KeithL
Administrator
Posts: 9,955
|
Post by KeithL on Sept 26, 2013 9:51:52 GMT -5
A few things to consider.... 1) Pretty well all modern DACs are of the oversampling variety (oversampling is simply upsampling to an even integer multiple). The delta-sigma process also involves mathematically "trading" bit depth against signal frequency. (It's not quite as simple as saying that the DAC chip "just" upsamples the audio then does a simple conversion). This is going to happen no matter what you do, although the parameters will be automatically chosen based on the input sample rate. 2) The process of upsampling itself CANNOT create new information; it does not "improve sound quality"; what it does is to simplify other parts of the conversion process (like the reconstruction filter), so that THEY deliver better sound quality (it avoids the necessity of other compromises). However, ALL upsampling processes involve rather complicated digital filters, and every one I've ever listened to DOES change the sound - although sometimes the difference is VERY subtle. My point here is that the results WILL sound subtly different if you use different software to do the upsampling (because their filters will be different); using FooBar to upsample may well sound different than using Audition, or Windows itself,or the Asus hardware or software, or some other program ( DO NOT assume that any program upsampling 16/44.1 to 24/96 will produce an identical result). There are even some players that offer you a choice of multiple types of upsampling filters (I think HQPlayer is one of them) - and they do sound noticibly different. 3) Personally, I find that the differences are most noticeable on "metallic" sounds, like well recorded wire brush cymbals, and on how certain voice sibilants sound, and not so much (if at all) on things like vocals and drums. The differences in how the various digital filters handle transients are on the order of microseconds or milliseconds (the changes usually involve pre and/or post ringing - but, again, on the order of milliseconds at most); they're more likely to affect the "texture" of the individual wires hitting the cymbal, or the texture of a high frequency sibilant, than to alter the shape of a (relatively long and slow) decay tail on something like a drumbeat. 4) Since virtually all upsampling involves SOME filtering, which is applied at the point where the upsampling is performed, it seems reasonable to expect that the effects would be cumulative, and so it would be best to avoid multiple or repeated upsampling steps. New experiment for XDA-1. Experiments with WASAPI vs regular using short samples. The XDA-1 upconverts audio to (assumed 24/192Khz) at the DAC chip itslf to process. My question....would there be a quality benefit to upsample my CD's 24/192 before transporting to the DAC. The goal being seeing whether there is a quality difference between the upconversion at the DAC vs the upconversion done at a different section of the audio stream. A few caveats. The conversion is done in a two step process. Foobar in WASAPI mode first upconvets the signal from 16 bit to 24 bit (assumed) 44.1 Then my asus xonar essence sets the output signal to 192 Khz. So there's actually two conversions. Audition track: Beatles every little thing she does. 1:00 for about 20 seconds. 2009 stereo remaster. There are two accent drums one slightly different than the other and the latter has a different decay and that's what I concentrated on. XDA-1 at 80. Passive pre-amp turned down to make volume listenable. So direct sound: no wasapi mode @ 16/44.1 vs WASAPI 44.1 Hard to tell a difference. Felt like WASAPI mode was slightly clearer/less warm (less smeared?) WASAPI mode 44.1 vs dither 16/44.1 Dither sounded slightly clearer. Hard to tell a difference. 24/44.1 vs 24/192: Okay here it is....hard to tell a difference using really short samples in nearly any output mode. So conclusion: I didn't hear any difference. Certainly couldn't pass a blind test. Also note: my ears are slightly blocked. Not too bad, but enough for me to notice. Now the "feeling". I felt there was a difference. A very slight difference. A teeny tiny one. But I had no idea what it was. Sort of like a space between the noise. Also some past history, over time, I experienced more enjoyment with WASAPI mode. It simply felt less loose but when asked to pinpoint why, I couldn't at all. Not even an inkling. Another thing to note: The beatles release some REALLY poor recordings in terms of recording quality. Not all, but gosh there is so much audible distortion....
