Distortion problems

[Radio Wrangler]

The distortion in human hearing has been known about and even exploited for a long time.
Look on a large (but not extremely large) and you will find a stop labelled "Quint" This brings into play a rank of pipes pitched at 1.5 times the frequency of the ranks of bass pipes.

Played loudly, a bass pipe and its quint intermodulate in the listener's ear causing the perception of a tone at 0.5 times the frequency of the bass pipe. In other words it fakes the octave below its longest pipes.

Sneaky, huh?

So there's nothing embarrassing or worrying about distortion in your ears. We ALL have it. But like any distortion effect it is level-dependent and will fade away except at high levels.
And therein lies the clue to one way of tracing distortion... change the levels at different stages in the chain and if the distortion doesn't change, it was most likely created in a place whose level did not change.

David

[Radio Wrangler]

If it sounds fine with different speakers, and yo tried interchanging your amps and the linn amps are still in use on the Tascams, then I begin to suspect the drivers in the Linn speakers. It the Tascams are active in their own right, then scratch that thought, or if you're using a different amp for the Tascams.

We crossposted so I hadn-

't seen your last post when composing mine abve

David

[mhennessy]

Hi Linnovice,

That's good to know - I hope the test goes well for you. I certainly know that I'm not as bothered by CRT line whistle (15.625kHz) as I used to be. In fact today I find it strangely reassuring! None of us are getting any younger

Now, those Tascam speakers are active, so they've presumably picked up a signal from the output of the Linn or Technics preamps?

If so, that does seem to point back to the Linn power amps or the speakers themselves.

Which is confusing because you've heard good sound from the system when playing back tapes. It's almost as if certain CDs are reaching a certain voltage, and that's enough to cause the distortion (that would likely be a type of clipping).

1. My suggestion would be to find a CD that reliably exhibits the distortion via either pre-amp, and without changing any settings on the pre-amp, change over to the Tascam speakers and see if the distortion is present on them.

2. Another thing to try - if you have the right adaptor leads, etc - is to try driving both the Linn speakers and Tascam speakers together. If the Linn amps are excessively loading down the preamp, then the sound heard though the Tascam speakers will change from clean to distorted as soon as you connect the Linn amps. You might wish to disconnect the Linn speakers from their respective amps while doing this* as it might be hard to judge the sound from just the Tascams.

Cheers,

Mark


* I assume that the Linn amps are happy to work with no load connected - most solid state amps are.

[Linnovice]

How odd is this? A random selection, all played at the same volume levels (around 60%).
The Tallis Scholars Victorian Requiem on CD = Distortion
Gregorian Chant on CD = Distortion
Gregorian Chant on Tape = Distortion (significantly less though)
Gregorian Chant streamed = Distortion (again much less than CD)
Boy in a Bubble, Paul Simon = No Distortion
Hendrix, All along the Watchtower = No Distortion
Richard Thompson, 1952 Vincent = No Distortion
Cat Stevens, Anglesey = No Distortion
Pink Floyd, Aircraft crash on The Wall = No Distortion
Metallica, and nothing else matters = No Distortion
Brahms, 4th Symphony = Distortion

What could possibly make the situation that the system introduces this distortion? Where I’ve indicated ‘No Distortion’ there is absolutely no trace of any.

I think this will turn out as something I have to live with. Am I expecting too much of my kit?

[Linnovice]

The Linn Kairn has three pairs of output to drive power amps. I have the Linn amps and speakers connected to one pair. The Tascam connected to a second pair (in a different room though).
I’m wondering if because the choral music is usually recorded in a church/chapel/cathedral environment whether the inherent back echoes in the buildings are what is causing the distortion on my system. There is a harmonic reflection that is vibrating in sympathy with something in the system and causing the problem.
A wild shot in the dark but I can’t think of anything else.

[PJL]

Might be worth getting a second opinion to see if someone else experiences the same effects?

If they do then it could be a resonance problem either something loose in the speakers or even something positioned close to the speakers.

[mhennessy]

Interesting list - distortion for anything classic, no distortion for non-classical. How peculiar. I doubt it's anything to do with the reverb - it will be at a relatively low level compared to the main event.

Question: When you heard the distortion, which speakers were you listening to?

[Linnovice]

The Linn Kabers.
Don’t worry, there isn’t anything untoward with the Classical/Non Classical scenario. Apart from, when you think of it, the arena in which it is recorded. The Non Classical is studio produced. The Classical in a much larger arena with accompanying resonances?
A thought.

[mhennessy]

Well, the electronics won't know or care how much of the acoustic environment was captured by the recording. The reflections cause reverberation - not "resonances" as such - and that gives us a sense of the acoustic space. But as I said, this is a relatively low-level, subtle thing, as anyone who's ever added artificial reverb to a singer or instrument will know.

What might be significant about the classical music is the generally wider dynamic range. It has quiet bits as well as loud bits. A lot of "contemporary" music has a much lower dynamic range - the quiet bits are not as quiet as the classical recordings. Having said that, the recordings you mention are actually quite good for their respective genres.

But without hearing it for myself, this might lead me down a path of excessive speculation.

So, the distortion was present on the Linn amps and speakers.

Question: Does this mean that the same music was completely clean when heard through the Tascam speakers?

[Linnovice]

Is there a method of connecting an analogue oscilloscope to the speaker inputs? ie. if I pull out the banana plugs far enough to get a crocodile clip in the shaft. Would it damage the scope and/or the driver?

[dave cox]

You need to be careful connecting a scope to anything unless you know which points are at ground potential otherwise there is likely to be smoke! For audio, I omit connecting the probe earth clip, at first, until I have verified the earth points with an ohm-meter. For many amplifiers one of the speaker connections will be earth but this is not guaranteed. It any case, it will need to be a gross effect to see it on a scope so sadly I would bet you will see nothing amiss

I posted earlier about hearing an 'occasional slight distortion' (due to a dry joint on an output inductor) but in that case it took a THD analyser with residual spectrum analysis to ‘see’ that it was likely due to a ‘diode like’ bad connection. In hind sight, pressing on said inductor while listening to the right source material could have found the problem …

dc

[dave cox]

Another thought, if you try to use pure tones to find the problem then you need to focus on 2KHz and lower. The 4th harmonic and above are much more obvious so for a 2KHz fundamental the any audible distortion will occur at 8KHz, 10KHz, 12KHz etc. etc. quickly moving outside the range we can hear!

A different possibility is interference, commonly cause by Wi-Fi, mobile telephones, microwaves, Bluetooth etc but I would have thought that this would have been obvious.

I think your selection of track where you can / can’t hear distortion is telling, I would happily wager that it is there in all cases but is masked. This is exactly my experience! Did you try speech? I found streamed R4 (not DAB, due the economy bitrate) the most revealing …

dc

[TIMTAPE]

Originally Posted by Linnovice

How odd is this? A random selection, all played at the same volume levels (around 60%).
The Tallis Scholars Victorian Requiem on CD = Distortion
Gregorian Chant on CD = Distortion
Gregorian Chant on Tape = Distortion (significantly less though)
Gregorian Chant streamed = Distortion (again much less than CD)
Boy in a Bubble, Paul Simon = No Distortion
Hendrix, All along the Watchtower = No Distortion
Richard Thompson, 1952 Vincent = No Distortion
Cat Stevens, Anglesey = No Distortion
Pink Floyd, Aircraft crash on The Wall = No Distortion
Metallica, and nothing else matters = No Distortion
Brahms, 4th Symphony = Distortion

What could possibly make the situation that the system introduces this distortion? Where I’ve indicated ‘No Distortion’ there is absolutely no trace of any.

I think this will turn out as something I have to live with. Am I expecting too much of my kit?

As a general rule, classical recordings have wide dynamic range: pop music recordings narrow dynamic range, perhaps especially these days. On this scenario, and assuming you have adjusted your volume for a comfortable average listening level, the pop music may just sit comfortably in the middle volume range, never causing overload distortion. The more dynamic classical music may also be OK until a "forte forte" section comes up, running your gear into serious distortion.


Radio Wrangler mentioned trying different levels. Try running the pop tracks louder, to the point that they also distort. The point of distortion may seem higher because of the way pop recordings are produced, but it should be exactly the same point that the classical tracks distort. Large dynamics are always a harder test for audio gear.

Tim

[Hartley118]

I recall encountering this kind of elusive high frequency distortion some years ago. It was absent on most music, but would appear on choral or 'bright' organ music. It turned out to be due to the voice coil rubbing on the magnet pole piece. This is easy to diagnose of course in a woofer or midrange speaker, but in a tweeter the rubbing noise components can easily be masked by much of the HF music content.

In my case, diagnosis was relatively straightforward (though still stressful!) because only one of the pair of tweeters was afflicted with he problem. The only sure test is to replace the tweeter with a known good example. For a start, it would be worth using a mono test source and checking whether both your tweeters output exactly the same amount of distortion.

Martin

[mhennessy]

Yes, I mentioned this ages ago. I've been working away from home this week, so have lost touch with the ins and outs, but if this is affecting both channels, let's hope it's not both speakers that are affected. In my case, it was both, but one was much worse than others - it was a Vifa D26TG. Perhaps related to ferrofluid drying out with age which, according to ATC (who made my speakers), is a known issue with these when they have had a hard life. You could re-position the voice coil within the gap, but I decided to order replacements from ATC while I could (the tweeter had been discontinued by this point). The replacements sounded very, very different to the originals. Now that I have a measurement mic setup, I do plan to measure the old vs the new. One day...

Thinking about the list of source material, I don't think it's as straightforward as dynamic range as the Paul Simon track has a decently wide DR (the original version, at least). Either way, we need to establish if this is a problem with peak levels, or something else. That's why I kept mentioning alternative speakers earlier. The active speaker connected to the output of the pre-amp appears to point to a problem with the Linn power amps or Linn speakers, but as I said at the time, tests with a passive speaker would be a worthwhile step.

By all means connect a 'scope to the speaker outputs - but as this is not strongly dependant on level, it doesn't feel like a clipping problem. A test CD in conjunction with the 'scope would be the best way to eliminate this...

[TIMTAPE]

Originally Posted by mhennessy

...Thinking about the list of source material, I don't think it's as straightforward as dynamic range as the Paul Simon track has a decently wide DR (the original version, at least). Either way, we need to establish if this is a problem with peak levels, or something else...

Yes the Simon track would have more dynamic range on the original vinyl as well as the original CD rerelease, compared to more recent rereleases. This is the story of "The Loudness War" developing steadily over the last 30 years or so.

Even so, overall, pop records still had less dynamic range than classical records, even back then before the Loudness Wars that later developed and are still with us today. Live classical music can have huge dynamics, whereas live pop concerts are more one level - usually a loud level.


I just read that the OP used about the same 60% volume setting for all tracks played. I had assumed he had adjusted for a good average listening level, being familiar with the tracks involved and anticipating crescendo's later in the classical tracks. But if he listened at essentially the same setting on the volume control, then it should be the opposite, with the pop records distorting more than the classical, since average pop recordings reach peak levels much more often than average classical recordings. So for me it's back to square one!

[mhennessy]

The thing is, the pop CDs will have the same peak levels as the classical CDs. More dynamic range simply means lower average levels - but the peaks still reach 0dBFS.

Random classical CD: http://dr.loudness-war.info/album/view/82320
Graceland (original release): http://dr.loudness-war.info/album/view/89352

They both reach 0dBFS (actually, the Paul Simon doesn't quite, which is typical of that era), but the average/RMS levels vary according to the track.

In the case of the Paul Simon, the average might be around -18dBFS, which is the same as some of the tracks on the Tchaikovsky. Of course, some are quieter - as you'd expect - but the quieter tracks have lower peak levels too, so their dynamic range is much the same, at around 20dB.

I disagree that live rock/pop concerts have limited dynamic range - I know this because I mix them and have the recordings to analyse afterwards. When we get work experience students from local colleges, we have a tough time because they've been taught how to record commercial music in the peak-limited idiom, and so don't understand why our compressor settings are so much more gentle than they've been taught at college. We then have to have a long conversation about dynamics! Since the arrival of loudness normalisation, the loudness war has been slowly receding, and these students leave me with the message that they need to make their mixes more dynamic to make them stand out from the crowd. Hopefully their college teachers will eventually realise this too...

