REL Sub Woofer DSP Versus Analog Filters

[Stix]

Love to see what the ASR crowd thinks of this REL blog:

DSP Versus Analog Filters - REL Acoustics

REL Acoustics T/7x Amplifier Panel We talk about the speed of bass reproduction a lot at REL. I recently had an epiphany of sorts after speaking with a

rel.net

I had compared a REL S5 SHO to a SVS SB4000 a while back. I played around with the SVS forever with all their settings and could not get it to sound good for 2 channel. The REL sounded better and I kept it. Now looking back it could have been my room and my set up skills that were the problem.

 

[escksu]

Love to see what the ASR crowd thinks of this REL blog:

DSP Versus Analog Filters - REL Acoustics

REL Acoustics T/7x Amplifier Panel We talk about the speed of bass reproduction a lot at REL. I recently had an epiphany of sorts after speaking with a

rel.net

I had compared a REL S5 SHO to a SVS SB4000 a while back. I played around with the SVS forever with all their settings and could not get it to sound good for 2 channel. The REL sounded better and I kept it. Now looking back it could have been my room and my set up skills that were the problem.

I yet to read the article. But from my own understanding, brands like REL were better in music but not so good in movies. The likes of SVS, Velodyne etc are the opposite.

 

[Tangband]

I yet to read the article. But from my own understanding, brands like REL were better in music but not so good in movies. The likes of SVS, Velodyne etc are the opposite.

There is NO magic in why REL often sound better with 2 channel music .

1. They are bass reiforcement speakers.
They are ment to be used without HP filtering for the main loudspeakers .
Such a speaker drops about 12 dB oct below its resonance frequency if its a closed box . The REL subwoofers has 12 dB / oct LP filtering - inverted its gonna be good integration with the mains .

So…. Often are subwoofers just wrongly integrated with the wrong order of HP / LP filtering .

One has to study and try different orders of crossovers with a dsp to learn how it all connects and how it sounds.
It took me three years of DIY with subwoofer integration setups with dsp and different subwoofers to learn how it works .

The thx crossover 24/12 dB inbuilt in home theathre receivers is a compromise, and not a very good one with music.

 

[Boes]

Actually I had a similar question.

If people are spending a good amount of money on a DAC like Holo May or Denafrips Terminator ... it doesn't seem to make sense to me that the Sub is using a DSP and basically converting the analog signal back to a digital signal, do the processing in the digital domain and then convert the digital signal back to an analog signal.

Introducing an extra ADC / DAC conversion in the audio path doesn't make sense. The ADC / DAC that used in the sub is inferior to e.g. Holo May and Denafrips Terminator.
A low pass filter and gain can be implemented using analog components like REL does.
And the other EQ / room mode features the DSP offers should in my opinion be done on the source in the digital domain, before it goes in to the DAC.

Any thoughts?

 

[Tangband]

Actually I had a similar question.

If people are spending a good amount of money on a DAC like Holo May or Denafrips Terminator ... it doesn't seem to make sense to me that the Sub is using a DSP and basically converting the analog signal back to a digital signal, do the processing in the digital domain and then convert the digital signal back to an analog signal.

Introducing an extra ADC / DAC conversion in the audio path doesn't make sense. The ADC / DAC that used in the sub is inferior to e.g. Holo May and Denafrips Terminator.
A low pass filter and gain can be implemented using analog components like REL does.
And the other EQ / room mode features the DSP offers should in my opinion be done on the source in the digital domain, before it goes in to the DAC.

Any thoughts?

Yes - its not a good idea to combine analog passive crossovers with digital dsp processing, when the dsp can do the crossover as well in a much better way.

The best sound result youre gonna get with an all digital solution, without analog coils , resistors and capacitors.

This means using an active loudspeaker with dsp crossovers and a digital signal from the source directly into the dsp crossover. No A/D converter needed or used. The room correction is also best to do in digital domain without conversion to analog being done anywhere in the signal path.

Not many customers have heard such a system , its only a couple of brands that uses this technique extensively. Genelec with the 83xx series , Linn products with exakt, and GRIMM audio is a couple of brands that uses this technique to get a better, clearer sound. And yes - the sound is better than using analog, passive loudspeakers.

I see a lot of people trying to combine passive loudspeakers with dsp roomcorrection for thousands of dollars and I always ask: why ? Massive gains can be had with an all digital solution, including the loudspeakers crossovers. Throw away those sound hostile passive coils in the signal paths ! A passive loudspeaker sounds less dynamic, less clear and less pitch accurate than a good, active dsp-loudspeaker with everything in the digital domain.

About combining dsp roomcorrection with passive loudspeakers : - The risk is rather high that you will end up getting the worst from both analog and digital, if you try to combine those.
This is just an advice on how to get real improvements of the sound.
Been there, done that.

—————

If 100% transparent is the number 10 , and worse is a lower number, it can be like this:

A passive loudspeaker with passive crossovers with an A/D-D/A dsp roomcorrection:
(100% transparent would be the number 30.)

Digital source : 10
Passive crossover: 6
dsp roomcorrection with A/D and D/A : 8
Sum: 24

or

An active dsp crossover-speaker with digital roomcorrection and no A/D conversion:
Digital source (1) with both dsp crossover (2) and dsp roomcorrection (3) done in digital domain: total = maybe 9.
Sum: 29

As you can see , a better more clear sound can be had with the all digital solution.

