Dependant on how much the room is doing changes will be more apparent upfront. Was a very big difference in my case almost as big as changing the speakers, which I eventually did in the end. But in some cases its very subtle. Thierry cant and won’t fix any nulls caused by SBR but he will minimise the affects by addressing areas he can manipulate.
I’m hoping this process obviates the need for me to consider new speakers which is where my thoughts were wandering.
.sjb
It put me off for some time using DSP but as I could not use it on the analogue side my system was unbalanced too much as digital sounded so much better down to DSP. Someone mentioned that likely my speakers where just too big for the room and where overloading it and I had little flexibility to position them to their best. So I decided to try out smaller ones and if all be said better ones. And this time I had the opportunity to try them at home. It was a revelation. But I feel they still would benefit from a little DSP as the room still is problem but it’s just not as noticeble with the ProAcs as they are easier to place.
I’ve spent the last 3 days listening to the crosstalk reduction filter.
When I asked Thierry about it he replied
"The difference b/w standard and crosstalk reduction is only about soundstage : with tracks having good binaural content the soundstage will expand in all dimensions (more 3D like).
The effect is depending on each track and on your sensitivity to spatial information. The ones that are very sensitive are always using crosstalk reduction."
I know up above I referenced my lack of jaw dropping moments and veils being removed etc. but this implementation does give some spectacular results.
While listening to
I couldn’t help smiling from ear to ear. Now this really is “speakers disappearing territory”.
The width and height (not so much depth as this is Naim and SL2s) of the soundstage on some recordings is breathtaking. It has an astonishing effect on jazz recordings. I’d imagine it will be similar on classical, I tried one string quartet and I am now hearing what I hoped for when I bought the CD.
On tracks without much binaural content they sound close to the standard filter. I understand from Thierry that the crosstalk can be tuned up or down. So far the hit vs miss ratio is about 90%. (here’s what I’ve listened to https://www.last.fm/user/sloopjohn-b/library). The odd track can sound weird - I’ll have to give Sweet Caroline a few more listens before I figure out if I like the drums coming from what sounds like 1m to the left of the left speaker.
I’d probably change my advice on test track @ChrisH to one that would show off the crosstalk to its fullest.
Here are some tracks that really benefited to give you an idea.
I’m always conscious that there can be “effects” that enchant and are entrancing only to be identified as sonic charlatans in the cold light of day. That’s one reason why I have given a full 3 days listening to this filter before coming to any assessment. At this stage I think I could go as far as saying I doubt I’ll listen to an ECM album without this filter in the mix.
.sjb
Interesting stuff, I was wondering why you had gone to radio silence
I’m not expecting too much necessarily, but I am looking forward to measuring/seeing the impacts of my room on my listening position.
It’s a relatively modest cost though, so let’s see what happens, definitely curious now.
To your point about the test track, I’m not big on classical, but do have a reasonable amount of jazz, though I’ll probably end up selecting something in the pop/rock direction (thanks also for your Last.fm link, that’s a proper mixed bag, and you really have done a massive amount of listening this last few days!).
My mic turns up next week so I’ll crack on and start the ball rolling as soon as I get a chance.
Ps. I assume you mean PCM instead of ECM?
No ECM is a well know Jazz label.
Interesting, I wish I had gone for X talk now sounds good. When I get round to measuring again and and then off maybe I will ask about them this tme.
I think what I was getting at CG was whether Sloop maybe had the intention to use the crosstalk filter on all PCM, it was a lazy late night post.
ECM it was. I think the crosstalk filter will shine with these type of recordings, with a much wider soundstage and the full nuances of the instruments and performances shining through.
My point really was that I’m not sure whether the crosstalk filter (and it can be modified) would be my go to for all music, but certainly it most likely will be for jazz.
.sjb
Mic received, measurements taken today and dispatched to HAF with a test piece of music.
Taking the measurements was easier than I thought, though as you say @Sloop_John_B it is a job that would be made much easier with a mic stand
Interestingly Thierry came back quite quickly to say that my measurements were showing quite a strong tonal identity (forward mid and high trebles) asking whether I wanted the filters generated for a more neutral tonal balance or as is to keep the current character.
That was quite surprising, I imagine my NDS/SN2/PMC25.23s would be quite neutral so it must be the room interaction I guess.
