Explaining analog preference

I noticed a number of questions on a previous post about bringing back analog hearing aids (as well as other related posts) asking why certain users consistently prefer analog hearing aids. This article of mine may help explain:

5 Likes

The problem may be that of the dozens of audiologists and technicians I’ve met or spoken with over the years, not one has been hearing-impaired…

Right on. Nor are they audiophiles.

It’s like the mass transportation systems in the USA are such a useless, demeaning mess because the people who run and manage them have never been in a bus or train.

I hope for your sake and all those like you that someone starts marketing analog HA’s some time.

I’m going to be thinking about this. I’ve been doing so much digitizing and signal processing lately that I may take a stab sometime at making a digital amplification/equalization system that uses techniques that preserve phase relationships of mixed signals. I may lean on people who have this gift of hearing that dimension in the music to test it and see if my ideas affect the issue. If someone doesn’t beat me to it.

But the simple solution is actual analog hearing aids.

1 Like

So here’s the basic problem.

HA’s are full of filters. Filters affect not only the magnitude of the sound in a certain frequency range, but they also cause phase shifts in the waves that have different frequency contents. I believe it’s these subtle alterations in the frequency content and interactions that sensitive people hear that negatively alters their perception of the sounds.

To illustrate, here are plots of an analog low pass filter and a digital “FIR” filter (Finite Impulse Response). The top plot in each case is the magnitude plot – the sound pass through level at each frequency. The bottom plot in each case shows how much the waves are shifted (0 - 180 degrees) at different frequencies.

All filters have phase shifts; it’s a natural effect of filtering. But as you can see, the character of the phase shifts are radically different. The analog phase shift is smooth and curved while the digital phase shifting has jagged small effects and a rigid-slope in the filter cutoff region.

Though one not skilled in electronics can’t interpret the direct effects of the difference, it’s easy to see that there is a significant difference, and I believe that sensitive people can hear it.

Analog Low Pass Filter:

mXRXo

DIgital (FIR) Low Pass Filter:

1 Like

Thanks – that’s very helpful and certainly explains a lot. I agree that analog HAs would be an easy solution, though I’d be glad to be a guinea pig for developing any aids that work better with music!

1 Like

Someone alredy beat you on the issue with the filters. It’s not your idea but of mine fitter, which I mentioned earlier. Some wise Pole.

Those comparisons aren’t like for like scales: at -40dB there’s no distortion evident on either scale - the digital one actually goes further (-60dB) and is stable, but the HF analogue isn’t shown.

Do yo actually know what the Phase is? it’s the offset (delay) of the leading edge of the soundwave relative to the original signal, the sawtooth of the upper frequencies isn’t necessarily an issue as the relative offset is a couple of uSec and is smoothed by the final comparator stage of the output relative to the input.

People don’t like digital for all sorts of reasons. Usually the real reason is that most analogue circuits use quite rounded peak compression, while the digital circuit retains the original peak shape more faithfully. Loudness growth is also an issue, lots of hearing aids are fitted to amplify adjacent octaves to similar power levels as there’s a belief (in some schools) that this gives the optimum intelligibility - however (my own personal pref.) this doesn’t necessarily sound ‘nice’ and I’m sure results in longer term Audio fatigue.

4 Likes

I am not one of those who believe that analog sound is better than digital sound. I listen to a fair amount of MP3s and they sound better to me that their analog equivalents. I remember the days of vinyl records when there were scratches and pops even on brand new records. I also remember the hiss of recorded tapes. I think people don’t like the sound of digital hearing aids because it is so processed. And, the reason it is so processed is because there are a lot more things you can do with digital than analog. Just to name a few, echo supression, feedback management, instant microphone directionality changes, instantaneous processing to pick a voice out of noise. When hearing aid engineers discovered digital signal processing, it was, to coin a phrase, a paradigm shift. They would be very resistant to go back to a technology they probably consider horse and buggy. I don’t like the sound of digital hearing aids either, but it isn’t because they are digital.

2 Likes

Got picked up by Boing Boing!

1 Like

Good observation about the scales. I wondered if people would interpret that.

It isn’t the jaggies I’m really concerned about (though I did pay lip service to it). It’s the shape of the phase shift function. That can alter the perception of the sound.

1 Like

Well that’s a very strong statement that “quality is plummeting” when most digital HA users are perfectly happy with better conversation comprehension and even some music lovers giving good testimony.

Not everyone can hear the difference. There is a HUGE swath of the population who enjoys listening to MP3’s and lossy streamed music services like Pandora and Spotify. Even SiriusXM.

This reminds me of the outcry when people started using transistor-based amplifiers rather than tube-based ones. Purists said (and still say) the transistor sound quality was awful.

But very few people are vinyl/tube purists. And I’d wager a few fewer of those can actually hear the difference.

So I don’t dispute people who say they can hear the difference in digital HA’s and those who don’t like them at all.

But it’s not a true statement at all that “quality” is “plummeting”.

This is why I don’t follow the sensationalist tripe in Slate or BoingBoing.

Thanks for the response. Well, as they say, “the devil’s in the details.” I’d like to learn more about who that is and what they did.

Do you mean your post in “Bringing back analog hearing aids” thread?

So you think there’s something about the Widex filters? What Widex model was that?

I read your article in Slate, and thought about this for a while. I decided to comment, but as a fair warning, this is probably going to be along post.

Hi Fidelity Audio Journey

As some background I thought it might be helpful to detail my journey through the high fidelity world of audio. I have not worn analog hearing aids, but I have kind of done the gamut in hi fi equipment.

My first experience was with 78 RPM record player when I was a kid. The stylus was like a sharpened up finishing nail, and the audio quality was simply terrible. I couldn’t understand how adults could listen to that kind of stuff. When they were not around we used the records as Frisbee’s, although we did not know what a Frisbee was at that time.

Next in my teen years I got a battery operated LP player, and it was pretty bad too. It could not keep a steady speed. There were 8 tracks around about that time, and they were pretty bad, and unreliable. You would see many of them thrown out on the side of the road. During university years I bought one of the first Phillips cassette players, no Dolby, basic tape, and terrible hiss. I knew that was not the answer either. Had friends that used Teak reel to reel players, but the only real source of audio input was still LP’s.

After graduation and newly married I bought my first real high fidelity system, which as funds were limited, it was pretty basic; Sony 30 watt receiver, Dual turntable, Shure moving magnet cartridge, and Pioneer speakers. It was easily the best sound I had heard so far, but was still not like hearing live music. I wanted more.

I became a reader of The Audio Critic of which there are still copies of the now defunct publication on line. They maintained a “A” reference system equipment list, as well as a “B” best buy list. I certainly could not afford the A stuff, but did manage to accumulate much of the B list, all with the objective of getting a live like reproduction out of LP’s (Analog devices).

Fidelity Research FR1 MK3F cartridge - They were hand built in Japan and featured silver wound moving coils along with a line contact diamond stylus. It was an Audio Critic reference B component. Later after kids, the cartridge got damaged, probably from curious little hands. I had it rebuild by Peter Leddermann of Soundsmith in the US. He used a line contact diamond and a ruby cantilever.

Grace G-707 Tonearm - I wanted simplicity and a straight tonearm as my view was that curved complex arms were only for sales purposes and made the tonearms heavy and perform more poorly.

Rega Planar 2 Turntable - I bought a bare turntable that again was very simple and basic. It used a weak syncronous motor, a rubber o-ring drive belt, and a very heavy glass platter. The idea was to make the speed control (wow and flutter) negligible, without the complexity of the electronic speed controls etc on some.

Marcof PPA-1h moving coil cartridge step up pre-pre-amp - Another Audio Critic reference B choice. It was claimed to be as good as almost all transformer systems.

David Hafler DH-101 Pre Amp - Another Audio Critic recommend amp. It is amazing that to this day it has accommodated my requirements, including now a CD player (which didn’t exist them) on the Aux input.

David Hafer DH-200 Power Amp - This was again another Audio Critic reference B component, with the main claim to fame being that it used MOSFET power transistors that were said to perform like a tube amplifier. My theory at the time as that good amplifiers could put more power into a low resistance load, and also have a high damping factor. This avoided clipping and produced a nice tight bass when used with low resistance speaker wire.

B&W DM7 MKII Speakers - Were highly recommended by Audio Critic but were said to be too expensive to include in the reference B system. They were claimed to have been developed with sonic excellence as the main priority and cost not considered.

That is it. That was/is my reference vinyl playback hi fi system. My point in this is that the equipment was all carefully selected and maintained over the years. It is not just a record player. I think it did, and still does reflect something right at the top of the heap when it comes to sonic excellence out of an analog vinyl LP recording.

Next the digital story:

I am on probably my third CD player. The current one is the Marantz CD-6004. It kind of fits into the most bang for the buck category as one can spend much more than this on digital. However, it is probably 1/3 the cost of my vinyl LP component system.

Conclusion?

Despite having a high quality vinyl playback system, the digital CD playback system beats it hands down. On a good recording the background noise is zero, the dynamic range of sound is much greater, the left to right imaging capability is better, and in general the sound has a much more open air sound to it. In my view digital in sonic quality is a step ahead, not back. That said I am not throwing out my vinyl or the system. It has a place and a lot of it is nostalgic. If I have a recording on both CD and in vinyl, I always reach for the CD first.