|
|
|
Post by garbulky on Sept 26, 2013 10:13:45 GMT -5
A few things to consider.... 1) Pretty well all modern DACs are of the oversampling variety (oversampling is simply upsampling to an even integer multiple). The delta-sigma process also involves mathematically "trading" bit depth against signal frequency. (It's not quite as simple as saying that the DAC chip "just" upsamples the audio then does a simple conversion). This is going to happen no matter what you do, although the parameters will be automatically chosen based on the input sample rate. 2) The process of upsampling itself CANNOT create new information; it does not "improve sound quality"; what it does is to simplify other parts of the conversion process (like the reconstruction filter), so that THEY deliver better sound quality (it avoids the necessity of other compromises). However, ALL upsampling processes involve rather complicated digital filters, and every one I've ever listened to DOES change the sound - although sometimes the difference is VERY subtle. My point here is that the results WILL sound subtly different if you use different software to do the upsampling (because their filters will be different); using FooBar to upsample may well sound different than using Audition, or Windows itself,or the Asus hardware or software, or some other program ( DO NOT assume that any program upsampling 16/44.1 to 24/96 will produce an identical result). There are even some players that offer you a choice of multiple types of upsampling filters (I think HQPlayer is one of them) - and they do sound noticibly different. 3) Personally, I find that the differences are most noticeable on "metallic" sounds, like well recorded wire brush cymbals, and on how certain voice sibilants sound, and not so much (if at all) on things like vocals and drums. The differences in how the various digital filters handle transients are on the order of microseconds or milliseconds (the changes usually involve pre and/or post ringing - but, again, on the order of milliseconds at most); they're more likely to affect the "texture" of the individual wires hitting the cymbal, or the texture of a high frequency sibilant, than to alter the shape of a (relatively long and slow) decay tail on something like a drumbeat. 4) Since virtually all upsampling involves SOME filtering, which is applied at the point where the upsampling is performed, it seems reasonable to expect that the effects would be cumulative, and so it would be best to avoid multiple or repeated upsampling steps. New experiment for XDA-1. Experiments with WASAPI vs regular using short samples. The XDA-1 upconverts audio to (assumed 24/192Khz) at the DAC chip itslf to process. My question....would there be a quality benefit to upsample my CD's 24/192 before transporting to the DAC. The goal being seeing whether there is a quality difference between the upconversion at the DAC vs the upconversion done at a different section of the audio stream. A few caveats. The conversion is done in a two step process. Foobar in WASAPI mode first upconvets the signal from 16 bit to 24 bit (assumed) 44.1 Then my asus xonar essence sets the output signal to 192 Khz. So there's actually two conversions. Audition track: Beatles every little thing she does. 1:00 for about 20 seconds. 2009 stereo remaster. There are two accent drums one slightly different than the other and the latter has a different decay and that's what I concentrated on. XDA-1 at 80. Passive pre-amp turned down to make volume listenable. So direct sound: no wasapi mode @ 16/44.1 vs WASAPI 44.1 Hard to tell a difference. Felt like WASAPI mode was slightly clearer/less warm (less smeared?) WASAPI mode 44.1 vs dither 16/44.1 Dither sounded slightly clearer. Hard to tell a difference. 24/44.1 vs 24/192: Okay here it is....hard to tell a difference using really short samples in nearly any output mode. So conclusion: I didn't hear any difference. Certainly couldn't pass a blind test. Also note: my ears are slightly blocked. Not too bad, but enough for me to notice. Now the "feeling". I felt there was a difference. A very slight difference. A teeny tiny one. But I had no idea what it was. Sort of like a space between the noise. Also some past history, over time, I experienced more enjoyment with WASAPI mode. It simply felt less loose but when asked to pinpoint why, I couldn't at all. Not even an inkling. Another thing to note: The beatles release some REALLY poor recordings in terms of recording quality. Not all, but gosh there is so much audible distortion.... A well thought out reply and a good read. Thank you Keith.
|
|
KeithL
Administrator
Posts: 9,955
|
Post by KeithL on Sept 26, 2013 10:13:48 GMT -5
One more thing I should mention..... Strictly speaking, you can't "convert" 16 bit to 24 bit at a given sample rate... since that change doesn't involve modifying the actual data content. Assuming that FooBar is actually reformatting the signal, what it's doing is re-tagging the data and padding it with zeros. (It's adding 8 bits of low order zeros to "make" the 16 bit data into 24 bit data, but it isn't changing the audio content - and it shouldn't be applying any filtering or other modifications. It's sort of like "converting "10.0" to 10.00".) New experiment for XDA-1. Experiments with WASAPI vs regular using short samples. The XDA-1 upconverts audio to (assumed 24/192Khz) at the DAC chip itslf to process. My question....would there be a quality benefit to upsample my CD's 24/192 before transporting to the DAC. The goal being seeing whether there is a quality difference between the upconversion at the DAC vs the upconversion done at a different section of the audio stream. A few caveats. The conversion is done in a two step process. Foobar in WASAPI mode first upconvets the signal from 16 bit to 24 bit (assumed) 44.1 Then my asus xonar essence sets the output signal to 192 Khz. So there's actually two conversions. Audition track: Beatles every little thing she does. 1:00 for about 20 seconds. 2009 stereo remaster. There are two accent drums one slightly different than the other and the latter has a different decay and that's what I concentrated on. XDA-1 at 80. Passive pre-amp turned down to make volume listenable. So direct sound: no wasapi mode @ 16/44.1 vs WASAPI 44.1 Hard to tell a difference. Felt like WASAPI mode was slightly clearer/less warm (less smeared?) WASAPI mode 44.1 vs dither 16/44.1 Dither sounded slightly clearer. Hard to tell a difference. 24/44.1 vs 24/192: Okay here it is....hard to tell a difference using really short samples in nearly any output mode. So conclusion: I didn't hear any difference. Certainly couldn't pass a blind test. Also note: my ears are slightly blocked. Not too bad, but enough for me to notice. Now the "feeling". I felt there was a difference. A very slight difference. A teeny tiny one. But I had no idea what it was. Sort of like a space between the noise. Also some past history, over time, I experienced more enjoyment with WASAPI mode. It simply felt less loose but when asked to pinpoint why, I couldn't at all. Not even an inkling. Another thing to note: The beatles release some REALLY poor recordings in terms of recording quality. Not all, but gosh there is so much audible distortion....
|
|
|
Post by garbulky on Sept 28, 2013 6:45:43 GMT -5
A good way to audition various equipment including dacs, speakers, amps etc are with bad recordings. One's with slight flaws but still one's that's attempt to capture actual vocals/instruments. The reason is that 1. How much detail a setup can pull out of one is more obvious on mediocre recordings. 2. if a system is somewhat off, these mediocre recordings can actually sound more off. Here are some from a group that I listen to called Cimorelli. What I like about them is that their different voices all have different tones and they all have different melodies. Being able to identify whose voice it is and how natural they come off has been a good way to guage my systems.
This recording has been my go to for testing as there are a lot more different melodies, it's TOUGH to reproduce as it isn't optimally recorded. See how your system does with it. Can you follow all the different voices? There are two voices per mic. How well can the dac differentiate them without sounding awful?
|
|
|
Post by jdskycaster on Sept 28, 2013 10:07:12 GMT -5
If you can find equipment that makes that performance sound good buy it! On second thought, it might make something good sound terrible. I get where you are going though. I used to have a couple really poor recordings that were used to demonstrate how important the source is.
|
|
KeithL
Administrator
Posts: 9,955
|
Post by KeithL on Sept 28, 2013 11:22:48 GMT -5
That decision is always sort of a dilemma for me. I used to have a pair of Spica Angelus speakers (interesting speakers, with sort of a cult following). However, what was most interesting was that they weren't terribly revealing. Good recordings sounded "quite good" through them, but even lousy recordings sounded "pretty good". The bad recordings sounded much worse on my other speakers, while really good recordings sounded only slightly better. For a while I found them "quite pleasant". Finally, as I accumulated more better recordings, I sort of drifted away from liking them... But I am also sort of a fanatic about accuracy; I calibrate my computer monitor; based on the idea that I want to see what's really there, instead of "what looks good"; then I take responsibility for finding content that looks good when seen accurately. Likewise, if a recording has "warts", then they are part of it; and I don't see it as a virtue if a certain piece of equipment isn't good enough to let me see them (to me that would be sort of like leaving off your glasses at an art museum because you don't like the paintings). When auditioning equipment, I think the important thing is to be familiar with your sample music. Whether that means knowing exactly what a great disc should sound like, or what flaws you should hear in a poorly recorded one; the concept is the same. And you should also try to be careful to not overemphasize either "direction". A good system or piece of equipment should LET the great recordings sound great without artificially "improving" them; it should also LET you hear the flaws in a bad recording without overemphasizing them. (I've heard equipment that, sort of like a magnifying glass, seemed to artificially emphasize both the good parts and the flaws; and that's not good either.) I just love the tweeters in our Airmotivs and Stealths (folded ribbons); I find that they do a great job of showing what's there in perfect detail without overemphasizing it. Of course, your choice of listening material can also bias this decision. If all you listen to is poor recordings of 50's radio shows, then a system that "makes them sound better" might not be a bad idea at all (and try a pair of those Spica Angeluses if you can still find them) If you can find equipment that makes that performance sound good buy it! On second thought, it might make something good sound terrible. I get where you are going though. I used to have a couple really poor recordings that were used to demonstrate how important the source is.
