DIY Audio General - stuff you're making, tips, advice sought, etc


I find rew a bit heavy handed for eq generation. I find it easier just to measure with it, add a filter in the mini dsp and measure again.

The mini dsp is really only parametric eq, so only useful where the room is modal really, under 200hz.

I also don’t try and make the bass flat at listening position, a slight rise from about 150hz to 30hz approximates the house/x curve.

Also don’t just look at the response graph, look st the waterfall/spectral decay, sometimes you actually need a bit of a notch in the response to get a clean decay, which usually sounds better.


I agree. But it’s a very good starting point. I know from experience that a truly flat in room FR doesn’t sound good to me. I like a very slight lift in the presence region, 1.5 to 3dB at most.


I have it flat 130-8khz ish 6-10db boost at 25-30hz. But that was with 18” bass drivers.


Helluva boost there.

I tried the house curve idea the other day, didn’t like it as I felt it swamped things somewhat. But my room is not big so that could be why.


Not really, since it’s done so low, it’s only a few dB by the time it’s at 70-80hz. The JBL room correction curve has even more boost.

It’s all down to taste/room etc though, no real wrong or right to it.


Very much down to taste, preference.

I like deep bass but don’t want to feel swamped by it.

What’s the JBL curve?


Good comparison of the various curves. Audessey is flat, but if you read what they say on avsforum it’s so it can be adjusted to preference with the receivers sub level setting.


And unlike most of the frequency band, bass is all over on recordings, depends on the studio used and their setup etc. I found about 6db up at 25-30hz worked well with most of the stuff I listen to.


Cheers for that. I’ll take a look when I boot the PC.

I have noticed a large difference in bass output across recordings. Some have lots and others very little.

Also music dependent. String quartet won’t have much obviously


I have golden ears, so I don’t need DSP. :rofl:


Golden years :rofl:


Lab sub and a bit of dsp on the sub would be nice in your system :smile:


I have thought about the possibility of having horn bass in my system, but the faff factor would be huge. CBA prevents me from doing this :grimacing:



Subs and speakers crossing in and out at 85hz - 18db/oct.

Andy suggested PEQing a low shelf on the subs at 45hz, a 2.5db lift, Q of 0.7. The low/high shelf feature had passed me by at that point.
Working the subs a bit harder at the bottom has improved the overall experience as my room is quite big.

I’ve also dropped 2K by 2db, Q of 2.

Ive done nothing to high frequencies yet as they’re subjectively ok to my ears.

I really must measure things properly, everything has been set by listening up to now. CBA at the moment though.


Thanks mate, interesting. I’ll give them a go.

This I found useful, skip to 19 minutes in.


I tried this and wasn’t quite sure, it sounded a bit muted. But rather than change back I increased it, it’s now -1dB at 2kHz with a Q of 2. This is working well for me :+1:


Have stripped back the paint and just a rough polish with a drill mounted wire brush wheel and think it has come out quite well so far.

Will have to do some more work to get rid of the scratches and get a nice mirror shine before lacquering but not a bad effort for a complete amateur.


Model a circuit of an ECC81 with CCS load at 4.5mA, 430ohm bias resistor. Expected output level from circuit given known input: c. 1V. Actual output: way less than that on both channels.

Sounds very good but just not loud enough. Gnaah…


The ECC81’s actual AC load is the CCS in parallel with the input impedance of the next stage. Is the model handling that next stage properly ? Could you be being stung by something like the Miller effect ? In fact are you driving the whole thing with a source which isn’t going to be bothered by Miller effect in the '81 ?

Sorry if I’m trying to teach my granny to suck eggs here :face_with_hand_over_mouth:.



Next stage is an ECC81 cathode follower via cap coupling so it really should be ok. Source is a ahem cascode ECC88 and a network of L C and R with a 6.2k impedance…:thinking: