Limitations and Advantages of DSP (digital sound processing)

​​

In the beginning, DSP’s didn’t exist, and is fairly recent in the HiFi timeline. Therefore, some people wrongly assume that it departs from the ultimate audiophile goal. Yet… DSP is just a logic continuum.

 

Sound engineers understood early on that sound correction was necessary for high fidelity reproduction. In the day of gramophones, a horn was used and no correction to the sound was applied. Both recording and playback were direct. Sure enough then, you end up with a device optimized for voice and lacking bass and treble (pass-band is ~500 Hz - 2 kHz) It worked a lot better than Morse code, but was not audiophile by any means. Gramophones were a good start for voice reproduction. Just not enough to rest on.

 

Fast forward 40 years, to newer/better disks.

With improvements in technology, correction was applied on LP disks to make them more linear. These corrections are still in use today on 33 RPM LP’s (RIAA curve corrects non- linearity of the friction based vinyl/needle interface medium).

• Vinyl applied RIAA equalization is reversed in the preamp to restore the original to the end user. Resulting in a better sound by minimizing the vinyl/needle interface limitations

• Of course, vinyl playback receives more than just simple RIAA equalization to facilitate information transfer. 

These include:

• dynamic compression filters. (hell yea, they do)

• mono summation of low frequencies @ 60 Hz – 150 Hz range (yikes)

• low pass filters

• rumble filters

• high pass filters

• and so forth.

Yes, the vinyl sound is highly engineered before it is even pressed and sold to you.

 

Like RIAA correction, the DSP can be seen as an improvement.  

 

Yes, once the sound engineer releases a mastered track on any format, there’s absolutely nothing he can do afterwards.

If you listen to his track on a bass heavy headphone like Beats by Dr.Dre, or use a cheap tabletop radio having no low frequency reproduction ability, what you hear is what you get. And often, it’s not a linear listening device that’s playing the song.

 

For the Audiophile enthusiast, the steps were simple:

• Buy the best speaker that fits the need, the budget and the room.

• Buy an amplifier that can effortlessly drive the speakers.

• Hope for good room acoustics to work like black magic (impossible unless the room is heavenly treated, a shelf in a corner is an insignificant acoustic treatment device)

• Endlessly change components to dial in the exact sound. (bright vs bass heavy preamp, soft feet tweaks, xyz “transparent sounding” interconnects, new better sounding amp, and so on)

• Frustrated that the overall sound is never good enough, and ultimately accepting an average listening experience.

• Considering leaving the hobby because every time an audiophile hears live music, especially an orchestra symphony, he realizes how far HIFI is from the real thing.

• Extremely far from real sound indeed, this should be a wake-up call.

​

One option is now available:  DSP. Change your sound and correct your room!

No endless box swapping necessary, the DSP tunes the sound more than an amplifier swap can ever do.

 

The ancestor of DSP was the venerable equalizer.

The EQ’s permit the user to dial certain frequencies band up or down, to achieve better overall sound balance. (hopefully correcting the frequency in order to obtain a flatter frequency response, please do not use the "smiley face" setting...)

Studio and mastering consoles for live sound make good use of analog equalization. Old albums you listen to were all passed through an EQ (Led Zeppelin, The Beatles…you name it). They’ve all  been “optimized” by the recording engineer before it was sold to you.

​

------------------

​

Side note:

If you buy Symphonic music of some prestigious city, recorded in 2015, expect the CD to be nicely balanced as the average target audience of such music usually have a high quality playback system.

On the other extreme, if the music is "dance music from the 90's", the average target audience then was broke teenagers using cheap playback systems, to compensate cheap loudspeakers. The sound engineer boosted the bass in order to give a "better sound" on crappy electronics.

Lots of 80's pop music is "mastered" on poor sounding Yamaha NS-10 monitors in order for the  engineer to obtain an acceptable sound in the target audience crappy early stage digital CD and harsh solid state amplifier of the days. 

​

Ever wondered why 1975's release of  "Layla" by Eric Clapton sounds like crap while 1956's recording of "Ella and Louis" sounds wonderful. That is why.

They were mastered with different "ideologies" and as a user, you are stuck with a great Clapton song that sounds like crap. 

​

------------------------

 

EQ worked in the old days so well that you can still get basic three band EQ’s in 2020.

In your car radio for example: bass/mid/treble are EQ's, and up to 40 bands is sometimes available on portable music players. The 20 band EQ’s are especially useful and popular, as they can do wonders, if it’s a quality brand. Of course, one must physically set the EQ correctly to improve fidelity. An incorrect setting will make the sound worse at great lengths. DJ systems blasting away at weddings is a prime example of bad sound and misuse of EQ’s.

​

Being analog, EQ’s do change the phase, create minimal distortion, and will consequently add delays through different phase outputs.

But they can't create "phase shift free" delay. This is essential for time aligning loudspeakers.

Audiophiles tend to snob EQ's as they “change the sound” and aren't true to the artist’s intent.

TOTAL NON-SENSE? I think so. Don’t forget that your music was already "processed" through either a DSP or an EQ.

​

EQ’s make up for what the sound engineer couldn't do… correct for YOUR room.

Phase a delay? 99% of speaker’s aren't launching a perfect waveform to start with, as they’re unable to recreate a square wave. DSP can fix that.

The secret with an EQ is to know how to dial it in properly, and to purchase an audiophile or studio quality unit.

They are great, and if it wasn't for DSP, I’d use EQ as well.

​

Now, enter DSP. (DSP are basically EQ’s on steroids)

• DSP’s can correct for the loudspeaker deficiency (up to the loudspeaker mechanical limit)

• DSP’s can correct for the erratic room acoustic behavior. Well, sort of. They help a lot but are not a “fix all” solution.

• DSP’s can be Phase Perfect (no phase shift being created like in the analog world)

• DSP’s in active mode, can correct time delays and make speakers produce a coherent wave launch, and reproduce square waves.

• DSP’s cause no harm when used correctly.

​

Take DEQX for example. They have been selling great DSP products for years and have a cult following. I recommend checking them out. It's a mystery to me why they aren't more popular. I know they’re not cheap nor dummy proof. They are intimidating actually.

Most DEQX offer six channels and can correct speakers in passive mode, or can go one notch above and you can use them in an active system, provided you remove the crossover in your speaker, and are willing to invest the time to reap the huge benefit of an active system upgrade.

 

Typical DSP has four possible configurations.:

1. Active crossover + speaker correction

2. Active crossover + room correction

3. Passive crossover + general speaker correction

4. Passive crossover + room correction.

5. Active or passive crossover + room correction +speaker correction. (the best you could do)

 

In their advertising, DEQX’s push more on point #4. Easy to figure out why as the potential for unit sale is the greatest in that category.  Rightfully so, as it’s easier to do, and cheaper as one doesn’t need to buy multiple amplifiers. It also preserves loudspeaker value by avoiding dismantling the internal crossover and keeping them “stock”. Results are the most limited of the five options, but significant improvements are possible.

 

Some people would prefer #3, Passive crossover + speaker correction, as they’re trying to get the speakers as linear as possible, often measuring them outdoors to eliminate room acoustic. This is more of a purist approach IMHO. Getting a speaker "perfect" is a noble goal, but once the speaker is in the room, anything goes. The outdoor measurement is useless once indoor.

 

I prefer to use option #5 . Active crossover + room correction + speaker correction. Because I listen indoors, and my room is part of my chain (so is yours), it only makes sense to correct for the room and aim for best possible results at the listening position. I don’t care what frequency response I get outdoors or at 1 meter away... I listen indoors @ ~2.5 meters and this what I work with. I only care about my listening position. For active configuration, if you've been reviewing my site, you’ll already know why I favor this approach.

 

Correcting for the room implies that one measure at the listening position, optimizes the frequency response and phase only in that very small spot of the room. (AKA: the sweet spot). The rest doesn't matter much, or as much.

​

• The funny thing is that a live concert event does the exact opposite approach. They disregard absolute fidelity at “X” location. Rather, they try to have uniform sound across the whole seated area. It makes sense, they have 1500 people to please, not just the ones sitting in the middle. No matter where your seat is situated, you want to hear the concert and total sound radiated should be balanced at any seat.

• Movie theaters are a bit of the same, as they vastly compromise absolute fidelity in a tiny location in order to get an "entire room sweet spot", so the full audience experiences acceptable sound.

• If movie theaters were maximized for one single seat, they would sound much, much, much better, but would go out of business quickly as one customer at a time isn't enough to generate profits.

• Goals are very different and require much different approach.

• This why Pro-Sound loudspeakers, with extra-wide dispersion horns, think Altec Voice of the Theater, are oriented for "public use".

• Such designs are generally a bad idea for use in a domestic environment.

 

In their man cave, the audiophiles are blessed, as they don’t have all the constraints that the professionals have (live event, theater, arena and big room overall sound uniformity being priority #1).

Using a DSP, the audiophile can optimize the sound at the very spot he listens to, and completely disregard what happen 3 meters away from his seat 

​

Life is good.

 

DSP IS NOT a replacement to room acoustics.

If there’s no treatment on the ceiling, there will be a reverberation, which will create a null at a certain frequency.

 https://mehlau.net/audio/floorbounce/

This null creates a void. It is simply a cancellation of energy. DSP’s can't do anything for that, with the only fix being room treatment.

This type of cancellation happens on four distinct areas (ceiling, floor, left wall and right wall)

With these four main issues DSP’s can't correct, one may question why use a DSP at all? They do a lot for other problems.

​

On the flip side, if there’s a summation (often bass) in a certain frequency, the problem is easy to fix with a DSP. Simply dial the said frequency down few dB and your frequency becomes flat again. DSP can correct peak but can't correct valley.

DSP’s can also compensate for uneven speaker response. Often, systems in a room are weak in the 60 Hz – 300 Hz range. DSP’s can correct those frequencies to obtain a flat response. This is HUGE, as it’s right in the human voice fundamental frequency. You wouldn't believe how much more realistic a system can sound when the vocals are bang on!

 

Fine line with DSP.

Speakers often have weak bass in the listening area, (20 Hz – 200 Hz). DSP’s can bring system bass very flat at the listening position.

That would be the ultimate gain but unfortunately, speakers often can't mechanically reproduce

20 Hz - 60 Hz at significant SPL. Even if they could, the harmonic distortion would be exceedingly high, if not excessive potentially resulting in the speaker voice coil cooking itself.

Boosting the small 5” woofer in the Kef LS50 can reproduce 30 Hz. However, the speaker will clip quickly as it reaches the mechanical limit. The Laws of Physics still prevail, as 5" can only move so much air.

The power amplifier will be taxed as well, as 10 dB or more of gain may result in electrical clipping, and could cook the woofer voice coil.

Lost of potential problems to watch for here. Use the DSP with caution and watch for potential out of capacity demands on the system.

It’s a trade-off between maximum SPL. Frequency flatness and distortion are to be considered when equalizing in the 20 Hz – 60 Hz range.

If woofer displacement is limited, cutting higher will bring a better compromise.

​

Yet, despite limitations, I couldn’t imagine having my system without a DSP ever again.

I'm sold :)

​

​

For setting the DSP, one rule trumps them all:

Treat the room as much as you can and apply DSP correction on remaining areas of concern.

The other way around doesn't work.