Is 50Hz the magic XO number wrt subs in a music system?

[Nyal Mellor]

I came across an AES paper recently (Martens, Braasch, Woszczyk 'Identification and Discrimination of Listener Envelopment Percepts Associated with Multiple Low-Frequency Signals in Multi-Channel Sound Reproduction' that shows the result of a study from which the conclusions were that keeping bass signals separate above 50Hz contributes to perceived envelopment and spaciousness of the sound.

The extension of this conclusion being that for ultimate quality the main LR speakers should NOT be crossed over to subs located physically separate (e.g. in the corner or behind the listener) from the LRs above about 50Hz. Also that the LR signals should NOT be combined monophonically and directed to sub(s) since the signals above 50Hz contain information that contributes to overall sound quality.

What are people's thoughts on this? How does this fit with the typical paradigm of combining musical information monophonically below about 80-100Hz? How does this work with the Geddes multi-sub approach for music?

[Steve Williams]

I completely agree

I have always felt that if a sub is crossed over too high you get some midbass information inthe sub which IMO always seems to localizethe sub in the room. Crossed over correctly the sub should not be distinguished as to its position in the room

Presently I am crossing over at 41 Hz

[cjfrbw]

I think the rule of thumb is that you don't want localizable sound above 200 Hz entering the sub in a volume that can be heard compared to the main signal.

Using this rule of thumb, a 12db/octave filter would dictate a pass frequency of between 40-60hz which would roll off in excess of 20db for 200 Hz,
You could use steeper filters, which some believe make the blending worse, or shallower filters which would require a lower cut off point.

Just playing it by ear, I find that a 80 Hz crossover with 12db/octave filter seems to work fine in my room, I can't hear the location of the subs unless they rattle some physical object. So some of it just has to do with what works in your room with your speakers through physical experimentation and measurements if you have that capability.

Also, some individuals with very small mains may choose higher crossover points just to remove the amplifier demand from the smaller speakers caused by the bass.

[RUR]

Nyal, a fair bit of discussion on bass crossover in this thread. Copy of the AES paper may be found here.

[Steve Williams]

I use a crossover of 41 Hz with a 12 Db filter. I was using a 24 Db filter but prefer the 12 Db now

[flez007]

I found it as the second best option for my specific setup, the first one was 86Hz with a soft rollout curve, but differences were VERY subtle. This could be because of my particular room and speaker characteristics. I used the equalisation programme of my Velodyne sub which really works better than going just by ear.

[DonH50]

I participated in a listening test long ago that had maybe a dozen people run trials to see when we could "localize" the subwoofer. IIRC, there were three subs set up, and which one (or 2, or 3) were on was determined randomly. We all took turns over a couple of days doing multiple (blind) trials. The results showed generally that everybody could tell where the sub was at around 100 Hz, nobody could below 50 Hz, and between was a continuum, with some not finding it below 100 Hz and some reaching to near 50 Hz. I was around 80 Hz, I think.

Another thing that came out of it was that the sub was easier to localize at very high volume, which we speculated was due to the percussive wave from some material (we used the Sheffield Drum record, pink noise, test tones, and several musical selections including rock/pop, orchestral, and organ -- and no, I don't recall all the selections!) I think we ran most of the tests at "moderate" volume, about 75 dB SPL.

It took longer than we expected to set up because we had to almost clear the room -- when something rattled anywhere people tended to focus on that and it throug off the localization results. We had to tie down and take down shelfs, tighten a few things, etc.

Fun times - Don

p.s. I am crossing over around 40 Hz from my Maggies to a pair of subs, treating them as "full-range" L/R speakers. I have preferred this in the past, before the time of LFE, and just returned to my default path. The xovr was determined from measurements and listening; mostly measurements as there's not a lot to hear down there.

[Nyal Mellor]

It also probably depends greatly on what type of music you mostly listen to. The authors of the AES paper used computer generated 'noise' rather than music...

Thanks for posting your findings Don, an interesting experiment! Was that something you did as part of a job or just as part of this crazy hobby?

[DonH50]

I worked at a high-end store in Columbia, MO that was started by some of the guys who started DiscWasher (remember them?) and had ties to some places in KC and St. Louis. At that time (late 70's/early 80's) we had quite a few high-power types stop through one or the other (typically Columbia or KC) as it was sort of the only high-end place in the "middle" midwest. For example, one of the Shure VP's made a special Greek/Roman style listening "room" sort of attached to his house outside KC, with a stacked HQD* system and all sorts of cool gear. The work got me into a lot of very interesting projects and rubbing elbows with a lot of folk I had no real business meeting, but it was a lot of fun! My job was mostly repair tech and installations, with a tiny bit of sales when things got busy. So, the short answer is a little of both. - Don

* HQD = Hartley/Quad/Decca = 24" Hartley subwoofer, a pair of Quad ESL-63's and Decca ribbon tweeters (4, I think) per side of a stereo system. At that time the Wilson of the audio world. This was long before the home theater revolution.

[FrantzM]

OT.
The HQD used the original Quad ESL, often referred to as the ESL-57 although Quad never called it that.The second Quad ESL was officially named the ESL-63 though, in the early 80's... Heard the HQD (by Mark Levinson, the Man) and a wonderful speaker for its time .. I think that DAW used ESL for the midrange in some (all?) of the WAMM iteration.

Interesting question.. later if time permits