Has it? Hildegard von Bingen, Johann Sebastian Bach, Wolfgang Amadeus Mozart and Philip Glass wrote/write very different music.
Hi Ken,
May I comment that your links concern R-2R voltage output Dacs.
I thought it may be interesting to add to this that TI has changed it's own view on I/V conversion for the 1792/1794 current output Dacs, see page 3 to 6.
https://www.ti.com/lit/an/sboa237/s...953343&ref_url=https%3A%2F%2Fwww.google.nl%2F
In the images below their suggestion for the I/V with a 5534 from 2003, but 14 years later in 2017 they came with a new circuit diagram.
A much faster OPA1612 with a much smaller cap.
That seems quite a lot better.
Hans
.
Attachments
Why stick to a bipolar op-amp when a FET input op-amp's input stage can have a much larger "linear" range?
A delayed answer to the question OS vs NOS in a short post. You may call it burn in or that I had to get used to the new sound, but I really needed to spend some time with it before drawing safe conclusions. The expectations for a "night and day" situation were a bit stressing and misleading. Now I can tell better. So for starters, it's true that both modes recreate the same soundstage in terms of dimensions, instruments size and placement. Also, the treble droop is perceived as such only for frequencies higher than 12kHz and only for constant tunes. Transients do not help to detect the difference in signal level. And all that said, yes, NOS sounds different than OS... Marcel's intersample overshoot test was revealing. OS is hard clipping and easy to distinguish from NOS. I just want to add my finding that in NOS mode at high volume there is a focused noise like in phase white noise which is absent in OS. Also not there in NOS when playing other files. Going back to music after I had trained my ears with that, I found that it easier to detect the difference in loud prolonged passages like for example brass. Especially with loud vibrato. Something like this: Ennio Morricone - Sergio Leone 30th Anniversary - The Trio [Remastered] ?? Audio - YouTube Vibrato is audible in both modes but only in NOS it can be felt live. Once I knew where to focus, I could tell NOS from OS with and without tweeters. OS makes an admittedly good try to convince for a live event but NOS just does it effortless!
Just turn down the digital volume control in your PC a few dB. If it is done in software before the dac then hard clipping should be taken out of the equation.
I use the digital volume control all the time in my system. I didn't see it to make a difference in listening and measurements too with this test. Perhaps the PC volume control has not the same effect as the DAC/usb interface control panel?
Regarding intersample overshoots, if there is a sample rate converter or something else involving filtering with no headroom before the digital volume control, then the damage is probably already done before reaching the volume control.
Can you just copy the audio files and halve the levels?
Can you just copy the audio files and halve the levels?
Hi Ken,
I have put together several days over the years most OS sigma Delta types. I have also tried a few cheap nos dac boards out, which I didn't think much of. But recently I picked one of Richard's phidecca nos face with the 7th order filter. I also have an AK 4493 dad. Both face have salas shunt psus and utilize jlsounds USB to i2s inputs, so everything except the race are the same and should make for a good comparison. The dacs; do sound different. To my ears the nos sounds more realistic on vocals and instruments, maybe it's a decay thing. The nos sounds a little softer but seems to have a wider, deeper soundstage. It's a very pleasant dad to listen too. The os is very clean and dynamic but seems more strident in it's presentation. I auditioned both extensively using fullface drivers, Jordan eikonas and class a amps, the nos stayed in my system.
Certainly, and also, Tchaikovsky, Stravinski, Korsakov and Rachmaninov are different in style, as are Holst, Hanson, Copland, Bernstein, Williams, etc. etc.
Yet, I think, if you played no more than 10 seconds of any of their works to the average person, they would immediately recognize and classify them as, 'classical music'. No matter, those of us who love classical of all forms will continue to do so 🙂, as I suspect, will lovers of rap and hip-hop 🙂.
Excellent report, Kostas. This is what I hear as well. That NOS and OS sound different even without instrumental upper overtones present. This adds evidence to the suspicion that the relative lack of image-suppression with NOS playback is not the reason for the subjective difference between it and OS playback. In your experiment, NOS is filtered acoustically by your ears, and also by your tweeters being completely disconnected. Yet, the OS/NOS subjective difference remained apparent.
By the way, feeling time pressure definitely is not conducive to comparative sound evaluation. However, your suggestions, tests, results and reports have all been extremely helpful.
The following statements:
... and this quote from abraxalito:
“The experiment with PhiDAC was bypassing a 3rd order LC (CLC) so not a particularly steep one.
In relation to filters and OS vs NOS I have one other observation which I'm not sure I've set out
before. You remarked on my slightly preferring 2X0S to NOS when listening on a PhiDAC hex (I think
it was that model). That result was a reversal of a much earlier experiment I did with an unfiltered
TDA1387 design when I preferred NOS to 2X0S. At that earlier time I noticed the 2X0S sound was
more 'greyed out', had more of a metallic flavour than NOS...”
… sum the NOS vs OS really nicely, I am of the exact same opinion as the blokes above.
My first NOS DAC was DIY TDA1541 used with Audiomeca Mephisto II CD Transport – this was by far the best analogue-sounding combination I heard. Pure bliss. It had that natural drive, emotion and overall presentation (extension, depth, natural colour of the harmonics) that was so much like analogue records playing on a decent system.
Then I tried dual differential 1704 design (Denon CD player), heavily modified throughout (the best I/V I arrived at was with AD811, which took me a long while to get right). It sounded great, “sophisticated”, but so far behind 1541 and Memphisto combo.
I tried DDDAC with very low noise regulation (I spent a lot of money on it!!!); it did sound great but required 2xOS to get rid off the high-frequency roll-off, and that was just ruining the listening experience for me. So, in pure NOS mode, the high freq. roll-off was an issue for me; with 2xOS –> I was getting just a bit of that metallic flavour…
I realised that DIY-ing will not get me where I wanted to be… so I started listening to various DAC’s at home, auditioning and comparing.
The moment I heard Holo May DAC was the moment I knew that I found what I was searching for all these years. Very natural, pleasant to listen to, able to re-create natural (or at least what I expect to hear…) ambience and harmonics’ richness. In addition, this DAC will also show what OS sounds like (hence, it is perfect to compare NOS vs. either PCM OS or DSD) – and it will handle anything and everything one can think of… certain HQ Plyer OS/dither/filters’ PCM combination, and certain DSD modulators/filters combinations will sound impressive (it seems that the sound is better than original!), but will all exhibit that metallic flavour.
I get the most listenable results with JRiver with its DSP disabled, playing raw (native resolution) files. And I do not get that feel of high freq. roll-off; actually, I feel so much top extension in NOS mode which is simply unbelievable (against all my previous experience) and the main reason why I got this DAC.
One thing I realised when I listened to Holo May DAC in NOS was how close its sound character matched the DDDAC – it was unbelievable. So, there’s something about NOS…
Audioscience Review has May’s NOS, PCM and DSD filters’ response plots.
So, it seems to me at least, that the minimum amount of DSP, coupled with letting a bit of hat ultrasonic noise through (no de-glitchers/steep, complicated / phase-relationship-ruining analog filters) is the best recipe
Last edited:
Hi, Hans,
During our discussions together, I've become impressed with your focused thinking toward this subject. As I had mentioned earlier, I'm not a S&H adept. I'm simply intrigued by what Frans Sessink has highlighted about the issue of settling-time error. While Frans' ideas are interesting, they may represent only a relatively small potential for subjective sound improvement given today's faster settling converters. Your suggestion to utilize an RC filter to suppress settling-time error may turn out to be the more effective and practical solution now. I would have no trouble with that outcome, as I'm not married to the S&H concept. Just intrigued by it.
Because of the above, I was wondering if you would consider becoming the technical leader for our thread on the subject of DAC and I/V circuit settling-time error mitigation? I feel that everyone would benefit from your lead in this area. 🙂
Last edited: