amplitude response and the ability to use with high-power amplifiers (since the speakers are not unloaded at low frequencies, as happens when working with bass reflex enclosures) are also a plus.
The disadvantages of the closed enclosure design are lower efficiency than when using a enclosure with a bass reflex. Typically, in a closed design, speakers with a quality factor of more than 0.3, a low Fs value and high Xmax and Vas values ​​perform well. Reducing the volume of the box will require lower values ​​of quality factor Qts and Vas.

Insufficiently damped system (box volume is small) and overdamped system (box volume is large)

Parameters of acoustic boxes used in calculations.
Vb- Internal volume of the box.
F3- Nominal frequency (Hz) at half power -3 dB. It is a point located 3 dB below the knee of the amplitude characteristic, at which the frequency response begins to decline in the low-frequency region.
Facebook- Resonant frequency for a case with a bass reflex (Hz).
QL- The value of the quality factor for the case is the sum of all losses. Cases with a volume of less than 11 cubic feet (311 liters) typically have a QL value close to 7. Cases with larger volumes have a QL of approximately 5.
Vap- A volume of air having an elasticity equivalent to that of a passive radiator suspension (cubic feet or inches, or liters).
Fp- Natural resonant frequency of the passive radiator (Hz).
Qtc- Quality factor value for a closed type housing.
Dv- The diameter or cross-sectional area of ​​a port or duct in a bass reflex housing.
Lv- The length of the port or duct in a housing with a bass reflex.

Charts


In this program you can access six graphs of various characteristics. These are the graphs:
- normalized amplitude-frequency response (often called frequency or amplitude response),
- amplitude characteristics when a signal of 2.83 V is applied to the input,
- maximum sound power,
- characteristics of voice coil resistance, phase and group delays.

Special Note
This remark concerns the transfer function of the car interior. The peculiarity is that the calculated amplitude-frequency characteristics of the system, displayed by the resulting graphs, very seriously depend on the specific car (size, design, etc.) in which the entire bass speaker system will be placed. The above graph demonstrates that the interior of a car leads to significant changes in the frequency response with the release of a “hump” at frequencies in the range of 30-50 Hz. The issue of the transfer function of the cabin was considered in "Master 12 Volt" N 1/98, and the experimental measurement results are presented in the next article in the same issue of the magazine.
Description of the chart: Characteristic "hump" due to the transfer function of the cabin
In most calculation programs, the transfer function is assumed to be some kind of universally averaged one, and SPEAKERSHOP is no exception in this regard. Although a point-by-point input of the transfer function measured experimentally is provided. The option of using experimental data can significantly increase the accuracy of calculations. Well, if there is no such data, then in the question of what will happen to the amplitude-frequency characteristics of the bass in various car models, Their Majesties the Experience and Intuition of the installer come first.

Crossover Module
This software allows you to calculate two- and three-way passive crossover systems from first (6 dB/oct) to fourth (24 dB/oct) order and a number of filter types: Bessel, Butterworth, Chebychev, Gaussian, Legendre, Linear-Phase and Linkwitz-Riley.

As a result of the calculations, an electrical diagram of the crossover system selected by the user will appear on the monitor screen, indicating the exact characteristics of its elements.

Question answer
[Q] I found a large speaker with no markings on occasion. How do you know if you can make a subwoofer out of it?
[A] It is necessary to measure its T/S parameters. Based on these data, make a decision on the type of low frequency design.

[Q] What are T/S parameters?
[A] Minimum set of parameters for calculating low-frequency design, proposed by Till and Small.

Fs - resonant frequency of the speaker without design
Qts - total quality factor of the speaker
Vas is the equivalent volume of the speaker.

[Q] How to measure T/S parameters?
[A] To do this, you need to assemble a circuit from a generator, a voltmeter, a resistor and the speaker under study. The speaker is connected to the output of the generator with an output voltage of several volts through a resistor with a resistance of about 1 kOhm.
1. We remove V(F) = frequency response of the speaker resistance in the resonance area. The speaker must be in free space (away from reflective surfaces) during this measurement. We find the resistance of the speaker at direct current (useful), write down the resonance frequency in the air Fs (this is the frequency at which the voltmeter readings are maximum:), the voltmeter readings Uo at the minimum frequency (well, for example 10 Hz) and Um at the resonance frequency Fs.
2. Find the frequencies F1 and F2 at which the V(F) curve intersects with the level V=SQRT(Vo*Vm).
3. Find Qts=SQRT(F1*F2)*SQRT(Uo/Um) / (F2-F1) This is the total quality factor of the speaker, one might say, the most important value.
4. To find Vas, you need to take a small closed box of volume Vc, with a hole slightly smaller than the diameter of the diffuser. Place the speaker firmly against the hole and repeat the measurements. These measurements will require the resonant frequency of the speaker in the Fc enclosure. We find Vas=Vc*((Fc/Fs)^2-1).
This technique was written in Audio Store 4 in 1999. I haven’t tested it.. There are others where the mechanical parameters of the head, mass, flexibility, etc. are measured.

