Without more dynamic characterizations in the datasheet, I very much doubt that DAC would beat TDA1541A in this particular application. But you're welcome to expose the flaws in my reading of the datasheet 
First up - SFDR is quoted at 100dB at a sample rate of 10kHz. How much would this be likely to degrade by 44k1? TDA1541A quotes -100dB THD whereas your suggested DAC sports a mere -97dB. The sample rates are rather different too (174k vs 10k).
The glitch impulses quoted in the AD5791 don't fill me with any confidence whatsoever as to how it will perform dynamically.
So - nice try, but no cigar After all, this is audio not industrial process control.
First up - SFDR is quoted at 100dB at a sample rate of 10kHz. How much would this be likely to degrade by 44k1? TDA1541A quotes -100dB THD whereas your suggested DAC sports a mere -97dB. The sample rates are rather different too (174k vs 10k).
The glitch impulses quoted in the AD5791 don't fill me with any confidence whatsoever as to how it will perform dynamically.
So - nice try, but no cigar
After all, this is audio not industrial process control.Attachments
Last edited:
I could spend a couple times more than the 10 min I searched for the AD5791 to come up with a better dynamic speced modern part
I assume you would object to any delta-sigma parts despite the fine audio # for the ESS DAC
AD part's SFDR is a potential issue, THD disappointingly low for the static specs
do you have the SFDR number for the TDA1541 - the original Phillips datasheet doesn't show it
but then it never shows the THD+N line in any graph better than 92-3 dB for the TDA1541
if AD part's only issue is glitch energy then reconstruction S/H would remove it
I assume you would object to any delta-sigma parts despite the fine audio # for the ESS DAC
AD part's SFDR is a potential issue, THD disappointingly low for the static specs
do you have the SFDR number for the TDA1541 - the original Phillips datasheet doesn't show it
but then it never shows the THD+N line in any graph better than 92-3 dB for the TDA1541
if AD part's only issue is glitch energy then reconstruction S/H would remove it
Up to you - I've searched a fair while myself and haven't turned up anything better spec-wise than the TDA1541A. Which was the reason for my suggestion.
The ESS DAC datasheet admits to DC noise modulation (assuming you're speaking of the 9018 Sabre). The point of the exercise is to uncover such artifacts in S-D DACs, so an S-D reference DAC where noise modulation is admitted would defeat the object.
ADI has some impressively spec'd parts for comms applications, but I haven't seen 16bits so far. Potentially their 14bit parts suitably oversampled might suffice, but the SFDRs do look rather marginal.
I've been speaking of the TDA1541A which does claim -100dB THD in the datasheet - page 6, right at the bottom, Feb 1991 edition.
Can you recommend a suitably spec'd S/H?
Semantic hairsplitting. If you want to get anybody to cooperate with you, you'll have to learn to control your tendency for timewasting.
Like I said earlier, get a good null (2 devices), chances are both are transparent. Null both those with a third, almost certainly transparent.
Let's not go looking for solutions for problems that haven't been shown to exist. We can deal with the integer delay and see what's left. Just include a command line parameter in the program to invert one channel into the other that allows sample shifting of the L channel against the R. Observe a waveform with some identifying features on 2 channels of a scope, one from each DAC. Move on from there.
Many DACs operate at multiples of Fs, and clocks have frequencies of multiples of Fs. If it becomes necessary an FPGA with pipelining will permit the shifting of the SPDIF signals with respect to one another. Better to avoid this and any synchronisation or jitter problems associated with it if possible though. In the last resort the data can be stripped from memory, but I'd rather avoid the effort of halfway-building the device scott didn't really want and I'd prefer to be using 2 real-world fully implemented (SPDIF or USB) devices.
It's just a question of how far down the residual is with respect to the signal. Frequency response is expected to be flat to a fraction of a dB. What are we about here, conducting an experiment or arguing about a thought experiment?
jcx, scott. It's hard enough keeping abraxalito focussed without two spoilers running interference. Please confine your remarks to the original thread topic, or at least to genuinely useful contributions to the solution of the problem he has requested help with.
Quite obviously you're wasting your time here - I'm done
you have just demonstrated his point, you just love wasting people's time with theoretical nit picking scenarios
..and the above statement goes so far as to even acknowledge this....
