That is correct but bear in mind that the delay needed by Tannoy DCs is only around 6 microseconds and that according to Tannoys own engineers the SyncSource circuit was a marketing-driven addition and DCs sound better without it.
My own Tannoys never had it fitted however I do know of a few people who have tried the SyncSource circuit but most removed it again. It tends to 'smear' the response ie you lose some fine details in reproduction.
I already proved you right Charles
Still enjoying my Tanny's without the sync source !
But he can achieve 0.938 mS delay with this circuit with the right components , so it's worth a try .
Cheers ,
Rens
I wouldn't advise you to go this way, but in the 1950's they used a rotating disc of magnetic material with a recording and a playback head to create delays for phased loudspeaker arrays. I've seen a classical article on how they applied that in a big English church, to make the sermon more understandable without needing to damp the reverberation.
Measuring and listening.
Basically, I first physically aligned the voice coils of my horns and did not use any delay. But since the bass horn is long and large, its mouth would start to shadow (or partially cover) the horns above it - mid horn and high horn, create diffraction that was clearly audible.
I moved the the mid/high horns forward so the were aligned by horn openings = about 12 inches = about 1mS of delay.
Further listening test and adjustments of delay pretty much confirmed that the best sound was when speaker voice coils were virtually in line. That was equivalent to 12 inches delay.
Herman
Basically, I first physically aligned the voice coils of my horns and did not use any delay. But since the bass horn is long and large, its mouth would start to shadow (or partially cover) the horns above it - mid horn and high horn, create diffraction that was clearly audible.
I moved the the mid/high horns forward so the were aligned by horn openings = about 12 inches = about 1mS of delay.
Further listening test and adjustments of delay pretty much confirmed that the best sound was when speaker voice coils were virtually in line. That was equivalent to 12 inches delay.
Herman
Certainly a digital delay is the best option here , whether it be with a FIFO or DSP as the delay element, the result is the same. DSP will be easier to implement though since it is quite commonly available, i.e. MiniDSP modules.
There is no reason, with modern ADC and DAC stages, that your audio will be altered in any audible way at all by this. 24bit is common and most DSP + ADC/DACs can run at 96k if not 192k, just to be sure! Of course, while you're using a DSP you might as well put your crossover in there too
There is no reason, with modern ADC and DAC stages, that your audio will be altered in any audible way at all by this. 24bit is common and most DSP + ADC/DACs can run at 96k if not 192k, just to be sure! Of course, while you're using a DSP you might as well put your crossover in there too
The acoustic center for a cone driver is not the voice coil. As sound propagates faster in the cone than in the air, the acoustic center is the ring on the cone where the area closer to the voice coil is equal to the area away from it.
In a horn the acoustic center actually moves with many factors including volume, so essentially this also moves the center out from the voice coil.
Now just to polish things off high frequencies travel faster than low frequencies.
But as to what you want to do to delay the signal to a given driver, you use what is called an all pass filter. Fortunately this is covered by a patent. Patent US6513622 - Full-range loudspeaker system for cinema screen - Google Patents Bernie really did most of the work but I assume Bill his boss contributed as both their names are on the patent. Also see http://en.wikipedia.org/wiki/Bridged_T_delay_equaliser
So that should answer your question.
But FYI before modern discrete time discrete level audio delays came the analog chips that were continuous level but discrete time. Before that were both tape loop and magnetic disc recorders, some with multiple heads for additional audio taps. And the first method was a loudspeaker driving a coiled tube into a microphone. There really is an Altec application note showing how to use a garden hose for that.
For what you want to do if you used DSP you would have some issues with the inherent A/D-DSP-D/A processing delays.
Last edited:
That's the issue with DSPs. Sound quality.
Put a DSP between the pre-amp and amp. No filters, just pass through. And you're saying it sounds the same or as good as no DSP in the SAME signal path?
If you know of a DSP that does -- please tell me.
I've tried various brands: XTA, Dbx, Lake. $3000 and over units. All kill the sound.
Herman
Put a DSP between the pre-amp and amp. No filters, just pass through. And you're saying it sounds the same or as good as no DSP in the SAME signal path?
If you know of a DSP that does -- please tell me.
I've tried various brands: XTA, Dbx, Lake. $3000 and over units. All kill the sound.
Herman
Simon7000 makes a good point, that the DSP may not process in as little as 1mS, so the typical solution is to put all channels via DSP where they are all coherent to begin with and you can time offset them precisely.
And yes, a DSP won't affect quality on direct pass though, nor should it when modifying the signal since it uses extremely accurate 32bit floating point arithmetic. If you are hearing any effects it will be from the ADC/DAC stage, but it still sounds unfeasible considering the accuracy of converters in those units. Most recordings are made using large amounts of DSP to begin with, be it hardware implementation or software running on a computer, it's nothing different or special, just 32bit floating DSP.
Ultimately it sounds like you've proven to yourself there is a lot to be gained from time aligning your horn system (typically there will be in such a setup) and DSP is really the only suitable route, certainly far better audibly than BBD, magnetic recording, hosepipes (!?) or all-pass filters.
And yes, a DSP won't affect quality on direct pass though, nor should it when modifying the signal since it uses extremely accurate 32bit floating point arithmetic. If you are hearing any effects it will be from the ADC/DAC stage, but it still sounds unfeasible considering the accuracy of converters in those units. Most recordings are made using large amounts of DSP to begin with, be it hardware implementation or software running on a computer, it's nothing different or special, just 32bit floating DSP.
Ultimately it sounds like you've proven to yourself there is a lot to be gained from time aligning your horn system (typically there will be in such a setup) and DSP is really the only suitable route, certainly far better audibly than BBD, magnetic recording, hosepipes (!?) or all-pass filters.
Last edited:
Looks like you'll need to invest in a decent-sounding ADC and DAC chain then. Probably not easy to find off the shelf, but hey this is DIY
<edit> Incidentally its not DSP which is the enemy of good (meaning digital with all the good qualities of analog and none of the bad) sound, rather lousy DACs and ADCs.
You don't need a digital delay if you need 1msec and your crossover point is 250Hz as long as you are OK with the delay falling back to zero at some frequency above that since this behavior is what characterizes an analog (all-pass) delay circuit of any order. So, if you want to "time align" the system then analog will not work, however, if your goal is to align the phase at the crossover point and nearby, then analog delay would be fine.
If you are going to bother to use DSP for digital delay then you might as well just implement the entire crossover using DSP.
.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.