1. Open the signal generator.
2. Start the signal.
3. Minimize the signal generator window.
4. Open the oscilloscope.
5. Start the measurement.
6. Move the oscilloscope to the right bottom of the desktop.

1. Open the signal generator: Alt+S
2. Start the signal: Alt+S
3.  
4. Open the oscilloscope: Alt+O
5. Start the measurement: Alt+ S

Measurement procedures are explained below.

1) Using the oscilloscope

When the input and the output signals are to be compared by the 2ch mode, connect the Left channel of the soundcard's output directly to the Left channel input of the sound card (for measuring the signal generator's output), connect the Right channel output to the amplifier (for feeding the signal generator's output to the amplifier), and connect the amplifier's output to the Right channel input of the sound card (for measuring the amplifier's output). Then, start the signal generator and the oscilloscope. You can see two waveforms (one is the input signal, and another is the output of the amplifier) on the oscilloscope screen at the same time. By changing two inputs of the soundcard, any signals can be compared similarly.

In DSSF3, the screen resolution of the waveform display is much improved (x16 over sampling). So the smooth waveform can be observed up to high frequencies. Attention should be paid when the RA series (RAL, RAD, RAE) is used for measuring the signals above 4 kHz. 

See this application note for the measurement setup. 

2) Using the THD analyzer

THD analyzer can measure the distortion of the amplifier very effectively. In this case, connect the soundcard's output to the amplifier, and connect the amplifier's output to the soundcard's input. You don't need to connect the soundcard's output and input directly. THD analyzer compares the input and the output signals automatically and gives the result (distortion). 

See this application note for the measurement setup. 

The impulse response is measured by the Realtime Analyzer by using MLS (maximum length sequence) or TSP (time stretched pulse) signals. Measured data is analyzed by the Sound Analyzer to calculate the reverberation time and other acoustical parameters. Measured impulse response can also be saved as wav files. Analysis result in SA can be exported as a csv format. For the detailed instruction, see the impulse response measurement guide in RA's program manual.

As for equipments, a vocal microphone SM58 seems somewhat unsuitable for the measurement. Condenser microphones for music recording or a high quality measurement microphone might be better. The accuracy of the measurement strongly depends on the hardware to be used. We highly recommend the use of high grade soundcards (e.g. wide dynamic range and flat frequency response). Some of our recommendation can be found in Additional devices for sound measurement. Before the measurement, you can test your soundcard by using the THD analyzer (read the application note for how to test).

For saving and printing the measurement screen as picture files, we recommend the image database software, MMLIB. Measurement data and graph images can be easily stored to construct a measurement database.

See the application note: Measurement data management in Realtime Analyzer for more information. 

Generally, recordings saved to the RAM memory, is more readily available and more quickly processed than recordings saved to the hard disk. With the faster computers available today, however, the difference between accessing RAM and hard disk is negligible. It will be no problem in treating 10 minutes sound on the Celeron 1GHz PC.

In the recorder there is no way to save sound as a text file. You can use other software that can convert a wav file to a text file and other formats. We heard that several freeware and shareware can do this. 

In the octave band analyzer in RA ver.5.0.3.4 or later, time series of frequency and level data can be exported into a csv file. Impulse response data can be saved as a wav file or a text file. See the following application note for how to use it. Measurement data management in Realtime Analyzer

Our software can work with any type of sound card, including PC's built-in sound chip. So you can start the measurement just after installing the software. However, it is important to remember that the quality of the sound device is the most critical part of the measurement system. Important things are: measurable frequency range, wide dynamic range, low internal noise, distortion, and other artifacts. Measurement results are much affected by the performance of the soundcard.

For example, suppose that you measure the frequency response of the speaker system using a low quality sound card, and the result is a severe reduction in the high frequency range. A problem is that it is not clear whether it is a response of the speaker or the sound card. So, we recommend you to check the performance of your PC's sound device before the actual measurement. See the following page for how to check the sound device. Operation guide of the Realtime Analyzer, 2. Adjustment before a measurement

If the performance is not good and if you prefer better one, you can replace the soundcard and microphone. See the following page to know about the additional equipments. Additional devices for sound measurement

In RAD (also RAL and RAE), sound device can not be selected directly. Sound device is selected in the Windows volume control. Check the control panel > sound and multimedia property > Audio tab. The device should be selected as a "Preferred Device". Audio interface can be used in our software only if the device operates compatibly with Windows' sound mixer.

In DSSF3, the sound device (e.g. built-in soundcards and USB audio interface) can be specified directly in the RA's main window. Different devices can be specified for input and output separately. We have been reported that the USB device (e.g. TASCAM US-122) did not work with RAD, but worked with DSSF3. 

More important thing in the measurement is to reduce distortion. Measured signal would be distorted when the input/output volume is too large. In most cases, the most problematic device is the analog amplifier after the D/A converter. You should adjust volume carefully before measurement. This can be done by measuring the response of the soundcard using pink noise, sine sweep, and the spectrum. See this operation guide for how to adjust before measurement. Distortion of the soundcard can also be measured using the THD analyzer. 

Using the sound recorder, you can measure the wav files directly. It means you can measure digital signal itself without passing through DA/AD converter. 

Check whether the Windows volume control is working properly. RAD (RAL and RAE too) inputs/outputs signal through the Windows volume control. So, the input/output device (e.g. wave, CD, line) must be able to be selected on the volume control. If those devices don't appear on the volume control, soundcard driver must be reinstalled. After reinstallation of the driver, install the program again.

Another possibility is that the soundcard does not use Windows' sound mixer but use its own sound mixer. In that case, reinstallation of the driver does not solve the problem. Try our latest program, DSSF3. In DSSF3, soundcard can be selected directly. We have been reported from our users that RAL/RAD/RAE does not work with some latest soundcard, but DSSF3 does work. 

Before the installation, please remember to backup the DATA folder, in which the measurement data is stored, before the installation. Otherwise, your previous data will be deleted because the installer program overwrite the DATA folder.

After reinstallation, your registration information is left as it was. When the program requires a Registration Number after reinstallation, input the registration number we sent to you before. If this number is refused, please contact us.

When you measure the impulse response, you can find its inverse filter (i.e. reciprocal of the transfer function) by clicking the "Save" button on the impulse response measurement window. Green curve shows the phase spectrum of the system except that the sign is reversed. Phase is shown between -180 and +180 degree. See the operation guide for a step by step instruction. 

If you measure the binaural impulse response and analyze it by the Sound Analyzer, phase delay between two channels input (Left and Right) can be measured using the cross-correlation function (CCF). The maximum peak of the CCF indicates the phase delay and corresponds to the angle of incidence of sound. In the audio systems, it is important that the sound comes from a center position. This can also be checked simply by using the Correlation meter of the Realtime Analyzer. See this application note. Speaker phase check

Phase meter has been added to the FFT analyzer in DSSF3. This meter displays the phase difference between two channels as a function of frequency. To test the phase delay of the audio equipments, input the *monaural* pink noise as a test signal, and measure the two outputs to be tested by the phase meter. See this Q&A too. 

Measurement data management in Realtime Analyzer

As for RA's FFT analyzer, data export function is available in the octave band analyzer. Time series of frequency and level data can be recorded at a specified interval (minimum is 1 s) and saved as a CSV format. 

For other measurement, please use PrintScreen function.

1) open the measurement window
2) press Alt + PrtScr keys of the keyboard
at this point, the screen image is saved into Windows' clipboard
3) paste it to other software (e.g. MS paint, MS word) by pressing Ctrl+V keys
4) save the image into files

MMLIB is also a convenient tool to manage measurement data as picture files. See the following application for more instruction. Measurement data management in Realtime Analyzer

If you are planning to perform only a real-time measurement, DSSF3 Light would be sufficient. Especially, power spectrum, octave band analysis, and spectrogram can help visualizing sound.

2) You need a microphone amplifier and a sound interface. See the following pages.
http://www.ymec.com/hp/signal2/tool1.htm
http://www.ymec.com/manual/install/equipment_en.htm.

3) Digitally recorded sound can be analyzed, but it is important to use the recorder that is equipped with a manual recording level adjustment. It is also important to record a reference signal for calibrating the sound pressure level. See the following link for how to use a digital video for the sound measurement.  http://www.ymec.com/hp/signal2/basic3.htm.

1) Measure the impulse response using MLS signal.

2) Output wave file as "Input signal". This is a recorded raw data during a measurement without normalization. You can calculate the sound level from this signal.

According to your web site, I set the Windows audio properties (Hardware acceleration: None, and Sample rate conversion quality: Best), but it did not work. When I tried the same measurement by another old PC with SoundBlaster DigitalAudio II, clean impulse response was obtained. So, is a problem in a ONKYO's soundboard ? I thought that it is a high level soundboard, so I am puzzled. Please give me some advice. 

In the impulse response measurement, a test signal is generated at 70 % digital level. Your soundcard should not cause distortion at this digital volume. Input volume can be adjusted in the main window of the Realtime Analyzer. Output volume can be adjusted in the Windows' volume control. 

If the distortion still exists after the adjustment, try to measure the impulse response by using TSP method with Auto Level and Auto Retry off. In comparison with the MLS method, clean impulse response can be measured easily.

Another solution is to use another soundcard. We recommend USB sound interface, such as EDIROL UA-5. This device is easy to use and has a good performance. Also, the use of the external microphone amplifier with a PC's line-in connector may solve the problem. If you have another PC, try the same measurement so that you can compare the results. This helps to find the cause of the problem.

See the following application note for a detailed instruction of how to measure frequency response using a sweep signal. Sweep generator

By using the running ACF measurement, the fundamental frequency and the formant frequency can be analyzed as well as the speech analysis. If you would perform this analysis, DSSF3 Full-system is required. 

