Playground for practical listening exercises

We use the Tune Method to evaluate performance

Moderator: Staff

User avatar
Matteo
Very active member
Very active member
Posts: 913
Joined: 2018-01-25 14:12
Location: Milano, Italia

Re: Playground for practical listening exercises

Post by Matteo »

998 way way more musical for me
Spannko
Very active member
Very active member
Posts: 2292
Joined: 2008-01-24 21:46
Location: North East of The Black Country, UK

Re: Playground for practical listening exercises

Post by Spannko »

I’m very wary of a difference which appears clearer, tighter, more revealing etc, and 998 falls into this category for me. Consequently I’ve taken great care to gather and double check my thoughts before commenting. I don’t know the track, so I could be wrong, but on first listening to 998, apart from the sound/hifi differences, my first thought was “this sounds like a live concert!”, it appears to have a live vibe. This wasn’t apparent in 997. For me, this sort of impressive sound often doesn’t come cheap. There’s usually a price to pay! This comparison feels quite an important one, which is why I wanted to delve deeper and try and sort out the quite small musical differences. An organ is playing in the background, buried in the mix. It’s audible on 997, and although the clarity and refinement of 998 makes it slightly easier to hear more of the notes being played, what’s important for me is that not only is the organ more audible on 998, it appears to be more tuneful too. This improvement in musicality makes me more confident that 998 would not only initially sound more impressive, it should also prove to be more enjoyable in the long term. It’s a win-win!

(I’ve gone into depth so that readers new to the Tune Method of evaluation can develop a better understanding of the way I listen so that they can try the method for themselves).
Last edited by Spannko on 2022-11-13 09:21, edited 1 time in total.
matthias
Very active member
Very active member
Posts: 2092
Joined: 2007-12-25 16:47
Location: Germany

Re: Playground for practical listening exercises

Post by matthias »

lejonklou wrote: 2022-11-11 22:08 I'll tell you a story that will make your balls shrink to the size of raisins. (quoting Spike in Notting Hill)
Fredrik,
please tell us the story...
Matt

MBP / Exposure pre + power (both modified) / JBL3677
User avatar
ThomasOK
Very active member
Very active member
Posts: 4358
Joined: 2007-02-02 18:41
Location: United States
Contact:

Re: Playground for practical listening exercises

Post by ThomasOK »

matthias wrote: 2022-11-12 22:22
lejonklou wrote: 2022-11-11 22:08 I'll tell you a story that will make your balls shrink to the size of raisins. (quoting Spike in Notting Hill)
Fredrik,
please tell us the story...
You sure you want that?

Let me find some tinfoil first. :-D
The LP12 Whisperer
Manufacturer, Distributor, Retailer and above all lover of music.
tokenbrit
Very active member
Very active member
Posts: 2038
Joined: 2012-03-22 19:47
Location: New England

Re: Playground for practical listening exercises

Post by tokenbrit »

ThomasOK wrote: 2022-11-12 22:44
matthias wrote: 2022-11-12 22:22
lejonklou wrote: 2022-11-11 22:08 I'll tell you a story that will make your balls shrink to the size of raisins. (quoting Spike in Notting Hill)
Fredrik,
please tell us the story...
You sure you want that?

Let me find some tinfoil first. :-D
Is it a story about 3 digit number fields: get it right by 999 or start over 😱
User avatar
lejonklou
Administrator
Administrator
Posts: 6524
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Playground for practical listening exercises

Post by lejonklou »

Thank you for your comments!

Will tell you what we listened to tomorrow evening. And don’t worry about raisins, it was probably only mine that were on the line.
Arjen
Very active member
Very active member
Posts: 431
Joined: 2021-06-10 13:23
Location: NL and Friesland

Re: Playground for practical listening exercises

Post by Arjen »

To my ears the 998 is more clear, more pronounced. 997 is slightly covered, tends to dullness. But longer listening 997 might be less fatiguing.
Lenco 76/S TJN, Dr. Fuss, Supernait2, CD5X, Slipsik7.1, Millon Phantom, Soundeck, AudioSensibility
User avatar
lejonklou
Administrator
Administrator
Posts: 6524
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Playground for practical listening exercises

Post by lejonklou »

Ok, so what we were listening to was Bobby Womack on Spotify, played by Källa, Giella ∏ 1.2, and JBL LSR308 on Linn Katan stands, in our living room.

Giella ∏ 1.2 is an experimental version that I intend to send to Headonist in Paris for evaluation before it's released, as Pierre at Headonist is such a headphone aficionado and responsible for the development of version ∏ in the first place. I want him to test it with all the hard-to-drive headphones he's got.

Version 1.1 is new chassis ground parts, identical with those on Slipsik 8 and Entity 1.2.
Version 1.2 is the addition of a new resistor, joining two ground points.

In the original version, there was no resistor at this position. It's perfectly possible to have zero Ohms - a short circuit - but in the listening tests I preferred no resistor (infinite resistance). The difference was really small between the two and I probably went back and forth at least twenty times before making my mind up. My conclusion was this was a rather unimportant parameter and no resistor was needed.

One and a half years later, I revisited this detail, as I had a vague feeling that maybe I hadn't drilled deep enough into that seemingly unimportant parameter. So instead of zero and infinite, I thought I'd be a little more precise. I tried 10 Ohm versus 1 MegaOhm. The differences were very similar to what I remembered from last time and I very slightly preferred 1 Megaohm.
I tried 100 kiloOhms versus 1 MegaOhm, preferred 100 kiloOhm.
I tried 10 kiloOhms versus 100 kiloOhms, preferred 10 kiloOhm.
I tried 1 kiloOhms versus 10 kiloOhms, preferred 1 kiloOhm.

An odd thing I noticed was that the differences appeared to always be the same: The higher number sounded a little sharper, the lower a little looser. And they were not really getting any bigger. They felt small and close to unimportant. But I now had a slight preference for a much lower resistance than infinite; 1000 Ohms.

At this point I asked a friend to help me out. I gave him a unit with a switch on and said "Please tell me whether you prefer the switch up or down". He didn't know what he was listening to. I compared myself and then he reported which he preferred. Our findings turned out to match 100%, which was inspiring. In this way we covered 500 Ohms to 5000 Ohms and eventually ended up where we began, at 1000. 976 Ohms appeared a tiny bit sloppy and 1020 Ohms was a tiny bit sharp, 1000 seemed to strike a balanced middle ground.

Still the differences were small and difficult. When I tried capturing them on clips, they were not easy to tell apart.

My intern Oscar has selected around 4000 resistors in the value of 1k (1000 Ohms) into 0.1% groups. They range from 993 to 1005 Ohms. The measuring equipment goes to 0.01%, but using that setting takes more time and so far I haven't found any need for that kind of precision in the particular value of 1k.

Now, just out of curiosity I took out the boxes of 0.1% selected 1k and rigged 995, 996, 997, 998, 999 and 1000 Ohms to begin with.

BOOM! Here was the peak! Or rather: The level shift.
995 sounded almost identical to 996.
At 997, maybe there was something starting to tighten up. Maybe.
At 998, things just tilted from loose to sharp.
999 and 1000 sound very close to 998, but a tiny bit sharper still. And this sharpness sort of pulls the sounds apart in time, making things seem more distinct but also stiffening up the rhythm. The flow of the song is sort of hardened and sounds appear less connected with one another.

So what made me quote Spike in 'Notting Hill' on Friday evening was that I find the difference between 997 and 998 Ohms bigger than anything outside of that range. Zero to 997 is a series of difficult choices, 998 to infinite is equally difficult, but there's a level shift right there in that 0.1% range where things flip between the two. Not the typical single peak of most fine tunings (which in comparison are much easier to determine, because the musical gradually improves until it suddenly gets worse and then you know that you've passed the peak) or the multiple peaks sometimes found when precision torque tightening mechanical fasteners (which are more work to optimize than the single peaks because you need to find them all and then compare the peaks).

I will now rig 0.01% for this resistor and am fairly certain the optimal value is somewhere in between 997.0 and 998.0 Ohms. And that value will go into Giella ∏ 1.2.

EDIT: I might add that I don't think it's possible to measure these differences in the audio signal coming out of Giella ∏. Neither the difference between zero and infinite nor between 997 and 998. The difference is likely buried far below the noise level on a very low frequency.
User avatar
markiteight
Moderator
Moderator
Posts: 834
Joined: 2012-01-13 01:50
Location: Seattle, Wa. USA

Re: Playground for practical listening exercises

Post by markiteight »

That is fascinating on so many levels. Thanks for sharing with us, Fredrik!

I'm curious to know if this discovery can be applied to your other products (possibly as part of an incremental '.1' upgrade, or if it's specific to the products mentioned. The difference seems subtle but important.
Spannko
Very active member
Very active member
Posts: 2292
Joined: 2008-01-24 21:46
Location: North East of The Black Country, UK

Re: Playground for practical listening exercises

Post by Spannko »

This is almost too much for my raisin brain to comprehend! Have you been able to repeat the experiment with different examples of Giella and resisters? (I’m sure you have, but I thought I’d ask anyway).
matthias
Very active member
Very active member
Posts: 2092
Joined: 2007-12-25 16:47
Location: Germany

Re: Playground for practical listening exercises

Post by matthias »

lejonklou wrote: 2022-11-14 00:50 Version 1.1 is new chassis ground parts, identical with those on Slipsik 8 and Entity 1.2.
Version 1.2 is the addition of a new resistor, joining two ground points.
Sorry Fredrik,
I thought for safety reasons the resistance on chassis ground should be as low as possible.
Please, can you shed some light on this?
Thx
Matt