[Q] I now have speaker parameters, what should I do with them?
[A] When designing each speaker, it is tailored to a specific type of acoustic design. To find out what exactly it is for, let's look at the quality factor.

Qts > 1.2 are heads for open boxes, optimally 2.4
Qts 85 bass reflexes
Fs/Qts >105 Bandpasses (bandpass resonators)

Elasticity, meatiness, dryness and other similar characteristics of the sound produced by a bass speaker are largely determined by the transient response of the system formed by the speaker, the woofer design and the environment. In order for this system to avoid overshoot in the impulse response, its quality factor must be less than 0.7 for systems with radiation from one side of the speaker (closed and bass reflex) and 1.93 for two-way systems (screen and open box design)

[Q] Where can I read about open design?
[A]Open drawers and screens are the simplest type of design. Advantages: ease of calculation, no increase in resonant frequency (only the type of frequency response depends on the size of the screen), almost constant quality factor. Disadvantages: large size of the front panel. Quite competent and simple calculations for this type of design can be found in V.K. Ioffe, M.V. Lizunkov. Household acoustic systems, M., Radio and communications. 1984. And in old Radios there are probably primitive amateur radio calculations.

[Q] How to calculate a closed box?
[A] Closed box design comes in two types, infinity screen and compression gimbal. Getting into one category or another depends on the ratio of the flexibility of the speaker suspension and the air in the box, designated alpha (by the way, the first can be measured, and the second can be calculated and changed using filling). For an infinite screen, the flexibility ratio is less than 3, for a compression suspension it is more than 3-4. As a first approximation, we can assume that the heads are sharpened with a higher quality factor for an infinite screen, and with a lower quality factor, for a compression suspension. For a pre-installed speaker, a closed enclosure like an infinity screen has a larger volume than a compression box. (Generally speaking, when there is a speaker, the optimal housing for it has a uniquely defined volume. Errors that arise during parameter measurements and calculations can be corrected within small limits by filling). Closed box speakers have powerful magnets and soft surrounds, unlike open box speakers. Formula for the resonant frequency of a speaker in volume design V Fс=Fs*SQRT(1+Vas/V), and an approximate formula connecting the resonant frequencies and quality factors of the head in the housing (index “c”) and in open space (index “s”) Fc/Qtc=Fs/Qts

In other words, it is possible to realize the required quality factor of the acoustic system in the only way, namely by choosing the volume of a closed box. Which quality factor should I choose? People who have not heard the sound of natural musical instruments usually choose speakers with a quality factor of more than 1.0. Speakers with such a quality factor (=1.0) have the least uneven frequency response in the low-frequency region (what does sound have to do with it?), achieved at the cost of a small overshoot in the transient response. The smoothest frequency response is obtained when Q=0.7, and a completely aperiodic impulse response at Q=0.5. Nomograms for calculations can be taken from the above book.

[Q] In articles about columns, words like “approximation according to Chebyshev, Butterworth,” etc. are often found. What does this have to do with speakers?
[A] The speaker system is a high pass filter. A filter can be described by a transfer characteristic. The transfer characteristic can always be adjusted to a known function. In filter theory, several types of power functions are used, named after the mathematicians who were the first to understand this or that function. The function is determined by the order (maximum exponent, i.e. H(s)=a*S^2/(b2*S^2+b1*S+b0) has a second order) and a set of coefficients a and b (from these coefficients you can then move on to the values ​​of real elements of the electric filter, or electromechanical parameters.) Further, when it comes to approximating the transfer characteristic with a Butterworth or Chebyshev polynomial or something else, this must be understood in such a way that the combination of the properties of the speaker and the housing (or capacitances and inductances in an electrical filter) is such that the frequency and phase characteristics can be adjusted to one or another polynomial with the greatest accuracy. The smoothest frequency response is obtained if it can be approximated by a Butterworth polynomial. The Chebyshev approximation is characterized by a wave-like frequency response and a greater extent of the working section (according to GOST up to -14 dB) in the region of lower frequencies.

[Q] What type of approximation should I choose for the bass reflex?
[A] So, before building a simple bass reflex, you need to know the volume of the box and the tuning frequency of the bass reflex (pipe, hole, passive radiator). If you choose the smoothest frequency response as a criterion (and this is not the only possible criterion), you will get the following plate
A) Qts 0.5 - you will have to allow waves on the frequency response, according to Chebyshev.
In case A) the bass reflex is tuned 40-80% above the resonance frequency. In case B) - to the resonance frequency. In case C) below the resonance frequency. In addition, in these cases there will be a different volume of the case. In order to find the exact tuning frequencies, you need to take the original formulas, which are cumbersome enough to present them here. Therefore, I refer those interested to the Audio Store for 1999, after this educational program it will be possible to figure it out there, or to Aldoshina’s books. And even Ephrussi’s articles in Radio for ’69 will do.