 

[DVDdoug]

I scanned the article and I'm pretty sure it wasn't written by (or edited by) a DSP programmer. I'm not a DSP programmer either but I've done some programming and I've studied DSP a bit.

There are trade-offs with any filtering (analog and digital). They talk about "speed" and delay and that can be one of the trade-offs, but a digital filter doesn't necessarily have more delay than an analog filter.*

Anything that can be done in analog can be done digitally. That's a trade-off too... If you are filtering an analog signal, with a "simple filter" it's probably not worth the cost of adding analog-to-digital and digital-to-analog conversion and a processor.

The additional A/D and D/A conversion doesn't hurt sound quality and it might be "technically better" than analog circuitry which is also imperfect. (But there's no reason for either to have audible side-effects.)

The main advantage of digital is that you make a more advanced filter simply by adding more code. More code doesn't cost more money (unless you "push it" to where you need a more powerful processor). Of course with hardware you have to add more components and that costs more money.

I had compared a REL S5 SHO to a SVS SB4000 a while back

Different speakers sound different.



* Computers add delay (latency) but that's related to the multitasking operating system. Computers can't process audio continuously so buffers are used on the input & output to keep the audio flowing in & out smoothly. Buffers add latency. A few milliseconds (or many milliseconds) of latency isn't a problem unless it's audio/video and the audio is out-of-sync with the video. (The digital filters in a subwoofer, or in a MiniDSP, don't have an operating system.)
.

 

[Tangband]

I scanned the article and I'm pretty sure it wasn't written by (or edited by) a DSP programmer. I'm not a DSP programmer either but I've done some programming and I've studied DSP a bit.

There are trade-offs with any filtering (analog and digital). They talk about "speed" and delay and that can be one of the trade-offs, but a digital filter doesn't necessarily have more delay than an analog filter.*

Anything that can be done in analog can be done digitally. That's a trade-off too... If you are filtering an analog signal, with a "simple filter" it's probably not worth the cost of adding analog-to-digital and digital-to-analog conversion and a processor.

The additional A/D and D/A conversion doesn't hurt sound quality and it might be "technically better" than analog circuitry which is also imperfect. (But there's no reason for either to have audible side-effects.)

The main advantage of digital is that you make a more advanced filter simply by adding more code. More code doesn't cost more money (unless you "push it" to where you need a more powerful processor). Of course with hardware you have to add more components and that costs more money.

Different speakers sound different.



* Computers add delay (latency) but that's related to the multitasking operating system. Computers can't process audio continuously so buffers are used on the input & output to keep the audio flowing in & out smoothly. Buffers add latency. A few milliseconds (or many milliseconds) of latency isn't a problem unless it's audio/video and the audio is out-of-sync with the video. (The digital filters in a subwoofer, or in a MiniDSP, don't have an operating system.)
.

I havent found a A/D - D/A thats 100 % audible transparent , but maybe 95% .
Yes - using 36 dB/oct crossover or 48 dB/oct filtering in a dsp can be easily done, and very hard to do passive.
Latency doesnt matter if you use an all digital crossover, in an active loudspeaker. If you combine digital with analog though, there might be a problem because the digital signal have some latency, the analog dont. If the latency of a dsp subwoofer crossover is , in reality maybe 6 ms , its the same as having the subwoofer 2 meters behind your main speakers. Can you hear this ? It depends on how high in frequency you use the crossover. Maybe inaudible at 40 Hz but certainly audible at 100 Hz.

 

[Tangband]

* Computers add delay (latency) but that's related to the multitasking operating system. Computers can't process audio continuously so buffers are used on the input & output to keep the audio flowing in & out smoothly. Buffers add latency. A few milliseconds (or many milliseconds) of latency isn't a problem unless it's audio/video and the audio is out-of-sync with the video. (The digital filters in a subwoofer, or in a MiniDSP, don't have an operating system.)
.

I wonder - why is it a problem with latency in audio/video and not a problem in music ? The opposite is infact true.
- Therès often more audible with to much latency in a dsp subwoofer than what you can see in bad lipsync in audio/video.

 

[FeddyLost]

Now looking back it could have been my room and my set up skills that were the problem.

We'd like to know more about your setup and your tries before making any assumptions.
Also "sounded better" must be expressed more detailed for analysis.

IMO any sub integration must be done with measurements, otherwise it's extremely troublesome to do it right.

I've found REL S/5 test data and most probably it's a decent unit, but nothing special for its price and size.
SB-4000 looks very good, but might have at least GD issues as smaller SVS subs.
There is a lot of possibilities to make any sub sound wrong with existing 2.0 setup.
Especially running 2.0 full range with sub augmentation.

 

[Boes]

The additional A/D and D/A conversion doesn't hurt sound quality and it might be "technically better" than analog circuitry which is also imperfect. (But there's no reason for either to have audible side-effects.)

Any conversion has and impact on the signal. The quality of the ADC / DAC also has a measurable and audible difference (please check all the threads about DACs).

My point is basically that in a 2 -channel system (streamer/PC -> DAC -> pre-amp -> power amp -> speaker). It doesn't make sense to add a DSP in the speaker/sub woofer that does another ADC / DAC conversion. The DAC in the subwoofer in most cases would be of lower quality than the main DAC in the system.
This doesn't make sense IMHO.

Having an analog low pass filter (speakers also have analog cross overs, be it static) in the Sub and any potential EQ or room correction done in the streamer / PC on the digital source, seems the way to go.