I do agree, however, about the volume being too high. But what many people fail to understand, however, is that often, there is little you can do about this in a smaller venue because you can't turn the drummer down! Typically, you can completely fade down the snare mic and it makes almost no different to what we hear in our local auditorium (~300 seats). If you've got an "animal" on a massive drum kit, then your only option is to turn up everything else, which isn't always ideal. We aim to average 85dBSPL at the mix position, which actually isn't all that loud (in H&S terms) for a 90 minute gig. The peaks will easily be 20dB or 30dB more, but that's OK (in H&S terms) because they are so very narrow.

Anyway, the music choice could well be a red herring, and we need some more objective testing. A regular speaker on the output of each of the amps is what I would be doing next - as I've probably said several times already. Failing that, a pair of headphones fed through a suitable series resistor would do

[Radio Wrangler]

Originally Posted by mhennessy

Anyway, the music choice could well be a red herring, and we need some more objective testing. A regular speaker on the output of each of the amps is what I would be doing next - as I've probably said several times already. Failing that, a pair of headphones fed through a suitable series resistor would do

I remember making both those suggestions some way back. Did anything get tried?

Full range phones or a full range speaker can be used to try any of the outputs of a triamp setup.

It would determine the cause between the amp/crossover assembly and the speaker and would work for all genres of music.

David

[mhennessy]

Agreed - you and me both, David. I'm not sure what else should be done at this stage that can actually move us forward.

I've just re-read this post:

Originally Posted by Linnovice

With just bass amp on the sound is clean.

With bass and treble there is a slight distortion.

With just the midrange on there is distinct distortion.

Same with both speakers (I’ve swapped them over).

...which seems to point to the midrange drivers or tweeters.

If nothing else, Linnovice could use the bass drivers to listen to the output of the 3 amps. They ought to go up to 4 or 5 kHz at least. IIRC, the bass and midrange drivers are the same on the Linn Kaber - at least, they look the same from the photos. If so, that means that he could connect the bass amp to the midrange driver to see what changes.

But it's worth saying that if the tweeters have rubbing voice coils (for example), you have to be careful when listening, because the acoustic energy from an adjacent driver could cause the tweeter to move and produce the distortion, even if the tweeter itself is not being driven. So get an ear right up to the drive units!

I first ran into that when I was a kid. I had been given a pair of old mid-fi speakers - they had the classic EMI elliptical drivers, but both of them had rubbing voice coils. No problem; I'll use them as speaker stands for my homemade speakers... As soon as I did, there was this terrible distortion. It took me rather longer than it should have done to "connect the dots" and realise that my good speakers were causing the EMIs to make all that noise! Valuable lesson learnt!

We know he has a pair of active speakers, but perhaps there are no more speakers available to test with? The concept of only having one or two pairs of speakers in the house is rather alien to me - though my wife would love it if I tried! At last count, there were 9 pairs in my music room alone. That's just silly

[jamesperrett]

Originally Posted by Hartley118

I recall encountering this kind of elusive high frequency distortion some years ago. It was absent on most music, but would appear on choral or 'bright' organ music. It turned out to be due to the voice coil rubbing on the magnet pole piece. This is easy to diagnose of course in a woofer or midrange speaker, but in a tweeter the rubbing noise components can easily be masked by much of the HF music content.

I had the same thing with one of my Tannoy dual concentrics. I had a feeling that something was wrong but found it very hard to isolate the problem. It was only when I fed low distortion sine waves to the speaker that I worked out that there was a problem above about 1.5kHz as the HF driver started to take over. It turned out to be very slight corrosion in the slot that the voice coil sits in - fortunately it is fairly easy to remove the diaphragm on the Tannoys and fix this.

Most music contained too much high frequency energy or sounds with natural harmonics to really isolate the problem. Even the sine wave generators on my mixing console and Roland synth had too much distortion to isolate the problem. It was only when I used a digital signal generator on the computer that I had a clean enough source to isolate the issue.