I understand that dynamic range, THD isn't everything, sometimes the IMD matter more and so on. I know the RME has received excellent reviews and while it might not be end all be all in the DAC wold, but if RME have the clock and jitter noise down to a science, wouldn't it be possible to extrapolate their SteadyClock FS and "hijack" the data signals and feed to an "emotional DAC" or simply use a similar clock architecture for a standalone solution ??? There is enough brain power around here for such a job 🙂
Is this a reasonable summary of the thread?
Some people have had informal and uncontrolled listening sessions involving various DACs and have shared their impressions.
Some others have stated they don't trust informal listening sessions. Some of those would prefer measurements.
It was then argued that it's not always clear which measurements are relevant, and that the usual measurements don't cover everything.
For some reason, we needed more than 650 posts for all this.
I think the question can be summarised in the following experiment: Suppose we have two DACs, identical with the exception of the clock. They are both exceptionally measuring but one has a decent clock and the other one a super duper femto clock with very low close in phase noise. The output difference between the two is measured. Is it a signal above a certain threshold, say -120Db? If so let us improve the "decent" clock.
If not, it should be completely inaudible by any human being. It would be interesting to see if a panel CAN hear a difference and what is the reason. Or an extremely low level signal with some particular structure can still be heard and why?
As stated before, the clock thread should have been moved to commercial sector very early in the piece. It would have solved a lot of problems.I think #5088 summed that up pretty Well 😉
https://www.diyaudio.com/community/...ystal-oscillator.261651/page-255#post-6843961
I always wondered why "well" in thread title was spelled with a capital W - last evening it finally dawned on me seeing Andrea coming back with a new user name or perhaps it is a ghost writer - language is a little to polished and correct - it was the name of a commercial endeavour and part of a gerilla advertisement scheme.. well o well.... ;-)
//
I have no idea why this was not done.... mods? The whole forum has experienced a slow decline to....
Carry on. 🙂
TCD
Lack of moderation is evident. Aggressive and obsessed members are allowed for a too long time to get away with behavior against the rules, resulting in many ruined threads. I find Digital Line Level section to be among worst in that regard and usually avoid visiting it.The whole forum has experienced a slow decline to....
Carry on. 🙂
TCD
I ottoh moderate myself ... Ja voll!
As far we do not organize a festival, all of this will stay words on non listened music (aka a jungle zoo)
As far we do not organize a festival, all of this will stay words on non listened music (aka a jungle zoo)
@mocenigo I come from studying loudspeakers and after some 7 years, a few things stand out in regards to what is or should be desirable to achieve an audible holographic reproduction. Aka, access the dimensional, space and time data available in many recordings. For this we need low amount of resonance modes in both the speaker and room, low amounts of harmonic and intermodular distortion as well as low to non existent diffraction. All of these contribute to loudspeaker and room coloration which is exactly what we want to stay away from.
Looking at the loudspeaker market we can see multiple interpretations of what a loudspeaker is or can be and while most designers and companies will claim their interpretation is a good one, we (the consumer) know this cannot be a correct interpretation, meaning there is a **** ton of BS out there.
Sometimes objectivity and subjectivity meet in agreement for both "sciences" agree.
If a loudspeaker is low in coloration and promote a good sound stage depth, width and imaging (which is a knowable science), the amplifier is controlling the drivers in a true push-pull configuration (differential-balanced or circlotron), the DAC have better than 120dB S/N which is the digital audio's dynamic range, then any form of transient reduction or smearing should be vanishingly low and lower than 99,99% of us are capable to hear.
Doede's DAC have, according to many users, an analog quality to it, which is the reason it got my attention in the first place. One of the reasons for the analogue sound property's is said to belong to the paralleling (error) correction which smooths out the "jitteryness" of the DAC chips.
But yet, people will argue and bicker. I've seen some audiophiles using subpar gear and complaining about others gear which from a technical standpoint perform better, aka cleaner aka more correct... and so forth. I think my solution is to simply get the RME ADI-2 Pro DAC, build my custom 1794 NOS dac and take it from there.
But thanks for the contribution ... everyone 🙂
Looking at the loudspeaker market we can see multiple interpretations of what a loudspeaker is or can be and while most designers and companies will claim their interpretation is a good one, we (the consumer) know this cannot be a correct interpretation, meaning there is a **** ton of BS out there.
Sometimes objectivity and subjectivity meet in agreement for both "sciences" agree.
If a loudspeaker is low in coloration and promote a good sound stage depth, width and imaging (which is a knowable science), the amplifier is controlling the drivers in a true push-pull configuration (differential-balanced or circlotron), the DAC have better than 120dB S/N which is the digital audio's dynamic range, then any form of transient reduction or smearing should be vanishingly low and lower than 99,99% of us are capable to hear.
Doede's DAC have, according to many users, an analog quality to it, which is the reason it got my attention in the first place. One of the reasons for the analogue sound property's is said to belong to the paralleling (error) correction which smooths out the "jitteryness" of the DAC chips.
But yet, people will argue and bicker. I've seen some audiophiles using subpar gear and complaining about others gear which from a technical standpoint perform better, aka cleaner aka more correct... and so forth. I think my solution is to simply get the RME ADI-2 Pro DAC, build my custom 1794 NOS dac and take it from there.
But thanks for the contribution ... everyone 🙂
It is always a good idea to have a reference...both to listen to music and have a benchmark to judge its own diy devices.
That's the truth. My friend insists the reference has to be analog, either vinyl or tape. I have seen for myself what vinyl can do at its best. Direct to disk lathe recordings such as Lincoln Mayorga have SQ that I have never heard from any direct to digital recording. Some hi-res digital can do pretty darn well, but still not beat the best of vinyl in certain SQ parameters. Other than that, vinyl still has its well known limitations.It is always a good idea to have a reference...both to listen to music and have a benchmark to judge its own diy devices.
Without a reference, people go off in all sorts of directions that aren't toward sounding 'real.' Fiddling with transformers, chokes, Mundorf caps, etc., won't get you closer to 'real' IMHO. Sorry. No offense intended to those who enjoy such things. Enjoyment is a really important goal, but for me at least real sounding is also a big goal.
I may have the devices needed to perform this type of measurement. ES9038Q2M DAC & ES9822PRO ADC and a full-duplex 32bit-perfect USB-I2S bridge that can use external clock for the DAC. Here is a loopback measurement of the setup. I have a lowish phase noise CSA309 based clock and some run-of-the-mill oscillators. What's missing is a SOTA clock.I think the question can be summarised in the following experiment: Suppose we have two DACs, identical with the exception of the clock. They are both exceptionally measuring but one has a decent clock and the other one a super duper femto clock with very low close in phase noise. The output difference between the two is measured. Is it a signal above a certain threshold, say -120Db? If so let us improve the "decent" clock.
If not, it should be completely inaudible by any human being. It would be interesting to see if a panel CAN hear a difference and what is the reason. Or an extremely low level signal with some particular structure can still be heard and why?
So I would play&record the same sound file using different clocks. Then the difference could be analyzed.
Some call the upgrades of an existing product for "the last mile", meaning it becomes the icing of a cake and not the freaking cake... LOL. And that is the problem I or we often face when we start talking. One, there are two many chefs trying to "force" you to replace the cake entirely rather than give you tips and ideas on how to improve the icing and two, they start with a horrible version or interpretation.
So I wanna take the opportunity to use an example.

See the principle operation for some of Yamaha's A-S series amplifiers. It is symmetrical ... s y m m e t r i c a l. Meaning there are TWO ACTIVE ARMS controlling the drivers in the loudspeaker vs. your standard amplifier which sends the entire signal via the (+) and reference the (-) to ground. This is asymmetrical or unbalanced operation - Out of these two, only Yamaha is capable of proper usage of the differential input. All asymmetrical amplifiers are BAL to SE converted internally. For those who are uninitiated, what Yamaha is using is known as Circlotron. I also believe Allan Wright of Vacuum State Electronics based his DPA (Differential Power Amplifier) on the Circlotron operation.
So for simplicity: Single Ended operation of a loudspeaker is inferior to Differential and/or balanced operation. This cover the underlying operational principle. A level up is which class of operation it is, aka class A, class AB and so forth. But this has nothing to do with how the loudspeaker is controlled, the class is a reference to how smooth the switching noise or lack off, is.. A level up from this could be something like type of feedback as in Negative, Feed Forward etc. Yet another level above feedback and we could start talking about which components are used, FET capacitance values etc etc...