|
|
|
Post by jdskycaster on Sept 28, 2013 15:22:26 GMT -5
Keith, Great points and funny to hear you mention Spica. I am very familiar with their products. In the shop I worked for in the late 80's early 90's they were always available for demo with tube gear and provided amazingly good sound for their price point.
|
|
|
Post by garbulky on Sept 29, 2013 1:30:06 GMT -5
If you can find equipment that makes that performance sound good buy it! On second thought, it might make something good sound terrible. I get where you are going though. I used to have a couple really poor recordings that were used to demonstrate how important the source is. That decision is always sort of a dilemma for me. I used to have a pair of Spica Angelus speakers (interesting speakers, with sort of a cult following). However, what was most interesting was that they weren't terribly revealing. Good recordings sounded "quite good" through them, but even lousy recordings sounded "pretty good". The bad recordings sounded much worse on my other speakers, while really good recordings sounded only slightly better. For a while I found them "quite pleasant". Finally, as I accumulated more better recordings, I sort of drifted away from liking them... But I am also sort of a fanatic about accuracy; I calibrate my computer monitor; based on the idea that I want to see what's really there, instead of "what looks good"; then I take responsibility for finding content that looks good when seen accurately. Likewise, if a recording has "warts", then they are part of it; and I don't see it as a virtue if a certain piece of equipment isn't good enough to let me see them (to me that would be sort of like leaving off your glasses at an art museum because you don't like the paintings). When auditioning equipment, I think the important thing is to be familiar with your sample music. Whether that means knowing exactly what a great disc should sound like, or what flaws you should hear in a poorly recorded one; the concept is the same. And you should also try to be careful to not overemphasize either "direction". A good system or piece of equipment should LET the great recordings sound great without artificially "improving" them; it should also LET you hear the flaws in a bad recording without overemphasizing them. (I've heard equipment that, sort of like a magnifying glass, seemed to artificially emphasize both the good parts and the flaws; and that's not good either.) I just love the tweeters in our Airmotivs and Stealths (folded ribbons); I find that they do a great job of showing what's there in perfect detail without overemphasizing it. Of course, your choice of listening material can also bias this decision. If all you listen to is poor recordings of 50's radio shows, then a system that "makes them sound better" might not be a bad idea at all (and try a pair of those Spica Angeluses if you can still find them) If you can find equipment that makes that performance sound good buy it! On second thought, it might make something good sound terrible. I get where you are going though. I used to have a couple really poor recordings that were used to demonstrate how important the source is. I do agree with you guys. But my point really isn't to make bad recordings sound good - though that is important to me to some extent. It's about how well can it balance the sound. The point is that when a bad recording has all sorts of electronics/acoustics, it puts a tougher job on the DAC to extract the good stuff - which is actually there - while still reproducing the bad stuff. For instance that recording has a thin sound to it, and some electronic issues. But with a good DAC speakers, you CAN hear the original voices and actually place them in their own space. On a great recording most DACS can do this and it gets had to differentiate but on certain cherry picked mediocre recordings, DACs can fall flat. Sort of makes the job of differentiating easier IMO. An extreme example - listening to this recording with the iphone DAC makes it very hard to tease out any details. Plus this one from them - On poor equipment makes their voices sound very tinny as they start rising up in volume during the chorus. The sound is more balanced with better DACS imo. On the XDA-1 however it does show the tinny bit that the recording does put but doesn't over emphasize the treble from what's there in the recording. A DAC that has a tip up or poor mid-range prescence will emphasize the tinnyness. Speakers with poor mid-range presecence will make the blurred bass more obvious. Also the little guitar sounds on the edge of the soundstage doesn't actually show up outside the edges of the speakers. Listen at 1:40, on the XDA-1 you are able to hear depth. You can even place the voices with the position they appear on the screen. Even with the autotune, you can also tease out the autotune vs the actual tone of their voices. This is impressive for a song that's produced to compress dynamics and with a treble bump introduced. In short, a good recording can sound great on most dacs but a poor recording doesn't sound all that great on poor DACS vs good DAC's. A caveat: a good dac will show up the flaws of the poor recording better. So, it's a mixed bag, I admit.