If you want to monitor snoring sound or any other noise while you are sleeping, Environmental noise Analyzer would be very useful. This module (available in DSSF3 Environmental noise) can record the sound automatically only when the sound level exceeds the specified trigger level. 

See operation guide of EA program manual for the step by step instruction of noise measurement by EA. 

However, it is no problem in the actual measurement. Once the input level is adjusted and saved in the calibration window, input volume is re-adjusted automatically when the calibration data is selected. See the operation guide for how to use this function. 

There are three volume controls in the Realtime Analyzer. The first one, "Input Volume" is accessed from the main window of the Realtime Analyzer. This volume is interlocked with the Window's recording control panel, and adjusts the input signal level before A/D converted. The second is the output volume, that is controlled by the signal generator's "Output Level". This volume is interlocked with the Window's volume control panel, and adjusts the level of the D/A converted signal. 

Third is the "Digital Output" on the signal generator. This is an original function of the Realtime Analyzer, that controls the amplitude of the digital signal (wave data) outputted from the signal generator. 

A procedure for the measurement of amplifier's phase characteristic is briefly described below.

1) Connect the left channel of the soundcard output to the left channel of the soundcard input directly. This becomes the reference.

2) Connect the right channel of the output to the amplifier's input, and connect the amplifier's output to the right channel of the soundcard input.

3) Generate a *monaural* pink noise as a test signal on the signal generator.

4) Start the phase meter from the FFT analyzer. The phase difference in each frequency is displayed. This is the phase delay of the amplifier.

Phase spectrum can also be obtained in the impulse response measurement. When you measure the impulse response of your system, setup the inverse filter by clicking the "Save" button on the impulse response measurement window. You can see the inverse filter of the system. Green curve shows the phase spectrum of the system except that the sign is reversed.

See the measurement report below for how to measure the impulses response.
Audio measurement report 2 http://www.ymec.com/hp/signal2/audio2.htm
Audio measurement report 3 http://www.ymec.com/hp/signal2/audio3.htm

After correcting the microphone's response, you can start adjusting the equalizer using the pink noise and 1/3 octave analyzer. Note that too low and too high frequency range can not be equalized perfectly. It is important not to boost the low frequency range. Keep in mind that between 200 and 1000 Hz should be flat. Between 2000 and 4000 Hz should be lower around 3-6 dB, because human ears are sensitive to these frequencies.

As for the performance of the A/D converter, you need to measure it before measurement. This is a very important test to obtain accurate measurement result.

You will see how to do it in this application note: Measurement of PC's soundboard by RA. Connect a CD player's output to a PC's line-in connector. 1/3 octave band analysis for a pink noise tells a frequency response of the PC's sound circuit. Oscilloscope display for a 1kHz tone tells about a distortion of the sound circuit.

See the measurement report http://www.ymec.com/hp/signal2/audio2.htm for how to measure the impulse response.

Time alignment of each unit of a multi-way speaker system can be adjusted using the impulse response. Impulse responses are measured for each unit at the listening position. The distance between the unit and the listening position is calculated from the time lag of the onsets of each unit response. To compensate time lag, each unit position is adjusted. 

In addition to the PC and the software, you need a microphone, amplifier, and soundcard for the measurement. See this page for information about the additional equipments.

When the "microphone" is selected as the input device, the input signal to the mic-in terminal is amplified by the built-in microphone amplifier and then converted to the digital signal. Generally, built-in microphone amplifier does not have a good performance for sound measurement (i.e. poor frequency response and high floor noise level). So we recommend using the "line-in" connector in the measurement.

When the line-in is selected, built-in microphone amplifier is not used. It means that the input signal itself will be measured. Line-in is supposed to be used to import signals from the audio equipments having the internal amplifier, such as CD player, MD player, and DAT player. So, an external amplifier is required when the line-in is used to input the microphone signal. For example, microphone amplifier, audio mixer, or portable DAT recorder can be used. 

If your PC does not have the line-in connector, PCMCIA or USB type audio devices are highly recommended. These devices have a high quality microphone amplifier and line-in connector. Using such a device, your PC turns into a high grade measurement equipment. More information can be found at Additional devices for sound measurement.

More important thing in the measurement is to reduce distortion. Measured signal would be distorted when the input/output volume is too large. You should adjust volume carefully before measurement. This can be done by measuring the response of the soundcard using sine sweep and the spectrum. See this operation guide for how to adjust before measurement. Distortion of the soundcard can also be measured using the THD analyzer. 

Microphone's frequency response can be compensated if you have its response data. Realtime Analyzer has a calibration function in which you can edit the frequency response curve for compensating the microphone's response. See the operation guide for how to do it.

If you need finer time resolution of the level data (i.e. order of 0.1 or 0.01 second), Running ACF measurement is available. This measurement is available in RAD and DSSF3 Full-system. 

Save all files in this folder. Then format your disk.

After you install the program, put the saved file in the same directory. You can use the previous data on new environment.

