Linn Klimax DSM Organik, impressions

General HiFi discussion, using the Tune Method to evaluate performance

Moderator: Staff

User avatar
PetterS
Member
Member
Posts: 45
Joined: 2017-06-03 15:11

Re: Linn Klimax DSM Organik, impressions

Post by PetterS »

I'm reading this discussion with great interest. I have some thoughts upon what's on stake when LINN has gone full-blown digital. Especially since LINN has been in the forefront of the HIFI-industry in putting "music" first. If my long post and my slightly disparate thoughts would be better off in another thread, admin please feel free to move the post.

Being in the classical “business”, it is more than worrying that younger generations are mostly listening to music on devices that are not able to communicate some of the most basic parameters and qualities of recorded performances. Me myself have struggled since the early 90s to enjoy music outside the concert hall. Mostly my listening has related to work with the score in hand, and then often with recordings from the 50s and early sixties. Naxos Music Library and Spotify has quite a stock of old recordings which are quite enjoyable streamed in terms of music-making, but not so much sound-wise.

To me it seems that studio-produced electronic music works very well in digital chains and on digital devices. Music that is organized around a click track do not suffer when it’s transferred into ones and zeros. Everything is put in perfect order and the recordings are clean from any “disturbing” sounds.

Question 1 - Timing
In classical music – as well as folk music and jazz – the timing is always slightly off. The Tempo, agogics (tempo handling within rhythmical patterns, in and between phrases), dynamics, phrasing and gestures is always on the move in classical music. The same goes for jazz and folk music, but then it’s mostly referred to as timing or feeling. For those familiar with the Golden ratio, it seems that “good” music making is within the golden ratio. Acoustically in how instruments and concert halls are built, but also in harmonics, intonation, overtones, and rhythm handling. When Bernstein composed West Side Story, he struggled with notating “swing-feeling”, usually notated divided in three (2+1), but instead he wrote it divided in five (3+2). As you see, it’s only the early steps of the Fibonacci-series (golden-ratio). Probably a foot-tapping swing feeling would have to be written more like dividing it in 55 parts, 34+21 aiming for the unachievable golden ratio 1.68033….The same could be find in how appoggiaturas are handled in baroque music, the groove in folk music, dynamic and agogic tensions in classical music and so on.

So how does this mathematically unachievable ratio transfer to ones and zeros handled by a clock, even a super precise one? To me it seems that the rugged landscaped in a plastic piece engraved by sound waves, is still superior in communicating this. It seems like digitalization is putting a grid over the music. All those myriads of bits are moved ever so slightly to be in line with the raster instead of letting them be slightly off, which is the ideal in music making.

Question 2 – Sound making
A vital part of the soundscape in acoustical music, is the noise that sets the tone in motion. The noise from rosin and horsehair that sets the string in motion. The breath, lips and tongue that create a micro explosion before the tone is created in brass instruments. The movement of bodies in the symphony orchestra setting the air in motion together with the sound waves. And so on. In fact, in contemporary “classical” music there is a whole modernistic genre with compositions based NOT on the tones produced by the instruments, but instead from the “noises” before and after the tone production. An example is Helmut Lachenmann’s Schreiben (2003) for big symphony orchestra which is – when performed in a concert hall – a real treat, but not so much on CD.

So, what happens with “noise” that is an essential part of music’s acoustical landscape? Is it a risk that it is interpreted as jitter or unwanted noise when transferred to ones and zeros, by filters in the pressing, in the streaming services or by our digital devices?
Primare Np5 & NP30, Genki, Boazu, Ninka w. Poly. st., ICs: Linn Si & bl incl. dig, tt TD46,
matthias
Very active member
Very active member
Posts: 1238
Joined: 2007-12-25 16:47
Location: Germany

Re: Linn Klimax DSM Organik, impressions

Post by matthias »

PetterS wrote: 2021-05-09 13:24 It seems like digitalization is putting a grid over the music.
There might be a simple test. You take some of your pure analog LPs to your Linn dealer and play through an LP12 with Urika II/NGKDSM analog output. The analog signal goes through four digital "nasties" (ADC, digital RIAA, digital volume control and DAC). If there is a grid you will easily hear it. Teatime for example enjoyed the experience at Tonlaget.

Matt
LP12 / Exposure / JBL3677
matthias
Very active member
Very active member
Posts: 1238
Joined: 2007-12-25 16:47
Location: Germany

Re: Linn Klimax DSM Organik, impressions

Post by matthias »

PetterS wrote: 2021-05-09 13:24 So, what happens with “noise” that is an essential part of music’s acoustical landscape? Is it a risk that it is interpreted as jitter or unwanted noise when transferred to ones and zeros, by filters in the pressing, in the streaming services or by our digital devices?
IMO, a good ADC does not care at all whether you feed it with music or noise or twittering of birds etc..
However, in a digital system the resolution gets lower with decreasing volume levels. But this should not be a major problem with a good 24bit ADC.

Matt
LP12 / Exposure / JBL3677
User avatar
PetterS
Member
Member
Posts: 45
Joined: 2017-06-03 15:11

Re: Linn Klimax DSM Organik, impressions

Post by PetterS »

matthias wrote: 2021-05-09 15:09
PetterS wrote: 2021-05-09 13:24 It seems like digitalization is putting a grid over the music.
There might be a simple test. You take some of your pure analog LPs to your Linn dealer and play through an LP12 with Urika II/NGKDSM analog output. The analog signal goes through four digital "nasties" (ADC, digital RIAA, digital volume control and DAC). If there is a grid you will easily hear it. Teatime for example enjoyed the experience at Tonlaget.

Matt
Thanks, I see the point. But if I then tell the dealer something is lacking, and I suspect it is because of the digital "nasties" you pointed out, I would have to try the same LP12 setup on a purely analogue system. Isn't that one of the problems today when you visit a Linn dealer, when you would like to do a head to head comparison?
Primare Np5 & NP30, Genki, Boazu, Ninka w. Poly. st., ICs: Linn Si & bl incl. dig, tt TD46,
matthias
Very active member
Very active member
Posts: 1238
Joined: 2007-12-25 16:47
Location: Germany

Re: Linn Klimax DSM Organik, impressions

Post by matthias »

PetterS wrote: 2021-05-09 17:49 Thanks, I see the point. But if I then tell the dealer something is lacking, and I suspect it is because of the digital "nasties" you pointed out, I would have to try the same LP12 setup on a purely analogue system. Isn't that one of the problems today when you visit a Linn dealer, when you would like to do a head to head comparison?
I get your point as well.
I think there are quite a few dealers who offer both Linn and Lejonklou, so you should be able to compare the mentioned Linn set-up to LP12 > Entity > Sagatun Mono or LP12 > Urika I > Sagatun Mono.
BTW, for playing LPs only it makes absolutely no sense to go for the NGKDSM.

Edit: For a preamp with one input only I would strongly consider Giella Pi.

Matt
LP12 / Exposure / JBL3677
User avatar
lejonklou
Administrator
Administrator
Posts: 5244
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Linn Klimax DSM Organik, impressions

Post by lejonklou »

teatime wrote: 2021-05-09 00:36 Wow, I still don't get it. I understand that data being transferred is always clocked. This is fundamental to the transfer of digital information. But this is a different time domain from the audio processing. Do we need to worry about anything before the last time the data was "at rest"? (and if we do, how far back? and why?)

Surely, if a file is copied from one server to another, it is then at rest? In the same way, when a UDP packet is transferred from somewhere, via the network, and stored in a memory buffer in (for example) Källa, isn't it then "at rest"? It will stay in this memory buffer for an unspecified amount of time, until read (and evicted). So why is this not when time starts? The time the data reaches the buffer is unrelated to the time it is read from it. (It enters when the network manages to get it there, it leaves when it is needed. If it's not there in time, we just get data dropout (and probably silence) - these two time domains are not synced).

Once the data is being interpreted as audio data and transformed (up-sampling and all the other stuff that goes on) it makes perfect sense that time is absolutely critical. But before that? I don't understand.
I have no simple explanations to share and to my knowledge there are none (simple ones). There is only empirical evidence, patterns that emerge and can be practically verified, theories that form and seem to hold up until they are suddenly proven false, at least under certain conditions.

Our basic premise is that what we hear is the truth. And then we try to make sense of it and understand how it actually works. Not the other way around.

Have you compared ethernet cables? And the direction of them? That's one case where there's lots of empirical evidence, but very few patterns and no theory or real understanding. Luckily not all things are like that or development would be hopelessly difficult.
User avatar
lejonklou
Administrator
Administrator
Posts: 5244
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Linn Klimax DSM Organik, impressions

Post by lejonklou »

matss wrote: 2021-05-09 01:01
lejonklou wrote: 2021-05-08 23:45 You asked me where I think source first is in a digital replay chain. I replied. And your answer is a sloppy attempt at an insult?
No intention to insult. You must have misread my response, or I expressed myself in a sloppy way.
OK, then I must have misread. I apologize.
matss wrote: 2021-05-09 01:01I usually start off with a simple clear idea. Then I wish to improve on initial results. Throwing complexity at a problem usually yields confusing results if you’re not fully aware of what exactly you are doing and their consequences. I believe many of today's digital systems suffer from this. Things changes and you don’t really understand why. If you lack perseverance to really get to the bottom with the challenges you face you are usually better off stepping back to the initial simpler idea. That’s what I do when I get confused.

Finding the true root cause to a behaviour is usually a really cumbersome process, but when found can help you to create the simplest solution possible yielding the best result. I understand digital systems have a lot going on and it can all affect the delicate recreation of the analogue signal within. Trying to keep digital interference to a minimum seems like a good strategy to me.

How many digital conversions are you down to using airplay in Källa? And how does sourcing files from NAS instead of streaming from a control point add digital complexity to the system? I’m at loss here. Sorry for OT, final paragraph should maybe be moved to another section.
I find this a difficult one to answer, because complexity comes in many forms and it's not always a bad thing. For instance, I find that in some cases an extra physical step can, if optimised, actually improve performance. For instance the network switch many add in between the router and the streamer.

Likewise, it can sometimes be beneficial to convert the data more than what is actually necessary. For instance sending the music data in packets across the network and then unpacking them and creating a digital stream that you feed the DAC. This can actually be made to perform much better than to send a digital stream directly to the DAC (even if you buffer and reclock it locally).

We have made some really simple streamers, but they haven't been good enough. Källa has quite a bit of added complexity that isn't needed for it to work, but it just sounds better with them added.

There is no oversampling in the DAC part of Källa. So far we have found that using the original sample rate gives the most truthful and engaging reproduction. But that seems to apply only if the delivery of the data has been carefully optimised.
Defender
Very active member
Very active member
Posts: 833
Joined: 2018-02-14 22:35

Re: Linn Klimax DSM Organik, impressions

Post by Defender »

PetterS
I really enjoyed reading your comment. It seems that some try to reconstruct everything which was created in nature in an analogue way with endless variations now with just ones and zeros and hoping if we just accelerate the ones and zeros (resolution, kHz usw.)and have enough data that we can recreate everything (virtual reality, AI etc.) Most of the time those things created with ones and zeros look appear the same just without soul ... but most will not see or hear the difference.
matss
Active member
Active member
Posts: 107
Joined: 2011-03-19 16:52
Location: Stockholm

Re: Linn Klimax DSM Organik, impressions

Post by matss »

lejonklou wrote: 2021-05-09 20:43
matss wrote: 2021-05-09 01:01 How many digital conversions are you down to using airplay in Källa? And how does sourcing files from NAS instead of streaming from a control point add digital complexity to the system? I’m at loss here. Sorry for OT, final paragraph should maybe be moved to another section.
I find this a difficult one to answer, because complexity comes in many forms and it's not always a bad thing. For instance, I find that in some cases an extra physical step can, if optimised, actually improve performance. For instance the network switch many add in between the router and the streamer.

Likewise, it can sometimes be beneficial to convert the data more than what is actually necessary. For instance sending the music data in packets across the network and then unpacking them and creating a digital stream that you feed the DAC. This can actually be made to perform much better than to send a digital stream directly to the DAC (even if you buffer and reclock it locally).

We have made some really simple streamers, but they haven't been good enough. Källa has quite a bit of added complexity that isn't needed for it to work, but it just sounds better with them added.

There is no oversampling in the DAC part of Källa. So far we have found that using the original sample rate gives the most truthful and engaging reproduction. But that seems to apply only if the delivery of the data has been carefully optimised.
Interesting. Thanks for sharing. Yes, my experience as well. Sometimes complexity can be made to be a good thing, sometimes the simplest solution works best. It all depends. How do you manage a native hires file from say Qobuz through Källa to reach the most engaging result?
KLP12, KDS, KK, 3xKCT, homebrewn active speakers, Sneaky DS, OD11, Axis, Linto, Classik, Klout, OA12
teatime
Active member
Active member
Posts: 155
Joined: 2007-02-11 23:37

Re: Linn Klimax DSM Organik, impressions

Post by teatime »

lejonklou wrote: 2021-05-09 20:12
teatime wrote: 2021-05-09 00:36 Wow, I still don't get it. I understand that data being transferred is always clocked. This is fundamental to the transfer of digital information. But this is a different time domain from the audio processing. Do we need to worry about anything before the last time the data was "at rest"? (and if we do, how far back? and why?)