But unless the foundation is of worthy quality, then it doesn't matter how good the icing is... It would be like trying to appreciate an onion cake and the baker ask you if you like the consistency of the icing ... LOL
My 2 cents
So I wanna take the opportunity to use an example.

See the principle operation for some of Yamaha's A-S series amplifiers. It is symmetrical ... s y m m e t r i c a l. Meaning there are TWO ACTIVE ARMS controlling the drivers in the loudspeaker vs. your standard amplifier which sends the entire signal via the (+) and reference the (-) to ground. This is asymmetrical or unbalanced operation - Out of these two, only Yamaha is capable of proper usage of the differential input. All asymmetrical amplifiers are BAL to SE converted internally. For those who are uninitiated, what Yamaha is using is known as Circlotron. I also believe Allan Wright of Vacuum State Electronics based his DPA (Differential Power Amplifier) on the Circlotron operation.
So for simplicity: Single Ended operation of a loudspeaker is inferior to Differential and/or balanced operation. This cover the underlying operational principle. A level up is which class of operation it is, aka class A, class AB and so forth. But this has nothing to do with how the loudspeaker is controlled, the class is a reference to how smooth the switching noise or lack off, is.. A level up from this could be something like type of feedback as in Negative, Feed Forward etc. Yet another level above feedback and we could start talking about which components are used, FET capacitance values etc etc...
But unless the foundation is of worthy quality, then it doesn't matter how good the icing is... It would be like trying to appreciate an onion cake and the baker ask you if you like the consistency of the icing ... LOL
My 2 cents
A speaker is normally a floating device. It doesn't know if it is driven from a differential or single ended source, hence this isn't true. Rather it is the nature of the amplifying source that determines the outcome.......Single Ended operation of a loudspeaker is inferior to Differential and/or balanced operation. ...
Though I would agree that symmetric balanced lines seems the way to go for advanced audio design, though not necessarily when driving a speaker.
Last edited:
What is missing here is the BEMF generated at the negative polarity of the voice coil which acts as a break or dampener and if the ground is not taking care of the force properly, you have a degrading transient performance. The active portion of a symmetrical amplifiers absorbs this in a better faction compared to if the neg is only reference to gnd 🙂A speaker is normally a floating device. It doesn't know if it is driven from a differential or single ended source, hence this isn't true. Rather it is the nature of the amplifying source that determines the outcome.
Without a reference, people go off in all sorts of directions that aren't toward sounding 'real.' Fiddling with transformers, chokes, Mundorf caps, etc., won't get you closer to 'real' IMHO. Sorry. No offense intended to those who enjoy such things. Enjoyment is a really important goal, but for me at least real sounding is also a big goal.
Absolutely everyone has a perfect and extremely familiar reference for what 'real' sounds like, it's all sounds they hear all day, every day. The most useful sound being the human voice, a primary component of most music.
The most useful sound being the human voice, a primary component of most music.
So you can hear where Topping D90 starts failing? At vocal frequencies and above?
A single ended amplifier can be represented as a voltage source with a characteristic output impedance. In balanced mode it can be represented by two symmetrical amplifiers each having the same characteristic output impedance. In this configuration the current path through the two symmetrical amplifiers (as looping back through the speaker load) now has double the characteristic output impedance. The ground is only a reference point. This is to state that a symmetric balanced amplifiers can be represented as a single ended amplifier with double the characteristic output impedance as far as the speaker is concerned. Damping normally deteriorates in balanced mode, as far as speaker connections are concerned.What is missing here is the BEMF generated at the negative polarity of the voice coil which acts as a break or dampener and if the ground is not taking care of the force properly, you have a degrading transient performance. The active portion of a symmetrical amplifiers absorbs this in a better faction compared to if the neg is only reference to gnd 🙂
Last edited:
How about going to the concert hall every other month and a visit to the local small jazz club once in a fortnight?Without a reference, people go off in all sorts of directions that aren't toward sounding 'real.' Fiddling with transformers, chokes, Mundorf caps, etc., won't get you closer to 'real' IMHO. Sorry. No offense intended to those who enjoy such things. Enjoyment is a really important goal, but for me at least real sounding is also a big goal.
For a reference, that is.
mk4, do you do this?
//
- Status
- Not open for further replies.
- Home
- Source & Line
- Digital Line Level
- The battle of the DACs, comparison of sound quality between some DACs