<*> If your software is RAD, this folder name is "RAD". Like this, folder name here is product name. 

Also see the Install guide for more information. 

However, if you format the hard disk, the registration will become invalid. A new registration number will be required, because the new hard key is created when you reinstall the program again. 

In such a case, we will provide a new registration number for special fee. Please let us know your name and your new hard key, and a reason you need a new registration. After checking your order information, we will inform you a new registration number. See the following page for how to find a hard key. http://www.ymec.com/store/en/register.htm

Also keep in mind that the measurable frequency will be affected by the performance of audio device and sound card. Generally, the audio device can measure signals in the audible frequency range of between 20 Hz and 20000 Hz. So, lower frequencies below 20 Hz will be much attenuated.

• Measurement of the noise level from the wav file.

The sound file (WAV format) can't be read directly in RAL/RAD/RAE. Use the sound player (e.g. Windows Media Player) to play sound and measure it by the FFT analyzer. In RAD, RAE, or DSSF3, wav file can be read and measured directly. See the program manuals of RA's running ACF measurement. Same measurement can be more easily performed in EA (Environmental noise Analyzer) ver.5.). See the EA's program manual.

For measuring the noise level, set the calculation conditions as follows. If you measure in RA, these settings should be done when the sound is analyzed later in SA (Sound Analyzer). 

1) Integration time: Set as 0.5 s to measure the noise level to emulate the sound level meter's FAST setting.

2) Running step: This is the calculation interval of the noise level. Set as 0.1 s. 

3) Weighting: Select the A-weighting to measure the level in dBA.

4) Max delay time: This parameter will not affect the calculation of the noise level. Set the short time (e.g. 0.01 s) to decrease the calculation load. (To analyze sound in more detail, this should be long enough (half of the integration time) to capture the lower frequency of sound.) 

• Calculation of Leq from the measured noise level.

The Leq is defined as "the steady sound level over a specified period of time that would produce the same energy equivalence as the fluctuating sound level". It means the averaged noise level measured for the time varying noise, and the SPL values that EA and SA calculate for every calculation step is exactly the Leq as shown in the following equation. The integration time T is the parameter that is specified in the calculation condition.

Calculating the Leq for longer period of time (e.g. 60 s), however, takes long calculation time and is not a convenient way. So the practical way to calculate the long time Leq is averaging the short time noise levels. To take an average, the noise levels measured in dB is first converted to the real values, averaged, and then converted back to dB again. This calculation is not available in the current version of DSSF3. So the calculation should be done in other software (e.g. MS Excel). To do so, measured data can be saved as CSV or TXT format. See the following instructions for how to do it.

1) Open the noise data measured by RA or EA. 

Measured data can be open by SA. If the noise was measured by EA, sound level has already been calculated. But it was measured by RA, calculation is now needed. Set the conditions above and start the calculation. For more information of SA, see the program manual. 

2) Data export.

 Select the "Output Files" from the "Parameter" menu on the main window of SA. Name the data and save as a CSV format.

From now, calculation is done in Excel.

3) Open the CSV file.

The data is shown as below. Use only the SPL values for the calculation.

・假ｼ韻onvert the SPL values (dB) to real values. Use the equation P = 10^(dB/10). 

9) Calculate the average for all values. Then convert to the dB again by dB=10*log10(P). The result is the Leq for a measured period. 

Also keep in mind that the measurable frequency will be affected by the performance of audio device and sound card. Generally, the audio device can measure signals in the audible frequency range of between 20 Hz and 20000 Hz. So, lower frequencies below 20 Hz will be much attenuated.

The upper limit is half of the sampling frequency. When the sampling frequency is set to 48 kHz, the upper limit becomes to 24 kHz. But note that the actual performance depends on the quality of the soundcard used.

Level Range decides a value of one grid of the vertical axis in the oscilloscope display. For example, when the Level Range is set to 1, one grid becomes 1. When the Level Range is set to 32768, one grid becomes 32768. Usually, this value is set automatically so that the signal is displayed in an appropriate range. 

If you want to measure the absolute voltage values, the input signal level can be calibrated by using the voltmeter. See the following application notes for how to do this. 

http://www.ymec.com/hp/signal2/probe2.htm

http://www.ymec.com/hp/signal2/probe3.htm

Click the reload button of your browser. You can see how the measurement window changes to an expanded graph.

* In this expanded graph mode, you can only start or stop the measurement with the Alt + S keys, although other buttons are not available.

For saving and printing the measurement screen as picture files, we recommend the image database software, MMLIB. Measurement data and graph images can be easily stored to construct a measurement database.

Measurement data management in Realtime Analyzer

Measurement data management in Realtime Analyzer

Second, a microphone is required for measuring sound. A microphone is connected to the "Mic in" connection of the computer. 

Third, you may need an external audio interface, because the input/output device of standard PCs are generally poor, that is not suitable for the audio measurement. You can find a lot of devices with PCMCIA and USB interface. 