Surely, if a file is copied from one server to another, it is then at rest? In the same way, when a UDP packet is transferred from somewhere, via the network, and stored in a memory buffer in (for example) Källa, isn't it then "at rest"? It will stay in this memory buffer for an unspecified amount of time, until read (and evicted). So why is this not when time starts? The time the data reaches the buffer is unrelated to the time it is read from it. (It enters when the network manages to get it there, it leaves when it is needed. If it's not there in time, we just get data dropout (and probably silence) - these two time domains are not synced).

Once the data is being interpreted as audio data and transformed (up-sampling and all the other stuff that goes on) it makes perfect sense that time is absolutely critical. But before that? I don't understand.
I have no simple explanations to share and to my knowledge there are none (simple ones). There is only empirical evidence, patterns that emerge and can be practically verified, theories that form and seem to hold up until they are suddenly proven false, at least under certain conditions.

Our basic premise is that what we hear is the truth. And then we try to make sense of it and understand how it actually works. Not the other way around.

Have you compared ethernet cables? And the direction of them? That's one case where there's lots of empirical evidence, but very few patterns and no theory or real understanding. Luckily not all things are like that or development would be hopelessly difficult.
Sure, I've made comparisons that shouldn't matter but apparently does. I don't dispute that the field of digital audio clearly holds quite a few surprises.

But what I got me going was that you wrote:
lejonklou wrote: 2021-05-08 21:42 The source in a digital replay system is where the data starts moving.
This seems like a very strong statement. It sounds like you have a very specific place in mind, and I was curious to know what this place was.

If "where the data starts moving" isn't when it was last at rest/unclocked, then I don't understand how you can draw that conclusion. The data has stopped/started moving many times. (It reasonably happens a few times even inside Källa) How far do we go back? (To the singer's lips?)

If you only mean that the data conceptually starts moving at the NAS or streaming provider, and everything between there and here seems to matter, so lets call it "movement".. then, respectfully, "the source" seems rather arbitrary.

Since I doubt you meant the latter, I'm left confused about what place you were really referring to.
sunbeamgls
Very active member
Very active member
Posts: 924
Joined: 2012-04-04 15:19
Location: North Wales
Contact:

Re: Linn Klimax DSM Organik, impressions

Post by sunbeamgls »

PetterS wrote: 2021-05-09 13:24 ...

So, what happens with “noise” that is an essential part of music’s acoustical landscape? Is it a risk that it is interpreted as jitter or unwanted noise when transferred to ones and zeros, by filters in the pressing, in the streaming services or by our digital devices?
Edited the quote down for sake of saving space on the page.

I have a question about what is possibly (but might not be) a parallel issue in the analogue world. What if the tape machine playing back the original master for cutting the LP lacquer is running at a slightly different speed to the machine that created the tape? What if either of these tape machines have a tiny bit of wow and flutter - and it would be different on each machine thereby adding more errors. What happens to playing the record when the player is spinning at 33.35 rpm rather than 33.33 rpm? Or the record player as +/-0.25% wow and flutter? What then? Are these issues manifested in a similar way to the "grid" idea of digital? Or will they have different effects that are less or more disturbing for the listener?

BTW can we better understand if the clocking effect of digital capture (the "grid" as you phrase it) actually affects the timing of the music? It cuts the waves into slices then rebuilds them in the same order. 44,100 times a second (or more). With a piece of music at 120bpm as an example (2 beats per second, or 22,050 samples per beat) what level of inaccuracy would be recorded, never mind heard? And what would that inaccuracy be? A second of music is 1/44100 different? Very open to it being possible, but need to understand better what mechanism may be causing the perceived problem.

On the other hand I'm fully on board with how incredibly important the accuracy of the clock is in a digital playback system. With Katalyst and now with the KDSM/3 used only as a system hub it is very clear that a better clock = better music.
AEDSM; AEBox/1; 3x Tundra Stereo 2.5; PMC 20.26. Blogger. Exakt Design. SO measuring.
User avatar
PetterS
Member
Member
Posts: 45
Joined: 2017-06-03 15:11

Re: Linn Klimax DSM Organik, impressions

Post by PetterS »

sunbeamgls wrote: 2021-05-09 22:48
Edited the quote down for sake of saving space on the page.

I have a question about what is possibly (but might not be) a parallel issue in the analogue world. What if the tape machine playing back the original master for cutting the LP lacquer is running at a slightly different speed to the machine that created the tape? What if either of these tape machines have a tiny bit of wow and flutter - and it would be different on each machine thereby adding more errors. What happens to playing the record when the player is spinning at 33.35 rpm rather than 33.33 rpm? Or the record player as +/-0.25% wow and flutter? What then? Are these issues manifested in a similar way to the "grid" idea of digital? Or will they have different effects that are less or more disturbing for the listener?