MBP / Exposure pre + power (both modified) / JBL3677
Arjen
Very active member
Very active member
Posts: 431
Joined: 2021-06-10 13:23
Location: NL and Friesland

Re: Playground for practical listening exercises

Post by Arjen »

Fredrik, I understand the methodology. But not the micro-technical details. Must be a precious work. Indeed an interested question in what way other devices like phono preamps will differ/improve using the same selecting and searching method.
Last edited by Arjen on 2022-11-14 19:20, edited 1 time in total.
Lenco 76/S TJN, Dr. Fuss, Supernait2, CD5X, Slipsik7.1, Millon Phantom, Soundeck, AudioSensibility
User avatar
lejonklou
Administrator
Administrator
Posts: 6524
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Playground for practical listening exercises

Post by lejonklou »

markiteight wrote: 2022-11-14 02:14 I'm curious to know if this discovery can be applied to your other products (possibly as part of an incremental '.1' upgrade, or if it's specific to the products mentioned. The difference seems subtle but important.
Giella and Superkikkin are likely the only two where it's applicable.
Spannko wrote: 2022-11-14 07:29 This is almost too much for my raisin brain to comprehend! Have you been able to repeat the experiment with different examples of Giella and resisters? (I’m sure you have, but I thought I’d ask anyway).
Confirmed once, but when I do the next I will be more certain it isn't a freak occurrence.
matthias wrote: 2022-11-14 08:53 I thought for safety reasons the resistance on chassis ground should be as low as possible.
Correct, mains ground to chassis is bolted down with minimal resistance.
That's not the position of this resistor. The changes in version 1.1 and 1.2 are not related to one another.
Arjen wrote: 2022-11-14 12:26 Fredrik, I understand the methodology. But not the technical details. Must be a precious work. Indeed an interested question in what way other devices like phono preamps will differ/improve using the same selecting and searching method.
Attention to detail is key when making machines that can bring the music to life.

I fine tune every aspect and detail to the point where either the differences become insignificant, or it's not practically possible to go any further. Many small details optimized create a big difference.

How this resistance "behaved", however, was something I haven't come across before. I hope it never happens again. Too weird and time consuming to optimize.
matthias
Very active member
Very active member
Posts: 2092
Joined: 2007-12-25 16:47
Location: Germany

Re: Playground for practical listening exercises

Post by matthias »

lejonklou wrote: 2022-11-14 12:58 Superkikkin
Cool name :-D
Matt

MBP / Exposure pre + power (both modified) / JBL3677
User avatar
lejonklou
Administrator
Administrator
Posts: 6524
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Playground for practical listening exercises

Post by lejonklou »

Tendaberry wrote: 2022-11-10 19:00 I just discovered that the size of the C-file is 3 times bigger... Haven't changed any settings since recording A though, weird.
As Charlie wrote this is likely due to the brighter TV.

What was were we listening to in A versus C?
tokenbrit
Very active member
Very active member
Posts: 2038
Joined: 2012-03-22 19:47
Location: New England

Re: Playground for practical listening exercises

Post by tokenbrit »

markiteight wrote: 2022-11-14 02:14 That is fascinating on so many levels. Thanks for sharing with us, Fredrik!
Indeed. Fascinating that this is a 1ohm change (& that may not be precise enough); that almost all can hear a difference; that almost all share the same preference; that our descriptions are so different :)

Will be fascinating to hear the "level shift" clips between 997 & 998, if/when you share those. Thanks from me too!
tpetsch
Very active member
Very active member
Posts: 342
Joined: 2020-08-17 18:46
Location: United States

Re: Playground for practical listening exercises

Post by tpetsch »

Very interesting Frederik, but in reality is it realistic to think that a capacitors tolerance considering time, temperature & usage can retain such a tight 1 ohm sweet spot for any reasonable duration? And then what happens when -over some relatively short period of time, say 6 months? ...if that?- the left channel shifts out of tolerance while the right is still retaining that 1 ohm tolerance for example?
Rega P10, Aphelion II, Aura. Naim CB 32.5/HC, Naxo 3-6/HC, 3x250 into K20/DMS.
User avatar
lejonklou
Administrator
Administrator
Posts: 6524
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Playground for practical listening exercises

Post by lejonklou »

tpetsch wrote: 2022-11-14 19:11 Very interesting Frederik, but in reality is it realistic to think that a capacitors tolerance considering time, temperature & usage can retain such a tight 1 ohm sweet spot for any reasonable duration? And then what happens when -over some relatively short period of time, say 6 months? ...if that?- the left channel shifts out of tolerance while the right is still retaining that 1 ohm tolerance for example?
Resistor, not capacitor. And super stable over time and temperature. In fact I have tried altering the resistance by baking them in an oven for days and they didn't change at all. Also didn't change the slightest after 10 years of continuous running in a powered up circuit.

Therefore I think it can be worth the effort to optimize this particular value to within 0.01%.
tpetsch
Very active member
Very active member
Posts: 342
Joined: 2020-08-17 18:46
Location: United States

Re: Playground for practical listening exercises

Post by tpetsch »

lejonklou wrote: 2022-11-14 21:21
tpetsch wrote: 2022-11-14 19:11 Very interesting Frederik, but in reality is it realistic to think that a capacitors tolerance considering time, temperature & usage can retain such a tight 1 ohm sweet spot for any reasonable duration? And then what happens when -over some relatively short period of time, say 6 months? ...if that?- the left channel shifts out of tolerance while the right is still retaining that 1 ohm tolerance for example?
Resistor, not capacitor. And super stable over time and temperature. In fact I have tried altering the resistance by baking them in an oven for days and they didn't change at all. Also didn't change the slightest after 10 years of continuous running in a powered up circuit.

Therefore I think it can be worth the effort to optimize this particular value to within 0.01%.
Oh resistors, don't know why I was thinking capacitors. But that's fantastic that these resistors are so stable even after baking them, really can lead to some interesting and long term results, keep up the good work! ...A few friends of mine that have a music studio sometimes play with resistors on various gear AC cable grounds to get different effects of "air" around the music, at the extreme end some resistors make the band sound like their playing in a cave, but like you said, seems you've really hit an interesting and musical "sweet spot" for playback.
Rega P10, Aphelion II, Aura. Naim CB 32.5/HC, Naxo 3-6/HC, 3x250 into K20/DMS.
User avatar
ThomasOK
Very active member
Very active member
Posts: 4358
Joined: 2007-02-02 18:41
Location: United States
Contact:

Re: Playground for practical listening exercises

Post by ThomasOK »

Very interesting information, Fredrik! It does seem rather unexpected that there is little difference between slightly off and a lot off in either direction but a definite difference between the two resistances where it flips. That difference was clearly audible so I can imagine the improvement that will happen when it is just right. So maybe 997.7 ohms? The work you do really is amazing. Thanks for the effort and the results it brings us.
The LP12 Whisperer
Manufacturer, Distributor, Retailer and above all lover of music.
User avatar
lejonklou
Administrator
Administrator
Posts: 6524
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Playground for practical listening exercises

Post by lejonklou »

tpetsch
Very active member
Very active member
Posts: 342
Joined: 2020-08-17 18:46
Location: United States

Re: Playground for practical listening exercises

Post by tpetsch »

lejonklou wrote: 2022-11-17 00:24 Ok, let’s see what you think of these two:

Lower: https://www.dropbox.com/s/zuuyg8h3uu5tt ... r.mov?dl=0

Higher:
https://www.dropbox.com/s/awrmf6bcmfbwc ... r.mov?dl=0
"Higher" seems to have the edge IMO.
Rega P10, Aphelion II, Aura. Naim CB 32.5/HC, Naxo 3-6/HC, 3x250 into K20/DMS.
Arjen
Very active member
Very active member
Posts: 431
Joined: 2021-06-10 13:23
Location: NL and Friesland

Re: Playground for practical listening exercises

Post by Arjen »

lejonklou wrote: 2022-11-17 00:24 Ok, let’s see what you think of these two:

Lower: https://www.dropbox.com/s/zuuyg8h3uu5tt ... r.mov?dl=0

Higher:
https://www.dropbox.com/s/awrmf6bcmfbwc ... r.mov?dl=0
Lower more clear and defined, Higher more smooth, probably more body, but less defined IMO
Lenco 76/S TJN, Dr. Fuss, Supernait2, CD5X, Slipsik7.1, Millon Phantom, Soundeck, AudioSensibility
User avatar
markiteight
Moderator
Moderator
Posts: 834
Joined: 2012-01-13 01:50
Location: Seattle, Wa. USA

Re: Playground for practical listening exercises

Post by markiteight »

lejonklou wrote: 2022-11-17 00:24 Ok, let’s see what you think of these two:

Lower: https://www.dropbox.com/s/zuuyg8h3uu5tt ... r.mov?dl=0

Higher:
https://www.dropbox.com/s/awrmf6bcmfbwc ... r.mov?dl=0
There's aspects of both I like but I lean toward Higher. P'raps something in between?
tokenbrit
Very active member
Very active member
Posts: 2038
Joined: 2012-03-22 19:47
Location: New England

Re: Playground for practical listening exercises

Post by tokenbrit »

Leaning towards Higher: Lower is rhythmic, but Higher feels more musical...
Post Reply