I would have to do some research (listening), but there is a Bob Marley song I have somewhere that has a rumble in it that sounds a lot like a truck doing a low speed maneuver, presumably outside the studio, or maybe warehouse...
To clarify, I don't think it matters how the reverb tails are produced. They can be natural or synthetic. The loss of audibility occurs either way. I happened be listening to music that sounded like complex reverb and delays as are commonly used for pop music recordings. The article I linked to where someone else noticed the same issue seemed be more focused more on the sound of naturally occuring reverberant spaces.
Not sure exactly what you mean by 'what sort of jitter.' Different types of dacs can have different sensitivity to jitter. Random jitter probably tends to be less objectionable than deterministic jitter. The amount of jitter at 10Hz offset from the nominal clock frequency is often considered very important for sound quality (mostly a human sensitivity factor, they say), and low jitter down at 10Hz offset and 1Hz offset is hard to achieve at the most commonly used audio data converter frequencies such as, ~24MHz, ~48MHz, and 100Mhz.
Hard to achieve in the stationarity of the air in a listening space too. http://www.davidgriesinger.com/ASA/Modeling_air_currents.pptx
Easy to measure put a speaker/microphone in your jitter measuring chain. Measuring this stuff as electrical signals plugged directly from one instrument to the other is basically a waste of time.
Last edited:
Push-Pull and Diff......cancel 2H. How do you cancel 3H? The more audibly annoying harmonic.
-RNM
Last edited:
I already covered that ground quite awhile ago here. But....
I want to design out all detectable distortions thru the entire system.
If I can now detect .04%..... I might detect .01? Havent tried yet now to see where My limit is at. But lets say .01% is the new level to which I (me) cannot detect a difference . I would want my gear to be better by some margin.... as others may be able to detect change lower than I can. Say 5-10 times better?
But, we also listen to the effect of added distortion thru the whole listening chain (and recording). So, we would now like to see < .001-005% from record to playback in order to know for sure
_)) no one can hear/detect any difference in the entire system (<.01%)..... worse case summing.
If some hear better than I do... we are below the .01% mark.
-Richard
Last edited:
Fundamental symmetrical compressive distortion, expansive compensation blows up eventually.
Of course. Where it may come into play is in ASRC where tricky mathematical upsampling to several tens or hundreds of GHz is performed. In a very rough analogy it is perhaps similar to an FFT vs DFT where the calculation can be cleverly sped up. In the case of ASRC, many of the upsampling calculations can be skipped so the problem becomes calculable in real time. But, the effect of the jitter's timeframe may be relative to the very high effective upsampled frequency timeframe. In that case very small time differences may equate to a significant number of bits of error in calculations of sample values. If you will pardon me for now, I haven't finished reading yet, but I'm working on it, and the part on jitter is coming up soon. I have a lot of material to go through which will probably take some time. Of course, you may already be up to speed on all that in which case I would be happy to have you explain it to me.
Last edited:
Not the best analogy, there is only one answer and it is exact the speed enhancement is just that there is no loss of accuracy at all.
Are you saying there are DACs made in last 20 years that produce audible jitter?
No, angle is not the angle between the images. It is the angle off center image.
The numbers to discern images an inch apart laterally at center stage is different from the numbers required for two images at a 30 or 45 degree placement off center.
In addition, depth discernment is also quite interesting.
Jn
Mark, one question. Jitter produce distortion (And FM distortion). The way you describe this reverb tail loss is level connected, right ?
Could-you explain ?
Don't know what all this tech sounding stuff is..
I do know Bob's bassist absolutely loved my k horns. Loaned him 4 with 1.2 kW of tigersauruses for a small concert. My friend was into the wailers, got some free tickets.. I was a dumb geek, who knew? I believe the bassist loved the fast bass.
Jn
Depends where one is. As I said, I don't necessarily stay on the centerline. I get up walk around in front of the speakers, turn my head in different directions, etc., so I am using a lot more information to estimate size than if I were staying on one place or using headphones. Still, it is an interesting point. I will have to try it again sometime and see if I still think I get the same results. Also, I may have used S1 to do the panning, don't remember, but it can do that. They say the shuffling is phase corrected, don't know about simple panning and placement. Unfortunately, there is way more interesting stuff to waste time on when I already have plenty on my plate as it is.
Don't know about FM distortion. Only think that the upsampling and converting from 44.1 CD audio to DSD64 sounds worse in virtually all respects than going to DSD256 instead. The exception is audibility of reverb tails is better at DSD64. Reverb tails are also better on an old, obsolete whatever that dac was that Abraxalito sent me (don't remember the part number at the moment). Seems much like what was reported with the Border Patrol TDA1541 dac in the article I linked to. I think the difference between all the above things that relates to audibility of tails is jitter, either jitter resulting after upsampling and DSD conversion or lower jitter sensitivity of old dac technology vs modern S-D dacs. Also, the thing about DAC-3 is that it is designed to work as well as possible in a studio or other professional environment where SPDIF compatibility is required. SPDIF tends to have high jitter, and tolerance of high jitter from SPDIF may result in DAC-3 being configured in a way producing a bit more jitter all the time vs dacs which don't have to support high quality SPDIF.
Similarly, adjustments I make to DSD DPLL bandwidth in my modded mobile Sabre dac affect the relative jitter contributions of external jitter vs internal vco jitter. I find that jitter from external sources tends to be the biggest problem there, and if I increase DPLL bandwidth, the influence of external jitter increases and reverb tails become inaudible.
Everywhere I turn, it seems the audiblity is linked to jitter levels or sensitivity to jitter. In the case of Allo Katana Sabre dac, for awhile it had the lowest jitter ever measured at ASR, even lower than DAC-3. And, on it jitter tails come through as well as any Sabre dac I have listened to. I find I can match Katana reverb tails with my modded dac by adjusting DPLL bandwith, or make the tails disappear with a different DPLL bandwidth setting. Again, in all cases the presence or absence of reverb tails is linked to jitter levels or jitter sensitivity.
So, the case I have pointing at jitter is entirely empirical and circumstantial. I don't have a theoretical explanation to offer at this time, but perhaps eventually something will start to click. Have to recruit my System 1 to work on it while I attend to other things. Maybe someday I will wake up in the middle of the night with insomnia and suddenly realize the answer when my when I happen to be in a receptive state of mind and my System 1 is ready. I never know how to predict things like that, but usually I get it eventually if I keep working on something and thinking about long enough.
Last edited:
Why would you convert to DSD anyway? There is no point that I can see? The ESS DAC is basically an array of 6 or 7-bit Sigma Delta converters.
Well, I actually don't even see a point for DSD to exist in the first place... but that's another discussion.
Perhaps the lack of errors is where the missing reverb tails went. They may be an artifact of jitter or some other digital error. It seems there is some correlation in your description between reverb tails and older, less linear DACs.
You could test the jitter theory by swapping in a voltage controlled master clock and modulating the voltage pin.
You could test the jitter theory by swapping in a voltage controlled master clock and modulating the voltage pin.
- Status
- Not open for further replies.