BTW can we better understand if the clocking effect of digital capture (the "grid" as you phrase it) actually affects the timing of the music? It cuts the waves into slices then rebuilds them in the same order. 44,100 times a second (or more). With a piece of music at 120bpm as an example (2 beats per second, or 22,050 samples per beat) what level of inaccuracy would be recorded, never mind heard? And what would that inaccuracy be? A second of music is 1/44100 different? Very open to it being possible, but need to understand better what mechanism may be causing the perceived problem.

On the other hand I'm fully on board with how incredibly important the accuracy of the clock is in a digital playback system. With Katalyst and now with the KDSM/3 used only as a system hub it is very clear that a better clock = better music.
Human hearing is a mystery. I recommend this article. https://phys.org/news/2013-02-human-fou ... ciple.html
"For the first time, physicists have found that humans can discriminate a sound's frequency (related to a note's pitch) and timing (whether a note comes before or after another note) more than 10 times better than the limit imposed by the Fourier uncertainty principle." I'm not a mathematician and can't put the prinicple into fractions, but it seems that we're moving from the domain of sample rates into quantum physics in narrowing the sensitivity of human hearing.
Primare Np5 & NP30, Genki, Boazu, Ninka w. Poly. st., ICs: Linn Si & bl incl. dig, tt TD46,
User avatar
springwood64
Very active member
Very active member
Posts: 389
Joined: 2008-10-13 18:19
Location: UK

Re: Linn Klimax DSM Organik, impressions

Post by springwood64 »

PetterS wrote: 2021-05-09 23:18 Human hearing is a mystery. I recommend this article. https://phys.org/news/2013-02-human-fou ... ciple.html
"For the first time, physicists have found that humans can discriminate a sound's frequency (related to a note's pitch) and timing (whether a note comes before or after another note) more than 10 times better than the limit imposed by the Fourier uncertainty principle." I'm not a mathematician and can't put the prinicple into fractions, but it seems that we're moving from the domain of sample rates into quantum physics in narrowing the sensitivity of human hearing.
Great article and thanks for sharing. This paragraph won't surprise anyone here:
"We were still extremely surprised by how well our subjects did, and particularly surprised by the fact that the biggest gains appear to have been, by and large, in timing. You see, physicists tend to think hearing is spectrum. But spectrum is time-independent, and hearing is about rapid transients. We were just told, by the data, that our brains care a great deal about timing."
The best timing precision of their subjects could resolve down to 3 milliseconds.
Pete

Linn Axis, Kinki, Hakai, Boazu, Espeks
Spannko
Very active member
Very active member
Posts: 1430
Joined: 2008-01-24 21:46
Location: East of The Black Country, UK

Re: Linn Klimax DSM Organik, impressions

Post by Spannko »

Sound travels about 1m in 3ms, yet some manufacturers claim that time aligning drive units, which are usually less than 100mm out of alignment, improves the performance. However, the research seems to suggest that this would be inaudible, so are we being mislead, or is the science of hearing still not fully understood?
User avatar
PetterS
Member
Member
Posts: 45
Joined: 2017-06-03 15:11

Re: Linn Klimax DSM Organik, impressions

Post by PetterS »

sunbeamgls wrote: 2021-05-09 22:48 BTW can we better understand if the clocking effect of digital capture (the "grid" as you phrase it) actually affects the timing of the music? It cuts the waves into slices then rebuilds them in the same order. 44,100 times a second (or more). With a piece of music at 120bpm as an example (2 beats per second, or 22,050 samples per beat) what level of inaccuracy would be recorded, never mind heard? And what would that inaccuracy be? A second of music is 1/44100 different? Very open to it being possible, but need to understand better what mechanism may be causing the perceived problem.

On the other hand I'm fully on board with how incredibly important the accuracy of the clock is in a digital playback system. With Katalyst and now with the KDSM/3 used only as a system hub it is very clear that a better clock = better music.
sunbeamgls, valid points, and the maths seems to indicate there should not be any problems whatsoever. But we have heard the same numerous times about "bit perfect". I'll try to exemplify what I mean with the grid in timing, and how it seems that blocks of info (slices) are moved ever so slightly. Here is an example of Duke Ellington performing with his quartet on two different media. The first clip is a Qobuz stream and the second clip is vinyl on my Thorens tt. It's a gigantic difference when heard in the room. The "inaccuracies" in the performance has a charming "feeling" and "groove" on vinyl, which sounds just right and it is also very captivating. When streamed it sounds unprecise and chaotic. I'm well aware that recording this on my mobile phone, uploading it to Dropbox, and then played on whatever listening device you are using, involves a lot of "digital nasties".
Despite all that, I think there is an audible difference in timing:
Qobuz:
https://www.dropbox.com/s/pefel8ll2e8io ... s.mp4?dl=0
Vinyl:
https://www.dropbox.com/s/g1znm6nn376io ... s.mp4?dl=0
Primare Np5 & NP30, Genki, Boazu, Ninka w. Poly. st., ICs: Linn Si & bl incl. dig, tt TD46,
sunbeamgls
Very active member
Very active member
Posts: 924
Joined: 2012-04-04 15:19
Location: North Wales
Contact:

Re: Linn Klimax DSM Organik, impressions

Post by sunbeamgls »

PetterS wrote: 2021-05-10 15:51
sunbeamgls wrote: 2021-05-09 22:48 BTW can we better understand if the clocking effect of digital capture (the "grid" as you phrase it) actually affects the timing of the music? It cuts the waves into slices then rebuilds them in the same order. 44,100 times a second (or more). With a piece of music at 120bpm as an example (2 beats per second, or 22,050 samples per beat) what level of inaccuracy would be recorded, never mind heard? And what would that inaccuracy be? A second of music is 1/44100 different? Very open to it being possible, but need to understand better what mechanism may be causing the perceived problem.

On the other hand I'm fully on board with how incredibly important the accuracy of the clock is in a digital playback system. With Katalyst and now with the KDSM/3 used only as a system hub it is very clear that a better clock = better music.
sunbeamgls, valid points, and the maths seems to indicate there should not be any problems whatsoever. But we have heard the same numerous times about "bit perfect". I'll try to exemplify what I mean with the grid in timing, and how it seems that blocks of info (slices) are moved ever so slightly. Here is an example of Duke Ellington performing with his quartet on two different media. The first clip is a Qobuz stream and the second clip is vinyl on my Thorens tt. It's a gigantic difference when heard in the room. The "inaccuracies" in the performance has a charming "feeling" and "groove" on vinyl, which sounds just right and it is also very captivating. When streamed it sounds unprecise and chaotic. I'm well aware that recording this on my mobile phone, uploading it to Dropbox, and then played on whatever listening device you are using, involves a lot of "digital nasties".
Despite all that, I think there is an audible difference in timing:
Qobuz:
https://www.dropbox.com/s/pefel8ll2e8io ... s.mp4?dl=0
Vinyl:
https://www.dropbox.com/s/g1znm6nn376io ... s.mp4?dl=0
Thanks PetterS

Do you find this issue regardless of which source you're using for digital (Qobuz, Tidal, local rips, etc.) and the different connectivity options you have between streamer and DAC?
AEDSM; AEBox/1; 3x Tundra Stereo 2.5; PMC 20.26. Blogger. Exakt Design. SO measuring.
Eli7
Active Member
Active Member
Posts: 90
Joined: 2020-05-24 18:22

Re: Linn Klimax DSM Organik, impressions

Post by Eli7 »

PetterS wrote: 2021-05-10 15:51
sunbeamgls wrote: 2021-05-09 22:48 BTW can we better understand if the clocking effect of digital capture (the "grid" as you phrase it) actually affects the timing of the music? It cuts the waves into slices then rebuilds them in the same order. 44,100 times a second (or more). With a piece of music at 120bpm as an example (2 beats per second, or 22,050 samples per beat) what level of inaccuracy would be recorded, never mind heard? And what would that inaccuracy be? A second of music is 1/44100 different? Very open to it being possible, but need to understand better what mechanism may be causing the perceived problem.

On the other hand I'm fully on board with how incredibly important the accuracy of the clock is in a digital playback system. With Katalyst and now with the KDSM/3 used only as a system hub it is very clear that a better clock = better music.
sunbeamgls, valid points, and the maths seems to indicate there should not be any problems whatsoever. But we have heard the same numerous times about "bit perfect". I'll try to exemplify what I mean with the grid in timing, and how it seems that blocks of info (slices) are moved ever so slightly. Here is an example of Duke Ellington performing with his quartet on two different media. The first clip is a Qobuz stream and the second clip is vinyl on my Thorens tt. It's a gigantic difference when heard in the room. The "inaccuracies" in the performance has a charming "feeling" and "groove" on vinyl, which sounds just right and it is also very captivating. When streamed it sounds unprecise and chaotic. I'm well aware that recording this on my mobile phone, uploading it to Dropbox, and then played on whatever listening device you are using, involves a lot of "digital nasties".
Despite all that, I think there is an audible difference in timing:
Qobuz:
https://www.dropbox.com/s/pefel8ll2e8io ... s.mp4?dl=0
Vinyl:
https://www.dropbox.com/s/g1znm6nn376io ... s.mp4?dl=0
I can hear what you mean. The Thorens is really much, much better. A very big difference.
MDSM3 / Netgear GS 108T-200 / BJC Cat 6a / K200 (Lejonklou Cut K400) /
M140
User avatar
Tony Tune-age
Very active member
Very active member
Posts: 1224
Joined: 2009-12-19 19:07
Location: United States

Re: Linn Klimax DSM Organik, impressions

Post by Tony Tune-age »

Eli7 wrote: 2021-05-10 17:01
PetterS wrote: 2021-05-10 15:51 Despite all that, I think there is an audible difference in timing:
Qobuz:
https://www.dropbox.com/s/pefel8ll2e8io ... s.mp4?dl=0
Vinyl:
https://www.dropbox.com/s/g1znm6nn376io ... s.mp4?dl=0
I can hear what you mean. The Thorens is really much, much better. A very big difference.
I like how the Thorens sounds as well, more musical and more enjoyable too!

Cheers
Tony Tune-age
User avatar
PetterS
Member
Member
Posts: 45
Joined: 2017-06-03 15:11

Re: Linn Klimax DSM Organik, impressions

Post by PetterS »

sunbeamgls wrote: 2021-05-10 16:44 Thanks PetterS

Do you find this issue regardless of which source you're using for digital (Qobuz, Tidal, local rips, etc.) and the different connectivity options you have between streamer and DAC?
No. I'm actually very happy with the digital setup now, when it comes to digitally produced music and music that is "beat-based". Streaming is way better than my Genki. My Genki is better through my DAC but still can't compete with streaming.
Here is an example with Esbjorn Svensson Trio on Spotify:
https://www.dropbox.com/s/u33aky5o8sndq ... y.mp4?dl=0

But the magic in classical recordings and old jazz is not there, either on the Genki or when streaming. I would actually love to do the comparison with the Ellington recording and maybe some good old classical recordings - like the Karajan Beethoven cycle from 1963 - on a first class streaming setup, but my suspicion is that something is lost in transferring those old recordings to streaming.

Edit: Btw, the video clips with Organik (Korea?) that was linked in another thread were impressively clear, but for me the classical (Mutter and Yo Yo Ma) sounded square and without the agogic magic.
Last edited by PetterS on 2021-05-10 18:22, edited 1 time in total.
Primare Np5 & NP30, Genki, Boazu, Ninka w. Poly. st., ICs: Linn Si & bl incl. dig, tt TD46,
Lego
Very active member
Very active member
Posts: 906
Joined: 2007-04-18 11:42
Location: glasgow

Re: Linn Klimax DSM Organik, impressions

Post by Lego »

teatime wrote: 2021-05-09 21:53
lejonklou wrote: 2021-05-09 20:12
teatime wrote: 2021-05-09 00:36 Wow, I still don't get it. I understand that data being transferred is always clocked. This is fundamental to the transfer of digital information. But this is a different time domain from the audio processing. Do we need to worry about anything before the last time the data was "at rest"? (and if we do, how far back? and why?)

Surely, if a file is copied from one server to another, it is then at rest? In the same way, when a UDP packet is transferred from somewhere, via the network, and stored in a memory buffer in (for example) Källa, isn't it then "at rest"? It will stay in this memory buffer for an unspecified amount of time, until read (and evicted). So why is this not when time starts? The time the data reaches the buffer is unrelated to the time it is read from it. (It enters when the network manages to get it there, it leaves when it is needed. If it's not there in time, we just get data dropout (and probably silence) - these two time domains are not synced).

Once the data is being interpreted as audio data and transformed (up-sampling and all the other stuff that goes on) it makes perfect sense that time is absolutely critical. But before that? I don't understand.
I have no simple explanations to share and to my knowledge there are none (simple ones). There is only empirical evidence, patterns that emerge and can be practically verified, theories that form and seem to hold up until they are suddenly proven false, at least under certain conditions.

Our basic premise is that what we hear is the truth. And then we try to make sense of it and understand how it actually works. Not the other way around.

Have you compared ethernet cables? And the direction of them? That's one case where there's lots of empirical evidence, but very few patterns and no theory or real understanding. Luckily not all things are like that or development would be hopelessly difficult.
Sure, I've made comparisons that shouldn't matter but apparently does. I don't dispute that the field of digital audio clearly holds quite a few surprises.

But what I got me going was that you wrote:
lejonklou wrote: 2021-05-08 21:42 The source in a digital replay system is where the data starts moving.
This seems like a very strong statement. It sounds like you have a very specific place in mind, and I was curious to know what this place was.

If "where the data starts moving" isn't when it was last at rest/unclocked, then I don't understand how you can draw that conclusion. The data has stopped/started moving many times. (It reasonably happens a few times even inside Källa) How far do we go back? (To the singer's lips?)

If you only mean that the data conceptually starts moving at the NAS or streaming provider, and everything between there and here seems to matter, so lets call it "movement".. then, respectfully, "the source" seems rather arbitrary.

Since I doubt you meant the latter, I'm left confused about what place you were really referring to.
I think we're back to what Linnofil originally said.The item closest to the music is the most important part of the chain .At the time that was construed as the Nas/network. Fredrik is now saying Spotify is the better source.Seems simple enough.
Although it's not really a great source when your Internet goes down .
I know that tune
matthias
Very active member
Very active member
Posts: 1238
Joined: 2007-12-25 16:47
Location: Germany

Re: Linn Klimax DSM Organik, impressions

Post by matthias »

Lego wrote: 2021-05-10 17:50 Although it's not really a great source when your Internet goes down .
Do you have issues with your internet?

Matt
LP12 / Exposure / JBL3677
Lego
Very active member
Very active member
Posts: 906
Joined: 2007-04-18 11:42
Location: glasgow

Re: Linn Klimax DSM Organik, impressions

Post by Lego »

matthias wrote: 2021-05-10 18:09
Lego wrote: 2021-05-10 17:50 Although it's not really a great source when your Internet goes down .
Do you have issues with your internet?

Matt
Not sure Matthias,don't use it enough to be 100% certain, generally it's been fine.
I know that tune
User avatar
lejonklou
Administrator
Administrator
Posts: 5244
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Linn Klimax DSM Organik, impressions

Post by lejonklou »

matss wrote: 2021-05-09 21:52How do you manage a native hires file from say Qobuz through Källa to reach the most engaging result?
The hardware in Källa can handle 24 bits, but Protokoll is limited to and optimised for 16 bits. So a 24 bit file will be converted to 16 when sent through Protokoll.

The reason behind our focus on 16 bit is that we have in general not been impressed with higher resolution audio. There are some recordings that are better in 24 bits, but the majority sadly isn't. They seem to gain some fine detail and loose some core qualities like bass timing. It's partly unclear to us why this is the case, but the problem lies within the files themselves, not how you convert them.
matthias
Very active member
Very active member
Posts: 1238
Joined: 2007-12-25 16:47
Location: Germany

Re: Linn Klimax DSM Organik, impressions

Post by matthias »

lejonklou wrote: 2021-05-10 19:30 So a 24 bit file will be converted to 16 when sent through Protokoll.
I suppose it is the same with sampling frequency?
192kHz to 48kHz or 44,1kHz?

Matt
LP12 / Exposure / JBL3677
User avatar
lejonklou
Administrator
Administrator
Posts: 5244
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Re: Linn Klimax DSM Organik, impressions

Post by lejonklou »

teatime wrote: 2021-05-09 21:53
lejonklou wrote: 2021-05-08 21:42 The source in a digital replay system is where the data starts moving.
This seems like a very strong statement. It sounds like you have a very specific place in mind, and I was curious to know what this place was.

If "where the data starts moving" isn't when it was last at rest/unclocked, then I don't understand how you can draw that conclusion. The data has stopped/started moving many times. (It reasonably happens a few times even inside Källa) How far do we go back? (To the singer's lips?)

If you only mean that the data conceptually starts moving at the NAS or streaming provider, and everything between there and here seems to matter, so lets call it "movement".. then, respectfully, "the source" seems rather arbitrary.

Since I doubt you meant the latter, I'm left confused about what place you were really referring to.
The data starts and stops moving many times, that's true. But the stops where it's actually resting within a powered and clocked system doesn't behave like static data. For instance, there's a small buffer of 2K in Källa and it's very sensitive to everything in its immediate environment. The data that briefly pauses there behaves like it's moving.

On the other end of the scale, you can store the data on a USB stick or on a HDD/SSD. There it retains its intrinsic qualities regardless of what happens around it. But as soon as you start moving it, you add a sonic signature. The character of that signature will depend on exactly how the movement is executed. And traces of that signature will remain even after many processes, conversions, buffers and the final conversion to analogue.
Post Reply