great thread here ! a couple thoughts..........would it be fair to equate those who have a problem with nwavg to those who have a problem with free energy scientists? what a problem it would be for certain segments of industry if we no longer forced to consume oil for energy, not just snake oil...........2nd and more seriously, I suffer from fairly severe tinnitus. ringing in the ear, but thankfully still have near perfect hearing of frequencies. And let me tell you, there is 99.9% snake oil being peddled out there in the unregulated market for proposed alleviations for reducing tinnitus. It's a bunch of nonsense. The analogy's are not the same since I believe most audio co's are in fact trying to put out a decent product, but we have to admit, alot of the gear being made today is substandard, not just from a measurement standpoint, but from a reliability/compatability/poorly engineered standpoint. Plus alot of it just doesn't sound good imo (good means natural and balanced to me). My god there are so many unatural sounding headphones and hifi gear imo. But it's not because of malintent, just people totally lost in a subjective world of what sounds cool to them, and it's usually just MORE. more punch, more fidelity, more clarity, more dynamics (the 'limited' dynamics of an LP is perfect for me thanks), more realism etc to the point that it's no longer in touch with reality. Like these hidef tv's where the people on the screens look radioactive. How is that more accurate or 'transparent' really than a good ol tube tv with it's slightly soft warm image? I mean if you sit next to a friend under an oak tree, does his image look more like a tube tv or a digital one? he looks softer and darker than if we were 'out in the open' and frankly some prefer life under an oak tree rather than out on an open plain.......Another not perfect analogy but transparency, "having the property of transmitting rays of light through its substance so that bodies situated beyond or behind can be distinctly seen" can also be taken too far and too much of a good thing maybe. I mean whose to say gear is not a part of the process of listening to music. But maybe it's nice to have the choice for those who believe it shouldn't be. Alot of money is being waysted by kids who don't have two nickels to rub together. enter the robinhoods of hifi. it's all good. I'm one who would like to see less choice of gear frankly. Finding a decent headphone rig shouldn't require years of internet and purchase searching imo. And the more poor choices one has to wade through, the longer it seems to take.
Analog or Digital? (1993) - YouTube
neil young" we got control of the sound, but we lost the sound"
objectives get control of it but lose it. subjectivists have no control of it but sometimes have it.
neil young" we got control of the sound, but we lost the sound"
objectives get control of it but lose it. subjectivists have no control of it but sometimes have it.
Wine spectator uses single blind taste tests. I know several people who actually work at wineries, and they occasionally use single blind taste tests, but mostly sighted tests. There are several studies which show cheaper wines beating expensive wines in comprehensive double blind taste tests. If what you said was true, then these expensive wineries are making all kinds mistakes.
Your strong claim that everyone everywhere has used double blind taste testing is utter nonsense.
The biggest threat to western audio is the Asian markets. There is no way we can compete with the likes of China on price.
I tried building an amp for less than the Chinese and failed miserably. I couldnt even get the components for what they are selling for. I have had pcb's made in China and they were a third of the price they are in the UK and that includes p+p half way around the world !
I sell software and was the only one selling that software on ebay, now there are numerous competitors selling much chepaer than me.
I am seriously wasting my time ion the general hardware and software markets.
The only way to make money is to specialise with technology above what the CHinese can make.
I tried building an amp for less than the Chinese and failed miserably. I couldnt even get the components for what they are selling for. I have had pcb's made in China and they were a third of the price they are in the UK and that includes p+p half way around the world !
I sell software and was the only one selling that software on ebay, now there are numerous competitors selling much chepaer than me.
I am seriously wasting my time ion the general hardware and software markets.
The only way to make money is to specialise with technology above what the CHinese can make.
Having worked in that industry for more than a decade, judged at competitions in both California and France, and personally set up, run, and participated in hundreds of DBTs of winemaking variables at my own company and at our customers and investors' facilities (including Beringer, Mondavi, Clos du Bois, Gallo, Kendall-Jackson...), I can confidently tell you that you're wrong. So can anyone with a MW or WCET certification, who must pass DBTs for their qualifications.
Part of your confusion is terminological- "single blind" and "double blind" have a different meaning in the wine world than anywhere else. A "single blind" for wine tasting can certainly be "double blind" in the usual sensory definition of "no one in the room knows which is which." In wine tasting, "single blind" means that the tasters know which wines (or variables in the same wine) are under test but not which wines are in which glass.
Edit: And I would be the last person to claim that expensive wines always taste better to everyone than cheaper wines. In fact, the opposite is often true- I find many of the VERY expensive wines out there (e.g., Super Tuscans, top-shelf Napa Cabernets like Screagle) are bloody awful.
Uh ... not quite ...
Well, abraxalito, it seems that folks at Schiit audio don't quite agree with your UNCONFIDENT claim ... ergo, they may've gone for that DAC IC in their their new(er) high-end D/A, the Yggdrasil.
Schiit exhibit Fulla USB and Yggdrasil DACs at Canjam '14 | DAR
Without more dynamic characterizations in the datasheet, I very much doubt that DAC would beat TDA1541A in this particular application. But you're welcome to expose the flaws in my reading of the datasheet
First up - SFDR is quoted at 100dB at a sample rate of 10kHz. How much would this be likely to degrade by 44k1? TDA1541A quotes -100dB THD whereas your suggested DAC sports a mere -97dB. The sample rates are rather different too (174k vs 10k).
The glitch impulses quoted in the AD5791 don't fill me with any confidence whatsoever as to how it will perform dynamically.
So - nice try, but no cigar
Well, abraxalito, it seems that folks at Schiit audio don't quite agree with your UNCONFIDENT claim ... ergo, they may've gone for that DAC IC in their their new(er) high-end D/A, the Yggdrasil.
Schiit exhibit Fulla USB and Yggdrasil DACs at Canjam '14 | DAR
Perhaps this is just a trolling outing, but as you've read my blog (DIYA stats show you've visited more than once) then is it possible you missed this post?
http://www.diyaudio.com/forums/blog...tts-flagship-dac-uses-not-audio-dac-chip.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.