• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Review and Measurements of Cambridge DacMagic Plus

tired_guru

Active Member
Joined
Feb 8, 2019
Messages
100
Likes
79
#41
especially noise floor is noticeable and less bass bloom more impact
This is what atom does with very low impedance (13ohms here) headphones with great bass possibilities. It simply has enormous control over the drivers thanks to its very low output impedance and lots of power for low impedance HPs without a hint of distortion. It is a must for high perfoming DACs when you use HPs (especially when taking into account the price of atom).

A few days ago I had possibility to check out fostex th-610 (25 ohms, very strong tesla magnets and famous biocellulose drivers know from its epic, organic bass). These HPs when driving by some high impedance output amps thus with poor damping factor, can sound dull, fluffy, muddy, boring. Once you connect them to something like atom this is completely different beast. Amazing impulse response, control of transients, clarity & impact.
 
Joined
Feb 23, 2019
Messages
7
Likes
1
#42
Any chance of any performance results with toslink? I always prefer optical connections - no ground loops etc...
 
Joined
Nov 23, 2018
Messages
32
Likes
30
#43
This is what atom does with very low impedance (13ohms here) headphones with great bass possibilities. It simply has enormous control over the drivers thanks to its very low output impedance and lots of power for low impedance HPs without a hint of distortion. It is a must for high perfoming DACs when you use HPs (especially when taking into account the price of atom).

A few days ago I had possibility to check out fostex th-610 (25 ohms, very strong tesla magnets and famous biocellulose drivers know from its epic, organic bass). These HPs when driving by some high impedance output amps thus with poor damping factor, can sound dull, fluffy, muddy, boring. Once you connect them to something like atom this is completely different beast. Amazing impulse response, control of transients, clarity & impact.
yeah , atom really is a beast of an amp, i would honestly pay $200 extra to have a solid aluminium enclosure + proper knob. I also have a beyerdynamic t90 and although those are quite easy to power using the dac magic, the improvements were still noticeable, though i felt it was not as drastic as on the aeon flow open.
 
Joined
Nov 23, 2018
Messages
32
Likes
30
#44
Any chance of any performance results with toslink? I always prefer optical connections - no ground loops etc...
if i recall correctly, kenrockwell said the optical was slightly better, but i don't know if he was using the Cambridge usb 2.0 drivers or not when he tested usb.
 

tired_guru

Active Member
Joined
Feb 8, 2019
Messages
100
Likes
79
#45
Any chance of any performance results with toslink? I always prefer optical connections - no ground loops etc...
Yeah I don't get it why amirm skips this very important test when testing DACs. It could show possibility of PLLs, FIFO or DAC itself if there is no extra jitter eliminator implemented to reject incoming jitter.

For me this is also crucial info. Only couple of DACs have been measured so far. Would be great to always see how jitter test goes on toslink input, just next to USB jitter results.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
3,584
Likes
1,842
Location
Zg, Cro
#47
Have you noticed same behavour as when one measure those IIR (minimum phase) filters they often can look a bad idea, or is that error in ones chain because often I/O clocks is not from same PCB or is there something about below.

Concern is when those filters is so steep a slope and being IIR they add a wide pass band phase domain distortion deep low into midrange area probably because they have to happen below nyquist limit which is not far apart filter target setting itself.

Examples is below where first (green) is a ES9023 USB DAC that will suppose is internal set as a FIR (linear phase) filter and second (blue) is a ES9018 USB DAC that will suppose is internal set as IIR filter, now look at grey phase traces which is asking REW how IIR phase should look for that amplitude response inside a 44kHz rate, DAC with FIR filter looks pretty close but DAC with IIR filter looks distort timing deep down and wide in passband and probably why some DACs can sound some weird or lets say stand out from any normal phase performance by the book.

View attachment 22406 .
Are you saying that an IIR filter with the same roll-off criteria cannot be designed with better phase characteristic?
 

tired_guru

Active Member
Joined
Feb 8, 2019
Messages
100
Likes
79
#48
About filters (from DM+ manual):

The Linear Phase filter is a filter offering low ripple in both the pass and stop bands, and what is known as constant group delay. Constant group delay means that audio signals of all frequencies are always delayed by the same amount when passing through the filter. All audio is therefore fully time-coherent at the output.
The trade-off with this type of filter is that due to internal feed-forward in the DSP, its impulse response will exhibit some pre-ringing. In other words, when excited with a theoretical impulse, the output has both a small amount of pre- and post-spike amplitude ringing (albeit well damped).

