1. Things you need to know about the new ‘Conversations’ PM system:

    a) DO NOT REPLY TO THE NOTIFICATION EMAIL! I get them, not the intended recipient. I get a lot of them and I do not want them! It is just a notification, log into the site and reply from there.

    b) To delete old conversations use the ‘Leave conversation’ option. This is just delete by another name.
    Dismiss Notice

Getting to know LS3/5As

Discussion in 'classic' started by Tony L, Dec 26, 2020.

  1. omers

    omers pfm Member

    The caterpillar on the top one seems healthier.
    ampedup likes this.
  2. ToTo Man

    ToTo Man the band not the dog

    Tony's step responses with a zoomed x-axis.

    I'm still not sure how to interpret these as they don't quite resemble the step responses that JA publishes or the ones I took of my Tannoys. Please chime in if you can interpret them?!

    1ms window:



    6ms window:


  3. omers

    omers pfm Member

    Please notice that the X axis is much more sensitive than that usually published in tests. All the action here happens within 0.2ms e.g. for the JR149, you should zoom out.

    Edit: Just saw you specifically wrote the X is zoomed-in. Usually JA does not go into these sub-ms resolution. This graph is the ripple that appears in many of his tests right after the main "rise".
    Last edited: Jan 18, 2021
  4. omers

    omers pfm Member

    One more thing re impulse response, not sure whether this was mentioned -
    Shouldn't the room be anechoic? or can the sampler filter out the returning waves?
  5. Tony L

    Tony L Administrator

    The thing I don’t understand is how REW can even generate a step/impulse response at all from a single frequency sweep measurement. I have a feeling Stereophile use a specific test tone to generate it, but again I really don’t understand this stuff.

    As stated previously I don’t personally think the odd db up/down in amplitude matters hugely, if at all. A speaker or speaker/amp/room combo is good or bad long, long before a 2db lift at 800Hz or whatever as human ears just aren’t very good at ascertaining that degree of subtlety in level, even relative level. My theory is what we really don’t like are unwanted artefacts such cabinet resonances, bad port loading, bad multi-driver time-alignment/phase/comb effects, driver distortion etc etc. I’d love to learn to use REW to better understand this. Especially the phase side of things, but as stated that just looks like a total mess to me, more ‘all over the place’ than I would have believed possible.
  6. ToTo Man

    ToTo Man the band not the dog

    REW users often gate their measurements to remove room effects but I don't understand enough about this to know how to implement it correctly.
    The further the mic is from the speaker, the more influence reflections will have in the measurements, but the closer the mic is to the speaker, the more vertical distance there is between the speaker's drive units relative to the distance to the mic, which is a problem when measuring multi-driver speakers.

    PS - I'd be very surprised if JA has an anechoic chamber!
    omers likes this.
  7. Tony L

    Tony L Administrator

    Here’s some explanation from JA regarding step response (Stereophile). I still can’t for the life of me understand how REW generates it from a frequency sweep. It just makes no sense to me. I would have thought it needed a single full-range (pink noise?) pulse, i.e. a single transient moment in time. I guess REW is doing some fancy maths by knowing how long the sweep takes to get from 20Hz to 20kHz and subtracting that time somehow to emulate an impulse, but I don’t see how that could possibly be an accurate indicator as to how the speaker would behave if confronted with a full-range pulse of a ms or whatever. I just don’t get it at all, though I certainly understand how such a measurement could be very useful.

    I’d also love to see how each speaker reproduced square waves etc, I know most multi-way speakers screw them up quite literally beyond recognition (the Gale 401 apparently being an exception), though planar speakers can do them. Given they are such an essential building block of electronica etc I’d have hoped a loudspeaker could reproduce them with some degree of accuracy.

    PA The common denominator between these two speakers, beyond the identical drivers, is a huge, huge amount of thought went into the cabinet design. Both game-changers in their own way. I think it is pretty clear from the plots posted so far that they are exceptionally well behaved, but I’m curious as to how that radically different construction shows itself.
  8. Yank

    Yank Bulbous Also Tapered

    I think it's more a matter of how many layers you can peel off an onion.
  9. DaveMc64

    DaveMc64 pfm Member

    It's been mentioned before but it has been suggested that the B110s for the LS3/5a were selected to minimise that peak.

    If you look at the BBC documentation for the LS3/5a it mentions that in the crossover "the group C5, L2, R2, compensates for a hump in this characteristic [of the B110]"
    If you look at the transfer characteristics of the woofer section of the LS3/5a crossover you do indeed see a dip at ~1Khz

  10. Rug Doc

    Rug Doc pfm Member

    Wow, the distortion plot showing the two is amazing, I can see why the LS3/5a has such a following - it’s super smooth and as close to distortion free in the 3khz area, I’ve never heard a pair, but really want to now.

    interesting also how close most other plots are between the LS and the JR, yet the distortion in the JR is much more significant... are the crossovers the same with the same components in both speakers? I would assume not seeing that plot, and would also presume that this is the reason the LS sounds superb on piano and simple acoustic styles. These have been designed to be monitors first and foremost and yes I can see why they are so good. Nice. I’d just want a pair with some balls in the bass dept.
  11. ToTo Man

    ToTo Man the band not the dog

    Hope Tony doesn't mind me posting these, but here's a more detailed look at the distortion plots of the two speakers.



    I'm not sure if it's a result of the different amplifier, listening environment, version of REW used or, more likely IMO, the different SPL at which the measurements are taken (I run mine at a lower level around 70 to 75dB as I'm usually running dozens of measurements at a time and am scared the repeated stress will eventually damage my tweeters!). However, the THD level is higher than I'm used to seeing, particularly in the 1kHz-10kHz region. I'll need to dig out my own JR149 measurements and see how they compare.
  12. Tony L

    Tony L Administrator

    I am using a 10 Watt Leak valve amp, so I guess it may have a little more measured distortion despite sounding absolutely superb. At some point I’ll measure again with a 303, but it would mean farting around digging cables out as both 303s are pretty well plumbed into their systems.

    I have to admit I’d be very surprised if the Leak was to blame as we are talking less than a Watt here, but it may make sense to double check. Certainly no sonic warning signs, the system with either speaker is as good as I’ve ever heard any mini-monitor sound. Just staggeringly good to be honest.
    ToTo Man likes this.
  13. ToTo Man

    ToTo Man the band not the dog

    I checked my own JR149 sine sweep measurements and they were all done at 70dB and 1m distance so the distortion plots cannot be compared to yours, I'd need to remeasure from further away and at a higher SPL for any meaningful comparison. Here's the nearfield distortion graph for mine anyway in case your interested.

    Measuring in excess of twenty HF2000 units has given me a fairly good idea of what distortion sounds like in the 2kHz-6kHz area. Even just a few dB increase above baseline is pretty audible for odd order harmonics, so if you can't hear any funky stuff going on then I wouldn't worry about it.
  14. Tony L

    Tony L Administrator

    I’m sure that would explain the difference. A 5” bass mid has distinct volume limitations, even modern ones with huge magnets and high-tech composite construction. Given the logarithmic scale 80db is over 3 times louder than 70db, and distortion from the bass-mid will really show up. I’m pretty confident my measurements would lie within expected margins and can largely be put down to the difference in driver displacement.

    PS You absolutely won’t hurt you drivers measuring at a more normal listening level. It is a quick sweep and no where near long enough duration to heat up a voice coil. Contrast and compare with some electronica, ambient dub etc where you have steady-state square-wave drones lasting half an hour or so! Even classical organ stuff like Bach is very steady-state/low dynamic range and I guess will heat up a voice coil pretty quick.
  15. robbyd

    robbyd pfm Member

    God this thread is going on!

    l'd love to own or try a pair but at the prices they fetch that's not going to put them on my list unfortunately :D
  16. Stunsworth

    Stunsworth pfm Member

  17. Martyn Miles

    Martyn Miles pfm Member

    Those AB1s weren’t as successful as Rogers hoped they’d be.
    I understand the Stirling Broadcast AB2s have been more successful when paired with their own
  18. Tony L

    Tony L Administrator

    I’ve obviously not heard the new active Rogers yet (very curious about their new LS3/5A too as it looks like it might actually be one!), but I haven’t liked the AB1s when I’ve heard them. I’m really not a fan of subs though, I can always hear them, and have always left with the impression the AB1 isn’t even a good one.
  19. Yank

    Yank Bulbous Also Tapered

    More of a one-note whistle...

Share This Page


  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice