The Lyceum: Audio Engineer Use Cases
In this article, I’ll be describing typical use cases for audio system engineers building a consumer electronics product.
As an audio engineer for just about 10 years, I’ve consistently run into these problems, wasting time and growing frustrated that there wasn’t already a SW tool that could automate and/or solve them.
Scenario 1: Comparing test measurement data to older products
When a company is developing a product to go-to-market, its required to baseline the products performance against its competition. An audio engineer does this by evaluating a metric such as frequency response, of a speaker they’re developing to current speakers being sold.
So let’s say the product I’m developing is called the “SpeakerX”, and my requirements is that it performs louder and more clearly than “SpeakerY” and “SpeakerZ”. In order to establish a baseline target, I develop a measurement experiment to acquire data of the frequency responses of all three speakers.
After data is acquired, I can upload each excel document to the Lyceum via the data ingestor.
The data ingestor attaches descriptors such as the environment I’m testing in, or the volume of the input signal I’m using, grant access to only my teammates in my company/team, as well as cleans the data so that each curve can be utilized for analysis in the future.
After ingesting all the data, I’m able to navigate to the analysis page and quickly search and find the data for comparison. The expand option will show all the available categories in the upload that I am able to pin to the right-hand “Pinboard” in order to start my analysis session with.
I would then choose the “Frequency Response” option of all three speaker uploads, and choose the button “Post-Process”.
I’ve just created a session. In this session I can select all three different speakers, and plot them on a graph. In the plot graph section, I will be able to see what the descriptions of each node are, as well as add any notes and/or observations to the graph for later referencing.
To share the results with my teammates, I can then choose “Share Session”, whereby a window will pop up where I can type in each teammates name. This session will then appear in their respective Sessions list, where they can choose to start a new session with more data to compare.
Scenario 2: Post-Processing
I’ve now got my baseline target, and would like to start setting up a 2nd test setup location at a factory, where rapid prototyping and design iteration can take place. This test setup will be one-half the distance acoustically, and therefor in order to correlate results of my baseline target experiment from my original setup to the new setup, I will need to add +6dB to the entire curve.
Once the data of the baseline and the new setup measurement is pulled into a session, I can then navigate to “Statistics”, whereby I will have the option of selecting a variety of mathematical calculations. These can be addition, subtraction, multiplication, division, as well as more complex arithmetic such as standard deviation and normalization. In this example I will choose addition, and choose “add constant”.
The next window is where I will input “6” in order to add 6dB to every point on the curve. After processing, I can then compare my measurement in the first location to the measurement in the second location.
Scenario 3: Generating Limits
Prior to a manufacturing build, one of the necessary steps is creating boundary conditions for the factory to use. The typical way this is done is by calculating a 6-sigma distribution of a data set, offsetting the upper and lower level of that distribution, and post-processing each curve against the upper and lower limits.
With the Lyceum, these limits can be generated automatically, saved and revisioned for future use. Limits may change from one point in the development process to another, and it might be useful to see how they shift over time. The Lyceum can also separate out all passing units, or all failing units in order to deep-dive into potential problems.
Next Steps:
In the future I will release an additional article as a follow-up to this, demonstrating the future capabilities of The Lyceum and its practical applications using real-world scenarios. These features will include the following:
Gage R&R
Yield Reporting
Massive data-processing for large-scale datasets
Before/After delta post-processing
Auto-correlation between different measurement environments and setups
“TestTube” attachment with procedures, lists of equipment/tools, and setup diagrams on how to reproduce measurements.
Please contact me at josh@thelyceum.io with any questions and/or requests for your project needs.
