No need for ad hominem attacks. If you don't have any sound stage measurements just leave it be.
So if we have DACs that have perfectly adequate frequency response for our sample music and the distortion measures identically over the audio band, BUT the soundstaging on one DAC is vastly superior to another, then there MUST be a difference in the analog output signals (otherwise they would sound the same) that should be measurable by the difference technique I mentioned earlier.
You could of course either listen or measure for this!
I wonder what “better soundstaging” actually sounds like as a difference signal?
I don’t know if my PC motherboard audio input has the resolution to measure this as this is the only ADC I have.
Or is the question that if the FR and distortion are same, CAN there be a difference in soundstaging? Assuming a modern well designed DAC. (Whatever that might be!)
You could of course either listen or measure for this!
I wonder what “better soundstaging” actually sounds like as a difference signal?
I don’t know if my PC motherboard audio input has the resolution to measure this as this is the only ADC I have.
Or is the question that if the FR and distortion are same, CAN there be a difference in soundstaging? Assuming a modern well designed DAC. (Whatever that might be!)
Absolutely, yes. IMHO and IME its a combination of things that would be harder to measure than FR and distortion, although they are factors too. In part its the correlated noise in the amplitude domain, and in part the correlated noise in the time-interval domain. There is some other stuff too, yet at this point those seem to be main suspects. Marcel has already done a remarkably good job with his reference voltage power supply (despite the 5XR bypass caps). Doing an equally good job with clock phase noise prior to the dac board is not trivial. There are other things that have to be done too, like rigorously avoiding ground loops, use of hum breaker circuits, etc. Then there is the output stage, which IME involves some tradeoffs which so far are not theoretically well modeled.
Again, there are issues with DSD dacs and with sigma delta modulators that are much more complex than what goes on in, say, a linear power amplifier. There are various kinds of weird dynamically changing intermodulation products that can look like some kind of noise on an FFT (because their energy is spread across bins over the time and FFT is acquired, and or because it is audio signal correlated). It takes a reasonably deep understanding of how DFTs and modulators, etc., work to start to get a handle on how that is different from steady state (PSS -- Periodic Steady State) distortion and noise measurements.
Last edited:
Thanks Mark. I am very curious to know what the scale of such differences would be as a difference signal. Anybody know this?
Well, we are removing or at least attenuating low-ish level dynamic noise sources that can mask low level audio details encoded on a well recorded CD. So it depends on whether the rest of the system after the dac is good enough to reproduce those unmasked low level details that carry perceptual localization cues including cues to perception of the recording space reverberation decays. Some systems, say, with low cost class-D amplifiers might re-mask some of those low level details. Even linear power amps with a shared power transformer which is too small can have crosstalk that obscures the cues. Ground loops in the system can mask low level details and make sound grainy or veiled (doesn't have to be just line frequency hum). So, the whole system including the speakers, room, any headphones, etc., should be as good as possible.
EDIT: I would also add that the listening position in the room should be located in consideration of the Critical Distance from the speakers.
EDIT: I would also add that the listening position in the room should be located in consideration of the Critical Distance from the speakers.
Last edited:
My system is a 3 way active with 4 power amps per speaker which’s of course have their own PSU. There is no audible noise from the preamps or active crossover at the listening position so I get pretty good imaging even from my modest DAC.
Based on your comments clearly the main localization cues for vocals, guitar etc will be at a very high level comprising almost 100% of the signal, but subtle room reverberation and fainter background sounds may be at say -60db or less below that level. I am fairly sure I can localize sounds that are at the level of master tape hiss on some tracks.
So the difference signal I am postulating would probably be at or below the base noise level.
Perhaps this may not be as useful an idea as I thought it might be!
Anyway - getting late here so goodnight from me for now!
Based on your comments clearly the main localization cues for vocals, guitar etc will be at a very high level comprising almost 100% of the signal, but subtle room reverberation and fainter background sounds may be at say -60db or less below that level. I am fairly sure I can localize sounds that are at the level of master tape hiss on some tracks.
So the difference signal I am postulating would probably be at or below the base noise level.
Perhaps this may not be as useful an idea as I thought it might be!
Anyway - getting late here so goodnight from me for now!
I would say you need to aim at least for the bit limit of a well dithered CD. Maybe as much as 18-bits, with the lower bits down in some dither noise. At least -93dBFS which is the top of the dither. If you can't hear low level details down to that or below, then IMHO there is a problem. If not, you probably couldn't tell the difference between CD and hi res. Also, many CDs are not always topped out. Sometimes there is some headroom such as in the quiet part of a symphony. There are definitely people who can hear dither noise without cheating by turning up the volume.
Anyway, part of the question is not if you hear noise when your turn up the volume and nothing is playing. Its more like if there is correlated noise produced in your system when music is playing. Ground loops, low-cost class-D, etc., can produce those types of dynamic noise problems. Sometimes it can be tricky to find and fix those types of problems.
If you want to PM, we could talk a bit more.
Anyway, part of the question is not if you hear noise when your turn up the volume and nothing is playing. Its more like if there is correlated noise produced in your system when music is playing. Ground loops, low-cost class-D, etc., can produce those types of dynamic noise problems. Sometimes it can be tricky to find and fix those types of problems.
If you want to PM, we could talk a bit more.
Last edited:
Would measuring channel separation be a start? ASR doesn't measure that.
I'm a bit short of time and peace currently but I'm curious to do a measurement where I do a standard FFT 1k hookup on one channel, measure and then feed the other channel a 7k signal and a 60Hz signal and have a look what changes. Has anyone ever done that? A channel separated IMD.
The reason for not measuring crosstalk at ASR is that it is typically very low in well measuring dacs. About -120dB for multi-channel DAC chips and -130dB (or less) for mono DAC chips. Actually the RTZ dac Markw4 has referred to in this thread has higher crosstalk which is another indication that claims of superior soundstage should be taken with a grain of salt especially since proper controls have not been used in listening evaluations.
Ok. But usually the output signal isn't taken directly from the chip. There could be I/V converters, output buffers, ... involved with various degrees of psu decoupling that would change the -120dB crosstalk figure? Also it seems a bit weird to do measured performance reports and then just go "we don't measure that, as usually it's at the edge of what we can measure and so assume this will be so with every single unit we encounter".
It might be wise to make sure you don't have asymmetrical crosstalk between channels. That can happen if there are layout problems.
That would apply to the whole system and not just the dac, BTW.
Please PM him, We are getting very bored with this load of rubbish. I don't care if the noise is -500dB or whatever, if the music sounds good I like it. Besides sound staging is a total myth. Demonstrate sound stage on a single speaker. Sound stage is created in the recording studio. If you don't believe this then go visit one during a recording in single booth or multi booth environment. I think you guys are delusional and think you hear things that is actually not there. I am done with following this thread as it deteriorates as it gets longer.
Not at all. If you record a symphony with a pair of condenser mics in a music hall, part of what the mics will pick up is the sound of reverberations in the room. There may be slap off that back wall that gets picked up. The question is whether or not the system can reproduce that low level information accurately. If it can, the the stereo illusion is one of a sound stage. You hear the sound of the music hall room between and behind the speakers.
You can reproduce depth information with only one speaker, but not lateral localization.
You are just recognizing time delay of the incident, nothing new or miraculous here. It has been evident since we developed ears.If I play you a signal of far away and then one of close-by you would not even know the difference.
- Status
- Not open for further replies.