• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Serious Question: How can DAC's have a SOUND SIGNATURE if they measure as transparent? Are that many confused?

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
Any possibility of a description of how the test was conducted and what the test setup was?
Google translate is your friend .

F4778AA4-7E4F-4D3D-8519-091E08F85AF0.gif
 
D

Deleted member 48726

Guest
The only (speaker) amplifiers that have such a poor DF are some tube amplifiers or 'specials' and in such case that poor output R is part of the 'charm' of the amp.
Those amps usually are not recommended for speakers dipping to 2 ohm in the bass area anyway.
Combine that with a wild varying speaker impedance and long or thin cables and yes it will be audible.

In practice most amps are above DF50 and people using speakers that dip that low use short and thick cables anway. Certainly when they are enthusiasts.

But... this thread is about DAC sound signatures and has no relation to speaker cables so perhaps discuss cable resistance in one of the many cable threads.
Well I questioned your claim about the need for an amplifier with a low damping factor of 10 for it to be audible. If you use higher than 14 AWG you get close to 1 dB error when using an amplifier with a DF <100 connected to a 4 ohm load.
1 dB is audible.
Many solid state amplifiers, especially vintage ones, have relative low DF.

But I won't discuss this anymore in this thread. :) Neither the definition of "lamp cord".
 
Last edited by a moderator:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Whether or not that error is audible or consequential depends on the impedance differences of the speaker. When that is fairly flat it is only an overall level difference and inconsequential.
Besides it has nothing to do with DACs.
 

ahofer

Major Contributor
Forum Donor
Joined
Jun 3, 2019
Messages
4,952
Likes
8,698
Location
New York City

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
Google translate is your friend.
Thanks, but no need for Google Translate - I have no problem reading Swedish. That picture doesn't help us much. Sure, it is a standard before/after switching setup, but it doesn't tell us anything about how levels were matched, and if the listening was single or double blind, or what statistical controls were used.
 

fpitas

Master Contributor
Forum Donor
Joined
Jul 7, 2022
Messages
9,885
Likes
14,191
Location
Northern Virginia, USA
But... this thread is about DAC sound signatures and has no relation to speaker cables so perhaps discuss cable resistance in one of the many cable threads.
Yeah. Then maybe we can understand why someone would run tiny wires to a speaker they care about :confused:
 

antcollinet

Master Contributor
Joined
Sep 4, 2021
Messages
7,409
Likes
12,294
Location
UK/Cheshire
Thanks, but no need for Google Translate - I have no problem reading Swedish. That picture doesn't help us much. Sure, it is a standard before/after switching setup, but it doesn't tell us anything about how levels were matched, and if the listening was single or double blind, or what statistical controls were used.
In other words we'd want the specific (and detailed) test descripition of the specific tests you are referencing. Not a generic description of a double blind test.
 
D

Deleted member 48726

Guest
In your first example output impedance (amp plus cable) is dominated by the amplifier.

In your second example, it is dominated by the cable - but only because you've changed your gauge from a reasonable 14awg to a less useful 18awg. In other words - thin lamp cord.

The only important rating of speaker cables is resistance - as long as you select low enough, then cables don't matter. I don't use smaller than 13, my current cables are 11 (2.5mmsq and 4mmsq)
Of course, but it matters.
DF = 160 returns an error of 0.5 dB (14 AWG)
DF = 80 returns an 0.7 dB error (14 AWG)
DF = 60 returns an 0.8 dB error (14 AWG)
Whether or not that error is audible or consequential depends on the impedance differences of the speaker. When that is fairly flat it is only an overall level difference and inconsequential.
Besides it has nothing to do with DACs.
It's audible compared to it not being there.

Which is easy to eliminate by having even length cables and going up in cable size and/or choosing and matching speaker and amplifier.

And most speakers have uneven impedance with peaks and troughs. I don't think I have ever seen a flat speaker impedance curve..?

It has nothing to do with DACs though.
 

Chrispy

Master Contributor
Forum Donor
Joined
Feb 7, 2020
Messages
7,751
Likes
5,910
Location
PNW
Of course, but it matters.
DF = 160 returns an error of 0.5 dB (14 AWG)
DF = 80 returns an 0.7 dB error (14 AWG)
DF = 60 returns an 0.8 dB error (14 AWG)

It's audible compared to it not being there.

Which is easy to eliminate by having even length cables and going up in cable size and/or choosing and matching speaker and amplifier.

And most speakers have uneven impedance with peaks and troughs. I don't think I have ever seen a flat speaker impedance curve..?

It has nothing to do with DACs though.
You really need to start a cable thread with that kinda stuff :)
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Just responding. I want to stop. :)


 

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,383
Likes
24,749
Location
Alfred, NY
Yes, DF being a product of output impedance and load impedance (cables included). Worsens with low impedance loads. My calculations with an amplifier with DF=10 and a 4-2 ohm load, 13 ft 14 AWG cable returns an attenuation of max. 3.12 dB. 18 AWG returns 3.42 dB error.

DF = 160 returns an error of 0.9 dB (18 AWG)
DF = 80 returns an 1.1 dB error (18 AWG)
The math seems off.
 

sq225917

Major Contributor
Joined
Jun 23, 2019
Messages
1,360
Likes
1,610
A very unstiff power supply regulation for an OP-amp : slightly higher distortion at full music levels, current clipping in extreme cases, possible lower power voltage driving the OP and for sure some minor compression sound effects . In the original Linn akurate DS the output OP amps lm4562 only have 9 volts because of this - read more here:

They had only 9v feeding them because that's the voltage linn had to play with given the penny pinching decision to continue using one of their existing smps, it wasn't a musical choice, it was a fiscal one.
 

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
They had only 9v feeding them because that's the voltage linn had to play with given the penny pinching decision to continue using one of their existing smps, it wasn't a musical choice, it was a fiscal one.
I doubt it was the reason . Its a very expensive player. They use a very soft power supply ( If you read the whole modification thread you will understand.) The reason for this is probably to gain less noise, and (maybe) to have a softer more mellow sound. The power supply in the Akurate DS is so soft it collapses from 16 V to 10,6 volt , and after the regulator its only 9 volt.

From Google translate:
”and I did some measurements on his Akurate DS. The situation is more difficult than I thought. First of all, it is actually clear what needs to be done. I count on:
The clock fraction looks decent, but the power supply of the clocks can clearly be improved - this power supply is quite crucial for the achievable cleanliness of the clock signals.

The exit stages. Oh man, same drama as in the O-Sneaky. Although LM4562 are already in there as OPs (good!), but then 300 ohms output resistance :roll: . Sure, there have to be output buffers. However, six pieces, two for the asymmetrical outputs and four for the symmetrical. The assertion in the audio test that discrete output buffers with Mosfets came after the OPs is a clear misinterpretation by the author: Both in the Akurate DS and in the Akurate DS/1, the Mosfets are only responsible for muting. This also applies to the Majik, which has the same circuit board but is not fully populated.

The RCA outputs of the O-Akurate are simply fed from the OP output, which is responsible for the positive half of the symmetrical signal. That's half-assed. That the RCA output is supposed to be the better one, as is sometimes claimed in the Linn forum (I'm not reading along there, but I was told that) is, from a technical point of view, sheer nonsense. You then only hear with half of the built-in DA converter. If you don't want to neglect the cinch outputs, you would have to do it like this: Take the symmetrical signal, calculate the difference with an OP and then feed the outputs via their own output buffers.

The output stages of the Akurate DS are only supplied with +-9V. That's a bit little in my experience. Above all, the output buffer should be supplied with at least +-12V. And there lies the problem: The original power supply does not supply enough voltage. In standby it's -19V and +16V, which come from the power pack (was already a dynamic or whatever the newer name is). But they collapse to -13.8 and +10.6V when the Akurate is turned on. This is just enough for the positive supply half so that the low-drop regulator (LM2941) can still work with the remaining difference of 1.6V. The power pack is designed to be extremely "soft". If I branch off 6x 15mA for the six buffers, 20mA for the clocks and 10mA for a double OP, i.e. a total of 120mA, the voltage of the power supply, which is much too meager for my taste, collapses completely.
Now you could say, ok, you need your own power supply for the additional installations. There's enough space in the case. However, this approach would torpedo the Linn concept that no 50Hz mains hum occurs in the electronics housing because the completely encapsulated switched-mode power supply already supplies DC voltage to the electronics compartment.”
——————

my thoughts:
I have tested driving lm4562 and the opa2604 with low voltage and the dynamics are clearly better with 15 volt. This is my (limited) experience .
 
Last edited:

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
Thanks, but no need for Google Translate - I have no problem reading Swedish. That picture doesn't help us much. Sure, it is a standard before/after switching setup, but it doesn't tell us anything about how levels were matched, and if the listening was single or double blind, or what statistical controls were used.

As an anecdote: When I participated (I think around year 1999) , the levels were matched +/- 0.01 dB at 1 kHz. A double-blind test between two CD-players, Denon DCD655 (from year 1999) and HK7600 (from year 1991). Testing place was Studio Blue in Stockholm. One person from the US had promised to travel to Sweden and join as a controller (from former AudioReview Forum) but he never showed up. The test itself was a real pain since it was not an ABX, but an AB-test with an initial training session. So no opportunity to switch and compare during the test. One person switched cable connections (hidden) according to a random scheme, and exited the room. Two other persons entered the room and one was the preamp switcher, and the other the listener. When the listener had decided, it was written down. The procedure was done 13 times. 11/13 correct.There was a short write-up published and some hot discussion about the results at AudioReview.

Other tests when I was there were always level-matched, but single-blinded (the switcher was however always out of sight of the listener). I would say that there were some doubts on that method, but some persons could really nail the correct source in consecutive series (more than 12-15 switches).

A final note is that the before-and-after test is not applicable on the CD-player test. I am also not aware of how current listening tests are performed; the above session occurred more than 20 years ago.
 
Top Bottom