The Minimum Phase filter is a filter that offers even lower ripple in the pass and stop bands. Unlike the Linear Phase filter, group delay is not constant so some timecoherence is lost; however, phase shift is low and the particular benefit with this filter is that the
impulse response exhibits no pre-ringing.

Steep filter is a linear phase filter that has been optimised for stop band attenuation of close-in aliasing images. Here we have traded a little attenuation of the very highest frequency response (-2dB at 20kHz) and a little more pre- and post-ringing for a very steep attenuation just outside the pass band. The Steep filter is able to attenuate aliasing at 22kHz by some 80dB.
Using specific digital filter is a kind of compromise to be accepted between best impulse response, transient handling, post/pre ringing, positioning & echo:

1550939015518.png


Always preferred slow roll-off, even having no idea what was set. Checking afterwards it was this filter. Especially in classical, raw piano with sharp attack of cello & violin. This reminds me what I can hear alive during chamber ensemble classical concert, sitting in first three rows, about 6-10m away from the musicians. By looking at this table I could go even higher - to super slow roll-off, but the dacs I have tested had usually only slow roll-off, fast & linear phase, so standard slow roll-off was the best I could get knowing my preferences.
 
Last edited:

tired_guru

Active Member
Joined
Feb 8, 2019
Messages
100
Likes
79
#49
My measurements posted today all were taken through optical cable input. Attached is another test through toslink.
But are you running J-test on toslink input with signal from Audio Precision device ? The point of the test is to have especially distorted clock during transmition. Then this clock will be retrieved from spdif stream and will affect the dac performance (or not). Having result measured on analog outputs of the DAC, we can tell how it handles jitter on toslink input.
 
Last edited:
Joined
Jan 16, 2019
Messages
23
Likes
12
Location
Tucson, AZ
#50
I've owned a DacMagic Plus since 2012 and DacMagic prior or that. It has one feature that was very important to me - digital passthrough, supposedly untouched.

What this enabled me to do was come out of my PC via USB and then connect to my headphone amp at my desk and then also send the signal via coax to my speaker system across the room from the DacMagic Plus. It isn't a common feature - at least I've not seen too may dacs that can do that.

I'm not utilizing the DacMagic Plus as the moment but I was glad to see the mostly positive test results.
 

pma

Member
Joined
Feb 23, 2019
Messages
12
Likes
28
#51
But are you running J-test on toslink input with signal from Audio Precision device ? The point of the test is to have especially distorted clock during transmition. Then this clock will be retrieved from spdif stream and will affect the dac performance (or not). Having result measured on analog outputs of the DAC, we can tell how it handles jitter on toslink input.
No, I am running either Dunn's 16-bit J-test sent by Toslink (which measures nothing, no worsening on DacMagic Plus) or 24-bit 1/4 Fs test, where jitter translates into widening of the spectral line. Typical jitter problem then looks like attached, it is a 16-bit Dunn's test. As said, this test shows nothing with DacMagic Plus. The previous version, DacMagic, was worse re toslink jitter, it was possible to measure it clearly.

Edit: the measurement shown is not the DacMagic Plus. It is posted to show the jitter issue.
 

Attachments

Last edited:

pma

Member
Joined
Feb 23, 2019
Messages
12
Likes
28
#52
Using specific digital filter is a kind of compromise to be accepted between best impulse response, transient handling, post/pre ringing, positioning & echo:

Always preferred slow roll-off, .
Slow roll-off automatically means mirror images. Of course, subjective preferences do not necessarily match with best measured performance.
 

tired_guru

Active Member
Joined
Feb 8, 2019
Messages
100
Likes
79
#54
Slow roll-off automatically means mirror images. Of course, subjective preferences do not necessarily match with best measured performance.
Interesting. How mirror image translates into hearable audibly artifact ? Really curious what gets me so much into that filter. Reflections from the walls in small chamber hall during acoustic concert ?

For example, from technical point of view these are 2 characteristics of built-in filters into ak4452 (grace sdac), default & slow roll-off:








Can you comment more what impact can it have on sound just by looking at them ?
 
