I actually asked them if they had an opinion on the suggested EQ on the Spinorama website. If you look at the measured response of the F6B versus the EQ’d response I would say the EQ’d response is smoother and flatter. I’m not saying there would be a huge audible effect, but if striving for absolute accuracy maybe the EQ is desirable. With a passive speaker there is only so much you can do and Ascilab have already achieved really excellent results. My question is whether the application of speaker correction using EQ is desirable from the manufacturer’s point of view. I imagine that with active speakers there is a level of fine tuning similar to what we’re talking about.
Don’t you think there is value in the idea of using EQ to finish the designers intended response curve of a passive speaker?
I don't think so. The point should be to correct for audible deviations, not to get a pretty graph for the sake of a pretty graph.
There is value in eq if it makes things better - in your listening room and listening position with your listening preferences. If not then not.
You misunderstand the data. This is a sample of one and the deviations are very small. They are small enough that the nornal error of the Klippel fixture and microphone are within thise bounds. It is also not known if the deviations will be consistent across speakers. Part to part variation is fairly wide and we already know that Ascilab puts a lot of effort into sorting and selection.
What is the point of the suggested EQ settings published on Spinorama?
They are a computed process to determine what is possible mathematically. I am not sure if people have tried them to see if the improvement is there or not. Long time ago I wanted to do that but have not found the time. We know the preference score is not perfect so it is possible the EQ makes things worse although less likely than making it better.
I did try it with my Revel M16, but wasn’t sure it was better. As Wiim can now do individual channel EQ I might try again and compare in mono with one channel with EQ applied.
As Amir said, they are mathematically generated to show a set of filters that improves the preference score to the extent possible. There's no guarantee that they are actually an audible improvement. My thoughts are that filters that are low Q and around 1dB in magnitude or more could be audible improvements (would have to A/B to be sure of course). High Q filters are not likely to be an improvement, particularly if the adjustment is low in magnitude. There's also always a filter to extend the low frequency (as low frequency extension has a large effect on the preference score). That would likely be an audible improvement if the speaker actually has headroom down there. If not, it will probably just result in a lot of added distortion with low frequency content which the preference score doesn't take into account.
www.audiosciencereview.com
Kudos to the driver's manufacturer and those selected by Ascilab ... doubled professional expertise.Case in point, @Nuyes measured (gated, nonanechoic) 10 samples of the A6B, plotting the difference vs. Ascilab's reference sample.
View attachment 461107
Tight tolerances.
AsciLab A6B Review
I sincerely apologize for the significant delay in posting this review. Thank you to everyone who patiently waited. This review marks a milestone, as it is the first time I’ve measured a total of 11 samples simultaneously, thanks to the collaboration of the manufacturer and...
@Lars Risbo Do you have any comments?
I think there is value. There is a sensible place to stop when designing a passive crossover, and the manufacturer absolutely should do a great job (as they are here), but that doesn't mean it's optimal if price and complexity weren't a factor. DSP can add a final polish, or subtle alternative tunings for different use cases. If a manufacturer has quality data, and the technical knowledge (again, they have that here) it doesn't look like much work to produce them. My passive speakers have convolution files from the manufacturer that improve group delay and something to in the crossover region, the changes are very subtle though.
Your thinking is in line with mine.
I just don't follow the logic here. When you already have very linear speakers, as Ascilab does, what is the point of that "final polish"? You even acknowledge that your speakers' convolution filters only make very subtle changes (and are your speakers as linear as Ascilab to start with?). As for "alternative tunings"... for what? If someone wants to EQ to taste, there's no need for the manufacturer there as that really has nothing to do with quality data or technical knowledge. That's about the end user's personal preferences.
I've had the same experience with that feature in kii and D&D speakers, I'm not sure it's never audible, but I'm convinced it's never important.
Meridian had these positional based DSP presets too. Free, Wall and Corner if memory served. They were audibly very different.
