So either one of three things happened:
1) Ascend production line measurements just aren't accurate, which is concerning and would love for
@AscendDF to comment on this. This is a matter of quality control.
2) Something happened to your speakers from the time they left the factory to the moment that Erin measured them.
3) Something is off with Erin's measurement.
This is very concerning to learn that the speakers measured by the NFS does not aligned with the production line measurements. Dave, really need you to comment on this.
Firstly, It is important to understand that our LX is not designed to be perfectly flat on-axis. If you look at our published measurements, there is a slight ~2dB bump starting at ~1.3kHz, and then a shallow dip at ~3kHz. The slight increase in woofer amplitude is to compensate for the narrowing directivity of the woofer, such that we better optimize overall directivity. The shallow dip has the same purpose, but just the opposite, to optimize the wider directivity as the tweeter begins to take over by lessening amplitude. You might need to look more carefully at the measurement to see this because the image itself is smaller, but it's plainly obvious.
We are a rare breed, we actually fully test each and every speaker we ship. Most speaker manufacturers only batch test, meaning they will do a production run and test only a few speakers. I personally test every Sierra and ELX speaker we ship. It is not remotely possible to get Klippel NFS like resolution in a production line environment, nor is it needed. As such, our QC testing (which is intense) - is still not going to show what an NFS is capable of resolving. We have a fully dedicated production line testing environment. There are various pictures of this online but the critical factor to understand here is that there is a HUGE difference and tremendous complications when testing speakers on a production line. We test frequency response, phase, polarity, impedance, rub and buzz, and we sinewave sweep and listen to every speaker we ship. You really won't find that level of QC anywhere else.
If someone wants NFS levels of precision, this is exactly why we offer this option. If we were to even attempt to test each speaker we ship on the NFS, we would, at most, maybe ship a pair of speakers per day. It is not practical, possible or necessary - and this is coming from someone with now over 40 years of speaker manufacturing experience at many different levels. Taking this further, let's say we do NFS test for production, and then come across something like Erin picked up, knowing full well that the driver in question passed SEAS's rigorous testing. Do we reject or pass? Too many rejects means higher costs, which translates to higher pricing. There isn't close to enough profit in any of our products to simply absorb this, especially knowing that it is extremely unlikely anyone could actually hear such a narrow +2 dB bump (this customer certainly couldn't, nor was I able to after extensive listening).
** Erin requested that I retract the comparison I used. Apparently, the FR5 Erin used to take his measurements had a problem and then was later corrected before Stereophile took their measurements. I was unaware of this, as such - and out of respect for Erin, I have retracted this comparison.
One of the reason we love SEAS is they fully test each driver we purchase from them, prior to shipping. Like all manufacturer's, they too stick to an acceptable +/- tolerance range. I don't feel comfortable publishing what this is (not sure how my friends there would feel) - but let's just say it is tighter than any other driver manufacturer I have worked with or sampled from. There is a reason SEAS has been as successful as they are for an amazing 60 years now.
As a manufacturer, we can not maintain higher tolerances than our driver manufacturers, nor can our driver manufacturers stick to even tighter tolerances without considerable increase in pricing due to higher rejection rates. This is how this industry works.
As such, the Klippel NFS creates an interesting dilemma for manufacturers who post NFS measurements, and likely the reason why so few bother to do so. With measurements becoming the norm, and with the general public becoming more educated, what exactly is an acceptable tolerance from a published NFS measurement? +/- 1dB? +/- 1.5dB? +/- 2dB? It used to be that high quality manufacturers tried to stick to a 6dB window (+/- 3dB) and that was with measurement gear nowhere near the capability of an NFS, while also being taken in poor environments. It is unrealistic for a customer to expect their speakers to measure the exact same - far too many tolerances involved. When we choose a speaker to be our "reference" - it isn't what measures best. We measure dozens of drivers and we pick the drivers that best represent the median.
In the case of this customer's speakers - Erin reached out to me because he picked up a high Q (very narrow) +2dB peak in one of the speakers that isn't evident in our published measurement. This is slightly noticeable on our production line measurement of said speaker, but as stated, a production line test can't come close to picking up what an NFS can. We also don't compare our production line measurements with our NFS measurements, we compare against our production line measurement of our reference unit and as long as they are within +/- 1.5dB of our reference, (which is considered a very tight tolerance) and the speaker under test passed our other tests, it is approved. Understand that we have a fixed microphone position, imagine having to adjust an extremely expensive microphone for each different speaker? Not possible or practical.
Our procedures work, and work exceptionally well - our LX has been and continues to be a huge success for us. We have had only 2 returns in all of 2025 so far, translating to a customer satisfaction rate that dwarfs any product I have ever been associated with through Ascend or any of the many other companies I have worked with or been employed by. I believe this rate is only matched by our ELX but it isn't a fair comparison as there are many many more LX's out in the field.
Erin reached out to me - he was curious about what he saw, as was I after he shared with me - and I wanted that one woofer back. I didn't feel that the measurements he took are representative of the average pair of LX and I am curious to compare with our NFS measurements so I offered to send replacement woofers. It is important to take raw NFS measurements of those woofers in order for me to determine if they fall within SEAS's acceptable tolerance range. SEAS records batch dates and measurements per batch, sometimes we order in huge quantities sometimes not as large. Now that I have the woofers back, these were from a smaller air-shipped batched that we needed to hold us over in lieu of a much larger ocean freight batch of LX woofers.
It is one thing if
we send a pair of LX to Erin for review. because, of course we would have tested it on the NFS beforehand. Any manufacturer would do the exact same - anyone thinking a manufacturer would just grab any pair from stock, knowing full well these are going to be NFS tested, is not being realistic or honest. At a previous employer, I was actually responsible for doing this (and that was with gear that could not compare to an NFS) But this pair sent to Erin was from a customer who absolutely loves the speakers, who also wasn't able to measure the bump using REW. I deeply care about every pair of speakers we ship so if something seemed a bit off to Erin when comparing his measurements to ours, you can be sure I am going to fully investigate regardless of whether the customer loves the speakers or hates them.
For full transparency and for the general knowledgebase, below are the raw NFS measurements of the two woofers being discussed as well as our reference woofer. These were taken in our dedicated sealed testing cabinet, no crossover, no port tube etc. This is so SEAS can compare with their data. You can see just how close they measure through their entire bandwidth, except for some minor differences in the 1.5kHz to 2kHz range. You can clearly see how one woofer is at one end of the tolerance, and the other is on the other side, with the woofer I pulled from our reference unit being right in the middle. This is what I suspected when I discussed with Erin, it's simply tolerances, certainly not any type of flaw or a bad woofer. Slight differences like this are always going to exist in drivers, especially woofers and unless measured in an anechoic chamber with the proper gear or an NFS, these differences you can see would be much less - if even detectable. With the NFS, we are looking at details through a high powered microscope, not much different than looking at what you thought was a perfect auto paint job until you see it under a microscope...
The replacement woofers I sent Erin are from 2 different production batches, so it will be interesting to see how they measure. Since I don't have this customer's speakers, I have not measured these woofers, only sinewave swept them.
Hope this all makes sense!