Looking forward to seeing what comes back.
So which did you say to go with?
I opted for the more neutral tonal balance.
It is only for the test track at this stage.
Curious!
Yes I got that with my old speakers they had a v shaped Signature. I also went for neutral. You will notice quite a big difference to what your used to. So listen to it for several days to get accustomed to it before trying without. But you need to try and volume match when doing it as the filters will reduce overall volume.
Looking forward to trying a tweaked set of filters tonight from Thierry. I have asked for “a sweeter tonal balance. There is a slight edginess to some music and vocals, that I’d like smoothed slightly.
Having said that I don’t want a mellow mush with all attack lost”.
From a thread on the roon forum I came across the tip to save each filter as a pre-set and to use headroom to equalize the volumes (the standard filter is louder than the crosstalk filter).
This makes a quick A vs B very easy to implement.
.sjb
A lot of music under my belt in the past few hours and I think I’m at or nearing the stage where I can say this is one of the most cost effective “upgrades” I’ve ever taken.
Listening this evening, every now again I went back to no filters and almost universally the difference is apparent and in favour of Thierry’s filters.
The ability to tweak the filters is great. I’ll probably ask Thierry to do a lesser crosstalk option as the odd time I nearly jump as the sound comes so far from outside the speakers. But the crosstalk option is probably my favourite but there are some times it’s a little bit OTT.
The option of using convolution in Roon is a fantastic addition which I’m just beginning to take advantage. This mixed with Thierry’s Home Audio Fidelity filters, which makes the whole process so simple really opens up so many options for us.
This really has the potential to be a truly disruptive feature.
Honestly if you’re thinking of cables or other “upgrades” costing over £250, try this first.
.sjb
Very happy it’s working out for you, Thierry is so pleasant to work with as well. I swear I need to get around to updating my filters again soon. Roon really does make it easy to switch between filters or no filters.
I have been using HAF filters since August.
My initial motivation was to do some low frequencies correction for room modes. Thierry has a dedicated s/w VTS plugin for bass management which cannot run in Roon DSP and prefers to have a “mild” correction for reflections in his basic (non x/talk) filters.
I find the x/talk filters quite impressive in terms of clarity and soundstage but I have some reservations:
-
Standard room correction tries to eliminate the effect of the room and make the end result as close as possible to the original recording. Cross talk though seems to attempt to improve on the original recording and make it -here I am oversimplifying- more “binaural”. Some might object to the “fidelity” of this approach
-
the x/talk effect reminds me a bit of a phase inversion effect (music more spread horizontally towards the edges); with some tracks which seem to employ phase tricks by design (like Elliott Smith on Either/Or) this may lead to strange results
Overall, I don’t use HAF for small ensembles (chamber music, jazz) but do so quite a lot on rock “muddied” tracks
I received my HAF filters last Saturday (HAF + Crosstalk) within a few hours of my order, really speedy service and excellent communication.
So have started my listening tests to see if I may need the filters tweaking.
I really thought it would be more difficult than it actually has been so far, adding them into the DSP engine and loading them in as Presets, absolutely seamless and very intuitive.
Im a couple of weeks behind @Sloop_John_B in this process so first impressions are that Crosstalk has quite an impact, think mine needs tuning a little as its more right sided in its effect.
But time spent so far mostly on the standard filter, think I need a few more days on this before switching back to without filters to see how I feel.
The one thing that I was slightly apprehensive about was whether my Innuos Zenith used as Roon Core would be able to run the Convolution filters, especially as it is widely written that if a lot of DSP will be done, probably better not choosing an Innuos as it wont have enough power in the CPU.
Pleased to say it manages perfectly well - see this screen shot showing processing speed on a standard CD.
The speed drops the higher res files are played, but even at 192/24, no problem, still comfortably over the 1.0 recommended in the Roon KB and no stuttering or playback issues.
Anyway, Ill carry on listening, and now we have 1.8 to look forward to next week also!
Ill post back once Ive come to a conclusion whether any tweakery is required.
Yes I really need to motivate myself. Something I dont have a lot of this lock down. The Tabs are waiting to be fully unlocked.
Has any one tried this on entry level kit - something like an Unitiqute?
My office system gets the most hours of use, but I’m keen to try it out with my NDS.