Last edited:

BYRTT

Active Member
Patreon Donor
Joined
Nov 2, 2018
Messages
160
Likes
181
Location
Denmark (Jutland)
#55
Are you saying that an IIR filter with the same roll-off criteria cannot be designed with better phase characteristic?
Guess i'm saying that when slopes are so steep as they become for some of those last decade of IIR filters its a yes when DAC is running at its native lower rates, where running DAC at higher rates say 192kHz problem is less or null either because they relax IIR filter at higher rates or mix it with FIR or because our interest is how look analog amplitude verse phase (coherence IIR) for audioband 20Hz-20kHz from input to output, and if that is a okay looking as a good old analog IIR domain we satisfied and probably don't care too much say phase has tiny mitchmatch way up higher.

Funny thing is its a IIR filter that distort phase at output to something non minimum phase coherence and FIR filter that output phase as is at input, so where in nature IIR filtering is the natural thing and FIR is not except for acoustic reflections and summing interference, for this particular case it looks distort phase at output, probably for the sake of 100% real time as @pos write about. Now advise others try take own measurements from input to output because mine was not using same PCB clock and that can move some numbers or precision in timing, but seen before some of the pro interfaces do the same scheme even they also have own input and same clock, and in that nature or enviroment is of minimum phase (IIR) or we can call it analog natural sound then how should a analog recording slip thruu such a DAC filter and be natural in original timing.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
3,584
Likes
1,842
Location
Zg, Cro
#56
Guess i'm saying that when slopes are so steep as they become for some of those last decade of IIR filters its a yes when DAC is running at its native lower rates, where running DAC at higher rates say 192kHz problem is less or null either because they relax IIR filter at higher rates or mix it with FIR or because our interest is how look analog amplitude verse phase (coherence IIR) for audioband 20Hz-20kHz from input to output, and if that is a okay looking as a good old analog IIR domain we satisfied and probably don't care too much say phase has tiny mitchmatch way up higher.

Funny thing is its a IIR filter that distort phase at output to something non minimum phase coherence and FIR filter that output phase as is at input, so where in nature IIR filtering is the natural thing and FIR is not except for acoustic reflections and summing interference, for this particular case it looks distort phase at output, probably for the sake of 100% real time as @pos write about. Now advise others try take own measurements from input to output because mine was not using same PCB clock and that can move some numbers or precision in timing, but seen before some of the pro interfaces do the same scheme even they also have own input and same clock, and in that nature or enviroment is of minimum phase (IIR) or we can call it analog natural sound then how should a analog recording slip thruu such a DAC filter and be natural in original timing.
Is your guess that IIR filters were used with ES9018 because there was not enough processing power to use FIR?
 

pma

Member
Joined
Feb 23, 2019
Messages
12
Likes
28
#57
Interesting. How mirror image translates into hearable audibly artifact ? Really curious what gets me so much into that filter. Reflections from the walls in small chamber hall during acoustic concert ?

For example, from technical point of view these are 2 characteristics of built-in filters into ak4452 (grace sdac), default & slow roll-off:


Can you comment more what impact can it have on sound just by looking at them ?
I am sorry but I would not like to speculate. Technically, mirror images (it is not "aliasing") in case of the DAC create new spectral lines above Fs/2, which can be very high in level. So, there is a possibility of HF intermodulations in the following amplifier, depending on its design and parameters. These intermodulations may transfer as difference tones into audio band. An example of the mirror image is shown attached. There should be only the left side spectral line. The line on the right side is a mirror image, created by DAC.
 

Attachments

anmpr1

Active Member
Patreon Donor
Joined
Oct 11, 2018
Messages
296
Likes
464
#58
As an owner of a Plus, I agree that the volume control has a strange kind of feel. However, I've not noticed the "feel" of the power switch, one way or the other. I have the silver face, which reminds me of '70s Japanese components, so I like that. No plastic. FWIW, complete instructions don't come with the unit, but are downloadable, along with so-called 'high res' (192) drivers. Cambridge Web support could use an updating; as I recall they show some Win XP configuration screenshots. But maybe that's changed. In spite of the roll your eyes 'Magic' in its name, the fact that decently implemented "old" tech does a good job shows you that, in digital, there's not really much new under the sun that makes a real difference.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,119
Likes
23,394
Location
Seattle Area
#59
Yeah I don't get it why amirm skips this very important test when testing DACs.
I skip them because USB is far more popular of an input and every DAC needs to get it right.
 

tired_guru

Active Member
Joined
Feb 8, 2019
Messages
100
Likes
79
#60
I skip them because USB is far more popular of an input and every DAC needs to get it right.
How much time does it take comparing to USB jtest ? Is it time consuming ? If it's not, maybe we could have a voting whether it is equally important to members and whether it should be included in following reviews ? (of course if you are willing to do it)
 
Top Bottom