The Audio Critic is now defunct and the originator, Peter Aczel, not even alive today. However, in one of their last editions of the magazine they included a “10 Biggest Lies in Audio”. It is worthwhile reading in particular lie #3 The Antidigital Lie.

Long story, but I can’t accept the idea that digital hearing aids are inferior and regressive. It simply makes no sense.

6 Likes

Here is a good article that talks about the issues in using digital for hearing aids. It does focus on the microphone analog to digital input part of the hearing aid, but I think is helpful in understanding the problems. It is by Widex and of course leads to why they do it best! The other limitation is that it is about 7 year old technology. Widex is proud they have achieved an input dynamic range of 113 dB, and claim they were first to do it. Today I believe a number of manufacturers can do that or better. I recall that the ReSound Quattro claims 116 dB. I also understand some are using more than 16 bit depth to achieve it.

Designing Hearing Aid Technology to Support Benefits in Demanding Situations

1 Like

Thanks for the responses here. I should stress that I was not trying to write an anti-digital manifesto. I do plenty of listening with digital, though ceteris paribus an analog recording will usually win out for me, and it’s not close. I should note though, that digital streaming and CDs sound substantially worse through digital hearing aids (presumably a function of the original signal having to be processed twice). Perhaps ironically, I was generally much happier listening to digitized music through analog hearing aids.

I have heard good things about the latest Widex model, and that may be the best available option at this time. That article, however, seems to make many of my points for me: namely, that digital conversion introduces a number of different problems unique to digital that the engineers then have to solve. And while it’s good to see that at least some manufacturers understand the headroom problem, it would be better if we could soon see digital aids with a greater bit depth and sampling rate than that of CDs.

The Audio Critic and pretty much all of audio rejected the Super Audio CD as unnecessary and having no real life improvement over the standard CD audio quality. See the article on pdf page 27 in this Audio Critic link for a detailed explanation. SACD to my knowledge has essentially died due to lack of recordings in the format, and lack of interest from the public.

There is nothing lacking or lost in a CD quality recording. When you sample at over double the maximum frequency to be recorded there is no loss in fidelity. And when the bit depth is sufficient for the headroom needed, there is no loss there either.

Agreed there is no loss in playback of CD, as long as the error correction keeps up with the hard error rate.

There is some limitation in dynamic range of content material given the 16-bit sample depth (effective 96dB dynamic range). SACD had an equivalent bit depth of 20 bits, which is more, but turns out to be lost on the vinyl purists. Vinyl records dynamic generally top out around 65dB, which is beat by standard CD audio.

All that being said, in the land of digital processing the phase relationships of the component frequencies are reshaped and cause phase delays which render different subtle qualities to the sound. Since SACD was processed with similar signal processing techniques as regular CD audio, these effects wouldn’t be mitigated.

I’m not surprised there was little interest in SACD given the costs and the incompatibilities.

But given that the dynamic range in human hearing is about 140dB, one might think that improvements on HAs’ 16-bit standards (96dB) improved HA performance. But given that most people’s correctable losses start at 20-30dB, 96dB range of 16-bit sampling should meet or exceed the limits of the rest of the ear’s processing abilities.

What is interesting, and I recall was detailed in that Widex article is that it seems digital hearing aid practice seems to be to use the 96 dB dynamic RANGE of the 16 bit depth, starting at 0 dB. The first 20 dB or more gets wasted on transmitting noise inherent in the microphones. So the easy fix, which a number have done now is more more toward the CD practice which is to toast the first 30-35 dB of microphone input, and then use the 96 dB range starting there to get 126 to 131 dB or maximum peak input without distortion - as in digital recording. This gets closer to what human hearing can digest. The reason I suspect this is not done with hearing aids, is that the ANALOG microphones used in hearing aids are far inferior to the microphones used in recording studios. From memory they may not be capable of more than 120 dB or less of peak recording level. So, it is pointless to use a bit depth to do A/D conversion of something that cannot make it through the microphone without gross distortion.

Thanks – very interesting. But isn’t this yet another example of digital creating problems that it then has to solve? I don’t at all deny that analog has its own problems (both with respect to hearing aids and generally), but again we presently have a situation where we are effectively stuck with digital. I don’t see why analog should have to conclusively demonstrate its superiority simply in order to exist at all.

1 Like

Electret mics can easily cope with that input range. The limiting factors are the noise floor due to Brownian motion and the A/D stage of the chip. Modern aids sample in plenty of range (16 Bit and above) to do it without input peak clipping.

This article from Widex, which is a bit dated says “today’s analog hearing aid microphones have a dynamic range from around 17-115 dB SPL”. This falls short of live performance peak levels, and levels that a properly recorded CD can have.

Evaluating Hearing Aid Processing at High and Very High Input Levels