You can find more information about the necessary equipments in this page. Additional devices for sound measurement

You can find more information about necessary equipments in this page. Additional devices for sound measurement

1. Connect a device to PC.
2. Plug in the power to the device and turn it on.
3. Start RAL.

You can find more information about the necessary equipments in this page. Additional devices for sound measurement 

 It is recommended to use microphones that are being sold as "calibrated", or "measurement" microphones. Such microphones are usually expensive, but these days you can find inexpensive microphones that have an excellent performance. The frequency response of such microphones might not be perfectly flat, but you can compensate its response if you have a frequency response data of that microphone.

For measuring a sound pressure level (SPL), you need a sound level meter or an acoustic calibrator for calibrating the input signal level to the computer. If you don't calibrate your system, you can measure only the relative sound levels. 

You can find more information about microphones and other equipments in this page. Additional devices for sound measurement 

Functions in the previous version of DSSF3 has been packed into the lower price of software and is released as RA series. For example, RAL consists of the most basic measurement capabilities including:

• signal generator
• peak level monitor
• power spectrum
• 1/3 octave analysis
• TEF (time-energy-frequency) display
• oscilloscope

Other advanced functions, such as the impulse response, ACF/CCF measurement, and environmental noise measurement are available in higher version of the program (RAD and RAE). 

New version of DSSF3 (version 5) has been released on February 2003 and almost all functions have been continuously upgraded to reflect users' request and to incorporate advanced measurement capabilities. Update history can be seen in DSSF3 News.

See the function list for comparing the available functions in each version.

Program manuals and user support documents can be found at http://www.ymec.com/store/en/manuals.htm

Even if you have few knowledge and experience of sound measurement, you can start the measurement by just clicking the "Start" button on the measurement windows. Complex settings can be ignored at first, because these are automatically set by the program. Signal generator, oscilloscope, FFT analyzer would tell you many things about sound. As a next step, follow the instructions in the operation guides and the measurement reports. You can learn how to use RA step by step.

Some of the application examples are briefly described below. 

1) measurement and diagnosis of motor and engine performance
It is said that some people can diagnose the performance of a motor only by hearing its operation sound. In our experiment, we tried to measure the motor performance using the power spectrum and the autocorrelation function, that represent the rotation speed (rpm) or the rotation stability. The motor sound actually changes by changing the power supply or by cleaning up fans. Read more...

The same measurement might be applied to the car and motorcycle engines. The exhaust note is very important for maintenance and tuning of the race car, because the sound directly reflects the condition and the performance of engines. Rotation speed and vibration frequency can be measured from the engine sound. Read more...

2) measurement and diagnosis of PC's soundcard
In any kind of measurement, it is very important to know the quality of the measurement system itself. When using the PC based measurement system, the most critical part is the sound card. Important things are: measurable frequency range, wide dynamic range, low internal noise, distortion, and other artifacts. It is important to know these performance before the measurement. It can be tested using the sweep tone and the spectrum analyzer. Any problem of the device, such as the distortion and the oscillation can also be revealed by the oscilloscope and THD analyzer. Read more...

3) noise measurement in a living room and a bed room
We are surrounded by many noises in our home. Almost all of the electrical appliances, for example, air conditioner, refrigerator, cleaner, emit noise. These sounds are not so loud, but are sometimes very annoying because the specific frequency component always exist. It might be caused by a mechanical breakdown or some accidents of the appliances. It can be checked using the 1/3 octave analyzer. Of course, the sound from outside, such as a traffic noise or a construction noise, can be investigated. Read more...

4) test and adjustment of home audio and car audio
Real-time analyzer (RTA) is originally used for the audio measurement. Basic measurement is the 1/3 octave band analysis of the pink noise response of the loudspeaker output. Combination of the sweep signal and the power spectrum is also a powerful measurement tool. It can reveal the distortion and noise very clearly. Read more... 

5) analysis of speech and singing voice, musical instruments
RAL can be used for tuning instruments, measuring frequency components of speech, or analyzing voice quality. Realtime spectrogram (note that this is a new function of DSSF3) is a valuable tool for a voice training and a speech therapy. 

6) Sound measurement for medical diagnosis
It is widely recognized that auscultation is a very effective method in detecting heart diseases and lung diseases in clinical diagnosis, because the auscultation sound contains much useful information about the activity of inner organs such as heartbeat, blood flow, breathing. We are trying to record the medical sounds using a stethoscope microphone and analyze its characteristics. Read more...

You can try measurement by yourself. Note that almost all functions of RAL have been upgraded and many advanced features have been newly added in DSSF3. For example, THD analyzer, Spectrogram, Phase meter, and Sound recorder are very useful in the audio measurement and analysis. Update history can be seen in DSSF3 News. For comparing the available functions in each version,  function list is available.

Q1. Which software should be purchased for the purpose of the following measurement?

1. The frequency characteristic both at 1m from a speaker and a listening point.
2. Reverberation time of a listening room for each frequency range. Time until sound pressure is halved in an impulse response
3. Time alignment of each speaker unit in the crossover frequency of a multi-way speaker. The purpose is to make a phase gap minimum. I plan to adjust a unit position to time alignment by the following method. 1) Impulse responses are measured on the crossover frequency of each unit at the listening position. 2) The distance between units is calculated from the time lag of the onsets of each unit response. 3) To compensate time lag, a unit position is adjusted. Is this method right? Or are there other methods?

Q2. The personal computer to be used is IBM ThinkPad X20, which has a stereo microphone input, a stereo line input (input sensitivity is adjustable), a stereo headphones output, and a stereo line output. Also, the comparatively high quality microphone amplifier and mixer are available.

In your Q&A site, it is said, "Simultaneous output and input is possible if the Sound Blaster compatible A/D and D/A can be used simultaneously. A PCMCIA correspondence sound card can also be used." Is a duplex type sound card required for measurement of the impulse response time as in questions 2 or 3 above?

1. The frequency characteristic both at 1m from a speaker and a listening point ... Measurement of the frequency response can be performed easily by using the pink noise and the octave band analyzer. The test signal can be generated by the signal generator of the Realtime Analyzer. 

2. Reverberation time of a listening room for each frequency range...  Impulse response measurement is performed by the Realtime Analyzer using MLS (maximum length sequence) or TSP (time stretched pulse) signal. Measured impulse response is then analyzed by the Sound Analyzer. For example, the arrival time of the direct sound from a speaker unit and the delay time of the reflective sound is obtained. The reverberation time and the inter-aural cross-correlation can be measured for 1/1 or 1/3 octave frequency band.

3. Time alignment of each speaker unit in the crossover frequency of a multi-way speaker... The method you mentioned seems right. Another possible way is to use a phase meter. 

A2. Full duplex type soundcard is necessary to perform the impulse response measurement for simultaneous output and input. 

The problems in designing the crossover network are described in Don Davis, "Sound system engineering". Some important points are quoted below.

1. Choosing the crossover frequency that match the directivity of driver units.

2. Choosing the proper filter slope that
  a) keeps high frequency transient distortion at a low level, and
  b) reduces the low frequency distortions.

3. Matching the phase characteristics of the crossover network and the driver. 

4. Matching the impedance of the crossover network and the driver.

We recommend DSSF3 Light for this kind of measurement. Internal signal generator and the spectrum analyzer might be enough. For a more advanced measurement and adjustment, DSSF3 Realtime Analyzer is recommended. This version includes the distortion analyzer and the impulse response measurement function.

In the measurement, you need to output test signals such as pink noise, 1 kHz tone, and sine sweep from your audio system. It is convenient to record these signals on a CDR. Set the microphone at the listening position, replay the CD, and measure sound. Adjustment of the equalizer, amplifier's volume and valance, positioning of the loudspeakers can be performed in real time. 

Before the actual measurement, check and adjustment of the measurement equipments (PC and microphone) is necessary. See the RA's program manual for detailed instruction. If the performance is not good and if you prefer better one, you can replace the soundcard and microphone. See the following page to know about the additional equipments. Additional devices for sound measurement

It is good to measure in a quiet room. Set the microphone close to the watch and increase the input level. If these microphone don't work well, another choice is to use a contact microphone. For example, we have made a stethoscope microphone to record a heart sound. 

However, we think you don't need to be too sensitive about the flat frequency response of the microphone. Usually, microphone's response is attenuated in low and high frequency ranges. It is no problem in measuring the mid frequency range, that is important for adjusting the channel divider's crossover frequency and equalizer gains. If you are using such an uncalibrated microphone for equalizing audio systems, replay pink noise and measure 1/3 octave response without compensation. Then, adjust your system until the 1/3 octave response becomes a smooth curve without sharp peaks. 

 L_total=10*log10(10^(L1/10)+10^(L2/10)+10^(L3/10)+...10^(Ln/10)).

For example, total noise level of two sounds with levels of 90 dB becomes 10*log10(10^(90/10)+10^(90/10) = 10*log10(2*10^9) = 93 dB. 

SPL of the wav files can be more easily measured using DSSF3 Light. See the following application notes. 

Measurement data management in Realtime Analyzer

Operation of a Sound-level Meter 2

Because the duration of the footstep noise is very short, it is difficult to measure it by the sound level meter with a usual setting. To measure such a sound, very short integration time of sound signal is needed. EA (Environmental noise Analyzer) of RAE and DSSF3 can measure it, by setting the integration time short, e.g. 10 ms. 

EA is also equipped with the automatic measurement function, that start the measurement automatically when the sound level exceeds the specified trigger level. This function can be used if the level of the background noise and the footstep noise are well separated. Some of our users are monitoring noise while sleeping. Even the averaged SPL of such an impact noise is low (e.g. 40-50 dB), its instantaneous level is sometimes very high (e.g. 70-80 dB). This corresponds to the car noise. Also, its relative level becomes high, because the background noise is generally low at night. So, it is very annoying and sometimes very dangerous for human health. 

For how to use EA, see the EA's operation manual. Step by step instruction is available.

Basically, noise measurement is performed by connecting the AC output of the sound level meter to the soundcard's input. Other microphones than the sound level meter can also be used as input. But in this case, the input level has to be calibrated using the acoustic calibrator. In any cases, calibration is necessary to measure the absolute sound pressure level. If the calibration is not done, you can measure only the relative sound level. 

See the operation guide in the RA's program manual for microphone calibration and how to use the sound level meter. See Additional devices for sound measurement for more information about the equipments.

Measurement software significantly extend the function of the standard sound level meter. For example, we can save the measurement results, perform the frequency analysis (octave band analysis), and so. Even the lowest version of DSSF3 (DSSF3 Light) is equipped with the data record function that can save the octave band analysis data as a csv file. DSSF3 is also equipped with the standard time constant setting (FAST/SLOW) and the 1/1 octave band analyzer. From the measured data, averaged noise level, min, max, or equivalent noise level (Leq) can be calculated easily.

To perform a long time measurement, EA (Environmental noise Analyzer) is recommended. For example, EA can perform the automatic measurement when the sound level exceeds the specified trigger level. The noise is saved as a sound file with the calculated acoustic parameters. The data can be further analyzed in SA (Sound Analyzer). EA is available in RAE and DSSF3 Environmental Noise.

EA can also measure impulsive noise, which is difficult to measure using a standard sound level meter because its duration is very short. To measure such a sound, very short integration time of sound signal is needed. EA can measure it very accurately, by setting the integration time short, e.g. 10 ms. 

RAD or RAE can load a wave file directly from the running ACF window. In this case, you don't need to worry about distortion. A wav file itself can be measured. Also, it is easy to play the loaded file and analyze it by the FFT analyzer simultaneously. DSSF3 is equipped with the sound recorder. It is much easier to analyze wav files than the previous version. We recommend DSSF3 Light for analyzing wav files. 

To measure two signals from different input terminal at the same time, you should connect one input to the Left channel, and another input to the Right channel of the same input terminal (e.g. Line-in). See the application note of the oscilloscope for the measurement setup. 

If you have an external mixer device, input signals can be switched easily. 

1) Open the FFT analyzer in RA (Realtime Analyzer), select the "microphone" as the input device, and click the start button. Sound is captured from the PC's built-in microphone, and the measurement starts immediately. Detailed settings can be ignored at the beginning. 

2) The measurement window can be captured by clicking the "Screen Copy" button. Image database MMLIB can also be used to manage the measurement data. 

3) Numerical data can be saved on the octave band analyzer. Click the "Data Record" button to open the data record dialog. 

As for the additional equipments such as the external microphone and the soundcard, see Additional devices for sound measurement. 

For example, balance of high and low frequencies can be easily checked by the octave band analysis using pink noise. Crossover point of the network can be checked more in detail by the power spectrum of the white noise or sine sweep. If you change some settings, it will be immediately reflected in the measurement results. You can adjust your system by seeing the results in real-time. 

Note that the output signal from the speaker is greatly influenced by the placement and positioning of the speaker. Perhaps you can find best positions by seeing the measurement data and listening to the sound. 

Single sine tone can be used to test the distortion or signal leak of the amplifier by observing the output signal on the oscilloscope or the spectrum analyzer. 

Basically, speaker's output sound is measured by the FFT analyzer for checking the performance of speaker network. But in this case the results are affected by the characteristics of the speaker unit. So, we need a high quality monitor speaker unit. As an alternative way, the electronic measurement stated above might be available.

See the following application note. Singing voice analysis by the spectrogram, Analysis of Japanese voice, Measurement data management in Realtime Analyzer

Basically, human voice consists of several harmonic frequencies. If the voice changes, it will appear as the change in the spectrum. We think that if you store results for each patient continuously, you can compare their spectra and find changes. 

If you need more precise measurement, you can use the Sound Analyzer. This module is available in DSSF3 Full-system. Some important acoustic parameters can be analyzed with a high temporal resolution. For example, time history of the sound levels, pitch, and pitch strength are important factors in determining the perceived voice quality. See Introduction to the autocorrelation function for more information.

Our software is designed to work automatically. Even if you have few knowledge and experience of sound measurement, you can start the measurement by just clicking the "Start" button on the measurement windows. Complex settings can be ignored at first, because these are automatically set by the program. Signal generator, oscilloscope, FFT analyzer would tell you many things about sound. As a next step, follow the instructions in the program manual and the application notes. You can learn how to use the Realtime Analyzer step by step.

C:\>taskkill /F /IM pcaudio.exe"

We have been reported some USB audio devices didn't work with RAL/RAD/RAE, but worked with DSSF3. So, we recommend you to try DSSF3.

Please remember to backup the DATA folder, in which the measurement data is stored, before the installation. Otherwise, your previous data will be deleted because the installer program overwrite the DATA folder.

