Support

First line support is provided by your local distributor.

noiseLAB support is available by email:

noiseLAB@delta.dk.

Upgrading from Version 2.0 to 2.1

Upgrading from Version 2.1 to Version 3.0:

• If you desire, you may leave version 2.1 installed on your PC.
• Version 3.0 consists of two programs: noiseLAB Classic (an enhanced version of version 2.1) and the Batch Processor for more advanced applications.
• First install noiseLAB Classic, and then the Batch Processor.
• If you use National Instruments Hardware, also install NI-DAQmx 8.5 (or later). noiseLAB 3.0 may require these newer drivers if you previously have installed DAQmx. Click here to download from National Instruments.
• NOTE: If you use noiseLAB 3.0 to analyze files recorded by noiseLAB 2.1, these will be modified and can no longer be read by noiseLAB 2.1. Therefore, you should make back-up copies of noiseLAB 2.1 files.
• Please read the instruction manual included with noiseLAB 3.0.

noiseLAB 3.0 FAQ

New Features: noiseLAB 3.0 has four significant new features:

• Tone Analysis: Automatically identify tones and their audibility according to the new ISO 1996-2 Standard.
• Batch Processing: Analyze a large number of clips with any or all measurement functions in parallel.
• .wav import: Import .wav files from external devices for subsequent analysis
• Train Analysis: Analyze time varying signals in consecutive "Train Cars".

Tone Analysis

1. What is so "cool" about tone analysis?

Visually, the tone analysis is cool. We think you'll like the selection of colors, the interactive cursors and reporting options. (And we also think it is cool that the new ISO 1996-2 standard for Tone analysis was developed by DELTA).



But more seriously, tone analysis automates what used to be a fairly complex manual process.

The basic steps performed automatically by noiseLAB are:

• Perform FFT analysis of the A-weighted time signal with a ~3 Hz bandwidth averaged over at least 60 seconds.
• Search through the spectrum to find "peaks" that fulfill the tone criteria of Slope Change and level above background noise.
• Slide Critical Bands through the spectrum and identify all combinations of tones inside a critical band.

In the above example you can see that there are five combinations of critical bands that include the three tones in various combinations.

• Now for each critical band, curve fit the background noise level to find Lpn (perceived noise level) assuming no tones are present in the selected critical band.
• Correct for the ear's masking characteristics.
• Compute the Lta, the tonal audibility level.
• Sort the Critical Bands by tonal audibility. The first critical band in the Figure above has the highest Lta, hence is the most annoying.
• Correct the sound level with the KT correction factor or penalty factor due to audible tones.

2. Is Tone Analysis fully automatic?

Yes. You simply select the Clips that you want to analyze, and the tones are automatically identified.
3. But the standard also says you should be able to do it manually.

Yes, this process is also described in the standard, but you really have to know what you are doing. Among other things you have to do a linear regression analysis to identify the noise in the critical band. (The narrow black box in the above screen shot shows the results of this done automatically by noiseLAB).


Note in the above example, the sliders to the right of the graph control the S (Slope) and O (Offset) of the critical band indicator. when you press "Save" the new tone values are stored and the Table of Critical Bands below the graph are automatically re-sorted by tone audibility. Note also that the Tone values in the table are now identified as being computed by the manual method.

4. What parameters should I set for tone analysis?

Typically you should set the FFT Line spacing in the range of 2 to 3 Hz, and the Delta Slope to 1 dB/line. Averaging time should typically be at least 60 seconds or so. If you use shorter averaging times, the spectrum will be very "noisy" or ragged in its appearance, and this may result in the detection of false tones. To prevent this, you need to increase the Delta Slope parameter to 2 dB or more.


Batch Processing:

1. Batch processing sounds like a complicated, not very cool thing. . . how does it help me?

2. It actually makes life a lot easier. Assume you have sound Clips. With batch processing you can analyze any or all of the clips at one time, fully automatically.

3. But isn't it complicated to set-up?



Just CTRL Click (as is normal in Windows) to select the individual Clips you want to process. (Use Shift Click to select a Range of Clips) Then hit the "Analyze" button, and you get the results.


4. Do I then have to run a new Batch process for each analysis type, or can I do FFT and Octave at the same time?

You can do any or all of the following in parallel for Batch Processing

Sound Level

Min Sound Level

Peak

Impulse

Takt Max (3 and 5 s)

L1, L5, L10, L50, L90, L95, L99

FFT

1/N Octave

Tone Analysis

So, for example, if you have 20 Clips, you can perform all of these analysis in one easy process.

5. And obviously, I can dump this data to spreadsheet with just one click?

Yes!

6. And if I want to store a graph, is it still one click?

Yes.
Train Analysis:

7. But I understand I also have this function called "Train Analysis". Can that be combined with the above processing?

Yes. Assume you want to break each Clip into "Cars" of 15 second duration. (See above Figure). Then all of the above measurements will be made for each Car in Each Clip.

8. That sounds a bit complicated - - -

Let's take a simple example.

Assume you have 10 Clips, each 5 minutes long. You then define a Car Length of 15 seconds, which gives you 4 cars/s x 5 s =20 Cars per Clip. And with ten Clips this becomes a total of 200 cars, for which all of the above measurements can be computed for each Car. So, for example, you would have 200 1/3 octave spectra, or 200 FFT Spectra.

9. And what can that be used for?

Time varying/non-stationary signals. You can see how tones evolve as a function of time, or more simply, how the sound level changes over time.

10. Which reminds me of the the classic level recorder plot of sound level. How does noiseLAB do that?

With modern computers we no longer have long rolls of level recorder paper, so we have to break the "roll" up into many individual graphs. This is where Train Analysis comes in handy.

First, set you Car Length to equal the desired duration of each graph (for example 30 seconds).

Next you perform the Analysis.

Note that the table that shows the Sound Level related values for each train car, and the graph shows the plot of Sound Level.

11. Which means that I have to save the graph individually for each of the 15 "Train Cars".

No, when Train Analysis is Active, the "Graph" button automatically saves the plots of all Cars.

and so forth up to the last Car

12. Sure that is really nice, but if a just want a full length plot of the entire Clip, do I then have to run a new analysis without "Train Analysis"?

No, simply hit the "Total Cars" button above the Graph, and the total plot will be shown.

13. But I prefer a white level recorder background, and an X axis with the actual time, not the relative time as shown above.

That's easy. Just Right Click on the upper left hand part of the graph to select a white background, and click on the X Axis to select an "Absolute" time scale.

Wave Import FAQ


1. Convince me, .wav file format has been around for ages. What's so cool about that?


Flash memory sizes have exploded in the last year. Today you can get 4 GByte flash for under USD 100! This means that the sound level meters that support this now can record up to 12 hours of raw time data! That means your portable tape recorded suddenly has moved into the SLM (The maximum size of memory card the various Sound Level Meters support varies from manufacturer to manufacturer).

All you do is simply import the .wav file into noiseLAB! And all of its processing capabilities are available, at a market leading price/performance.

2. That is quite cool . . . I had never really thought about a Sound Level Meter as a tape recorder before....

Good point. And now that you get that capability, you perhaps need to think about how much you want to invest in an advanced sound level meter with all its functions. From a convenience standpoint, it may well be very nice to have your sound level meter do all your processing for you. But from a price and performance standpoint, you may be better off doing the analysis externally in noiseLAB.

From our experience, for simple measurements, the sound level meter is the most convenient. For more complex measurements, noiseLAB is the preferred solution.

3. But noiseLAB also can function as a tape recorder?

Yes, with a suitable external digitizer such as the NI USB-9233, you get a great portable solution with up to 4 channels. (Desktop computers with PCI cards support up to 8 channels).

4. What are its advantages over a sound level meter or external tape recorder?


Keeping track of your data!!!

We have worked extensively with users making complex measurements in the field. Typically, they use multiple microphones and one or more hard disk recorders. The book-keeping of all the different channels, calibration histories, identification of noise events, is complex, and error prone. With the integrated solution of noiseLAB, the house-keeping is significantly simpler, much less error prone, and much faster.

5. Any things I should be careful with in terms of .wav import?

Yes! The .wav format requires that you make sure that you have recorded in a format with a high enough sampling rate, number of bits, and compression free! In addition you must be careful to transfer you calibration information to noiseLAB.

You should also be aware that if you use an external piece of audio editing software, then the calibration chain is broken since non-of these programs have been certified from a measurement standpoint. (noiseLAB also provides editing, and is certified).




General Questions

1. Which hardware (digitizer) is supported?

The following National Instruments hardware is supported to be able to provide Class 1 (as defined by IEC 61672-1) Sound measurements
PCI plug in cards:
4-Channel Digitizer with built-in Microphone Powering (IEPE): PCI-4474
8-Channel Digitizer with built-in Microphone Powering (IEPE): PCI-4472
4-Channel USB based digitized with built-in Microphone Powering (IEPE): USB-9233

National Instruments hardware has a significant advantage over Sound Cards, in that it is provided with calibration certificates.

Only one card at a time is supported.

2. Can I use noiseLAB with my PC Sound Card?

Yes, but most Windows XP sound cards will most likely work, but are not supported by DELTA. Therefore they are used with noiseLAB at the customers own risk.

3. How do I calibrate my system with a sound card?

If you use a Sound Card, DELTA provides traceable calibration of cards sound cards.


If you use a National Instruments card (see question 1 above) these are provided with factory based calibration certificates, and National Instruments or DELTA can both provide traceable calibration of these. Calibration cost sound cards for up to two channels is €195, up to four channels €295. If the product fails to conform to relevant Type 1 sound level meter standards, it will be returned to you and you will not be invoiced.

4. Which operating systems is noiseLAB supported on.

noiseLAB is supported on Windows XP home and Windows XP professional. DELTA provides no support if you attempt to operate noiseLAB on any other than these platforms. noiseLAB is not yet supported on VISTA.

5. Does noiseLAB run with Service Release 2 of Windows XP?

Yes. Service Release 2 is required.

6. What are the computer requirements?

noiseLAB requires minimum of a 1.5 GHz Pentium 4 processor with 256 MByte of RAM.

Disk Requirements: approximately 1.5 GByte of disk space is required per channel per hour.
Screen Requirements: noiseLAB is designed to run on a screen size of 1024 x 768 pixes or larger. noiseLAB front panels are not rescable in size.
Note: noiseLAB 2.1 does not run on multiple core processors. For hyperthreaded processors, HT must be disabled.

7. How fast must the disk be?

Generally speaking, most modern machines with a Pentium 4 processor will be able to sustain a minimum of 4 channels real time streaming to an internal disk without problems.

Before performing a critical measurements, DELTA strongly recommends that you make a half hour recording of the number of channels required in your final measurements, to verify that there is adequate disk and CPU performance.

DELTA makes no representations of that the software will record to disk reliably on any PC. Although there is a very high probability that most PCs that meet the above stated requirements will perform reliably, there are so many variable in PCs that an individual test must be performed. Therefore, the user may return the software for a full refund within 30 days after initial download.

8. What if I use the software on a machine that does not conform to DELTA recommendations?

DELTA provides no support for these systems.

9. noiseLAB seems to require that I enter a lot of information before starting a measurement. Can’t I just start the measurement if I’m in a hurry.

No. The only way you can "save" the uncalibrated channel is to let the recorder continue to run while you get a calibrator, apply the calibration signal to each channel of that recording. You can then apply that using "Post Calibration", which will only apply to the recording and channel which contains the calibration signal. This convention is made to ensure integrity in the measurement chain in calibrations.

10. Can I retroctively calibrate a channel?

Yes. This can be done using the .wav Calibrate Utility in noiseLAB 3.0.

11. How fast is the third octave analysis in noiseLAB?

A Typical PC will compute the third octave analysis of the recorded data four to 10 times faster than real time. This means, for example if you are analyzing 60 seconds of recorded data, that the analysis be 4 to 10 times shorter, i.e. from 15 to 6 seconds.

12. What is the dynamic range of noiseLAB?

This depends entirely on the dynamic range of your digitizer. The computational dynamic range of noiseLAB is more than 300 dB for most measurement, and hence well beyond the capabilities of any digitizer known now or in the foreseeable future.

13. I need a 1024 point FFT transform. Does noiseLAB support this?

Yes, noiseLAB supports any arbitrary size transform up to the maximum specified limit (10,000 lines for the Advanced Edition or 100,000 lines for the Professional Edition)

14. But the transform size does not appear anywhere on the front panel?

That is correct, because the most relevant measurement related parameter for the user is the frequency resolution. You simply enter the desired frequency resolution, and then noiseLAB automatically selects the right FFT transform behind the scenes to give you the right result. FFT transform sizes are a leftover from the past, where the industry concentrated on the mechanics of the transform instead of the desired frequency resolution.

15. Does noiseLAB have zoom FFT?

No, because noiseLAB provides such large FFT transform sizes, which gives very high frequency resolution, that zoom is not necessary for most applications. To keep it simple for the end user, noiseLAB only provides one setting, the frequency resolution. You don’t have to worry about span, center frequency, transform size etc. as on traditional instruments.

16. What overlap does noiseLAB provide for FFT spectra?

The overlap is always 75% in noiseLAB Classic, and programmable in the Batch Processor. thus ensuring all recorded data is processed even when using the Hanning window. This parameter is fixed, so you don’t need to worry about it.

17. How fast is the FFT?

Again, typically up to 10 times faster than real time. This, however, also depends on your PC speed.

18. What are the limits on the recording length?

Up to 8 hours for one project (Professional Edition only)

19. What happens to my recording if my PC crashes in the middle of a recording?

noiseLAB actually breaks up your recording into a number of smaller blocks, of 5 minutes duration each. Thus you will maximally lose 5 minutes of data in case of a crash.

20. What is the cost of an upgrade from one Advanced to Professional Edition?

The price is 50% more than the list price difference.

21. Can I connect a DAT tape recorder to noiseLAB?

Yes. At present you can connect the analog out of the DAT tot he analog in of the digitizer.

22. Is noiseLAB supported on Windows 2000? If not, why?

noiseLAB is only supported on Windows XP. To keep the selling price of the software low for the end user, DELTA has chosen to minimize support costs limiting support to Windows XP.

23. Can I use noiseLAB with versions of Office earlier than Office 2003?

Only Office 2003 Professional contains all the XML facilities necessary to support the report generation capabilities of noiseLAB. Earlier versions of Office are not supported. Also if you buy the stand-alone version of Word, it also includes XML capabilities, but as mentioned, when it is bundled with Office 2003, it requires the Professional editing to support XML.

24. I need to make ultra-low frequency analysis of infrasound signals below 10 Hz. Does noiseLAB support this?

Yes, if you choose the proper National Instrument card in DC coupled mode. In addition, the best frequency resolution when operating at 51 kS/s sample frequency will be a 0.2 Hz.

Please note that these capabilites are not possible with PC Sound Cards.

noiseLAB 3.0 also provides A and G weighted 1/3 octave spectra for infrasound and low frequency analysis.