After reinstallation, your registration information is left as it was. When the program requires a Registration Number after reinstallation, input the registration number we sent to you before. If this number is refused, please contact us by email.

Please remember to backup the DATA folder, in which the measurement data is stored, before the installation. Otherwise, your previous data will be deleted because the installer program overwrite the DATA folder.

After reinstallation, your registration information is left as it was. When the program requires a Registration Number after reinstallation, input the registration number we sent to you before. If this number is refused, please contact us by email.

The update information of the DSSF3 can be found at the DSSF3 news. 

FFT size: Integration time (decided in the calculation condition window)

percentage of overlap: Running step (decided in the calculation condition window too)

window function: rectangular

The FFT size and the overlap size is set as time (in second), so the actual FFT size in sample is Integration time * sampling rate. The overlap size in percent is (Integration time - Running step) / Integration time. The data portion that is used for the calculation is indicated by the blue area in the waveform display. The figure below shows the example, in which data was calculated with the integration time of 0.5 s and the running step of 0.1 s. The waveforms at 0.1 s and 0.2 s are shown here for illustration.

The procedure is summarized in How to purchase the license. 

We are trying to send registration numbers from ymec.com to avoid such problems.

Check the Windows' volume control panel. If the recording device such as mic or line-in does not appear on the panel, the device does not work with RAE (also RAD and RAL). The reason is that the sound device does not use Windows' sound mixer but use its own sound mixer.

In the case of UA-3, we have been reported that Windows volume control is not necessary, because the input volume is controlled by the hardware volume of UA-3. In that case, the measurement has become available after selecting the "USB audio device" as the Preferred device on the Sound and Multimedia control panel. The signal come from UA-3, although the "Input Device" column was blank and the Input Volume could not be controlled from the program.

However, we have been reported from many users as for the problem of Edirol products. It seems UA-1X can't be used for sound input. UA-3 is also problematic in many cases. UA-5 works fine. We can't say about other products. 

We recommend you to try DSSF3, because we have been also reported that some USB device did not work with RAE, RAD, and RAL, but worked well with DSSF3. 

If you use a digital video camera or a DAT recorder, it is important that the recorder has a manual recording level control. In addition, you need a sound level meter or an acoustic calibrator for measuring the sound pressure level (SPL). Without calibration, you can measure only the relative sound intensity. For how to perform a measurement, see the following application notes.

DAT and digital video recording for outdoor measurement

Operation of a Sound-level Meter 2

Additional devices for sound measurement

Measurement equipment example

Even DSSF3 Light can perform the spectrum and the octave band analysis. But we think you need Sound Analyzer for your measurement, because you measure a lot of items. It is more important to construct a measurement database than the measurement itself. It may be difficult to manage the measurement data by DSSF3 Light. 

Cooperation of Realtime Analyzer and Sound Analyzer helps multidimensional sound analysis by use of various acoustic parameters. In addition to the conventional FFT analysis, a縲€newly developed ACF / CCF is used for analyzing several acoustic qualities and spatial qualities of sound. Analyzed information is stored in the huge database, and is smoothly displayed in an intelligible set of graphs. This is the original and the advanced functions of DSSF3.

For saving and printing the measurement screen as picture files, an image database software, MMLIB is also recommended. Measurement window and graph images can be easily stored to construct a measurement database.

Realtime Analyzer http://www.ymec.com/manual/era/index.htm

Sound Analyzer http://www.ymec.com/manual/esa/index.html

Environmental noise Analyzer http://www.ymec.com/manual/esa/index.html

We are receiving a lot of request for pdf manuals, but at the moment only the online manual (html version) is available. 

We found that GoldWave Digital Audio Editor http://www.goldwave.com/ can do what you want. You can open a raw data and then save it as a wav format.

Instmsia.exe for Windows 98 or ME:

Instmsiw.exe for Windows NT, Windows 2000 or Windows XP:

[Back] [Up]

1) Impulse Response
Open the Impulse Response window from the Realtime Analyzer. Click "Load" button. Folder and Measurement data are displayed. To delete a specific data in a specific folder, select the folder and then select the data, and click the "Delete" button. To delete a whole folder, select the folder and click the "Delete" button. 

2) Running ACF
Open the Running ACF window from the Realtime Analyzer. The following procedure is same as above.

3) Noise Measurement
There is no way to delete the specific data of this type. Data is stored in the files "NMS.BIN" and "NMS.DBF" in the DATA folder in the program folder. If you don't need whole data, you can delete these two files. But if you need some of data, you have to backup these files to the another hard disk.

Finally, when you want to clear all the data, delete the DATA folder. After this operation, the new DATA folder is created so that you can start the measurement again. If you are using DSSF3, setting utility (DSS.exe) is available for the management of the measurement data. See the following page for more information. Measurement data management in Realtime Analyzer (Introduction to Sound Measurement)

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]

[Back] [Up]