|
|
|
Post by repeetavx on Sept 29, 2013 1:50:09 GMT -5
Yes, this is an awful recording. It makes my ears hurt when I listen to it. I think I hear what your getting at. I hear most of the voices individually. My question is how many emergency vehicles do you hear go by at the 2:10 to 2:20 mark?
|
|
|
Post by garbulky on Sept 29, 2013 1:59:32 GMT -5
Yes, this is an awful recording. It makes my ears hurt when I listen to it. I think I hear what your getting at. I hear most of the voices individually. My question is how many emergency vehicles do you hear go by at the 2:10 to 2:20 mark? I heard that too! Hilarious!
|
|
|
Post by garbulky on Sept 30, 2013 9:24:50 GMT -5
I saw wizardofoz's signature of enjoy the music while it lasts. Immediately inspired (it doesn't take much! ), I put on Ringo Starr's Only you. I got to say I love this XDA-1! Nice pleasant bass line on that sound. Waiting on Emo's reference dac....wherefore art thou?
|
|
|
Post by garbulky on Oct 12, 2013 1:08:18 GMT -5
Gentlemen, I have in my temporary possession........an Emotiva DC-1! Call me crazy but I haven't hooked it up yet. Guess what I have planned for this weekend?!
|
|
|
Post by garbulky on Oct 12, 2013 2:06:39 GMT -5
^^^Who am kidding I hooked up the DC-1 three minutes after posting !
Okay Immediate initial impressions. Holy moley warm up required batman!
Well it's obviously very different from my XDA-1 hooked up to the passive pre-amp. It's also very different from the XDA-1 hooked up to the USP-1. Interesting considering that all three measured flat and inaudible THD and with both a low SNR....hmmmm?
It's MIDDLE mid-range is the thing that jumps out and talks about itself. It's there, It reveals detail and it's smooth compared to the XDA-1 mid-range which can't keep up to the DC-1 in that very specific middle-mid-range segment. Unfortunately it feels like nearly all the other frequencies are missing in volume.
I wrote a bunch more stuff. But then I realized. It would be unfair on this unit to write more about it so soon. This unit requires time to stay turned on till I can properly evaluate it. So stay tuned as I listen to it a bit more.
A few things I noticed. The headphone amp has a bass boost on it as default. By my initial headphone impressions, it really needs it. But...does this affect the actual analog outputs as well?
Interesting: The first song I played was "Strawberry Fields". My wife was asleep on the couch and instantly said "wow that sounds AMAZING".
Headphone listening with the Sennheisser HD600! Initial impressions: ( Remember I said INITIAL!) The sound signature is similar to what mentioned before though not the same. But honestly, this headphone amp (at least on initial listening) is not great. Audible clipping and strain is reached early. The HD600 is not a suitable load for it for critical listening. Remember this is from a cold start. I am going to be patient and give it a fair shake
IGNORE ALL OF THIS. TECHNICAL ERRORS INVALIDATE THE ABOVE.
|
|
|
Post by garbulky on Oct 12, 2013 3:52:23 GMT -5
More initial impressions updates to the DC-1: I connected it to the passive pre-amp. Now that's more like it. Something tells me this unit was designed to run in full on DAC mode at around +6 volume or at least most of it was....
|
|
|
Post by Boomzilla on Oct 12, 2013 6:05:11 GMT -5
Hey, Gar - For the sake of comparing apples to apples, could you audition (and review) sans the passive preamp? Use the Stealth as intended (combo DAC/Preamp) and tell us how it sounds once warmed up.
Thanks - Boom
|
|
|
Post by garbulky on Oct 12, 2013 6:28:24 GMT -5
I will try!
|
|
|
Post by GreenKiwi on Oct 12, 2013 12:06:42 GMT -5
Hey Gar,
I found that you can plug the xda into the DC-1s inputs. Then you can effectively compare the dac portions directly. And one can adjust the xda-1s volume slightly to match.
Looking forward to hearing what you think.
Interesting about not finding the hp amp able to drive the 600s. I didn't seem to have any problems with the 650s.
|
|
|
Post by garbulky on Oct 12, 2013 19:07:00 GMT -5
STILL INITIAL IMPRESSIONS!
Okay so some time has passed. I'm still going throuh the passive pre-amp right now as the thing doesn't have enough bass for me yet. But I will try direct soon. I already did before but the result was a little mixed. So my wife was surprisingly more emphatic than me. She normally doesn't care about this stuff but since being awaken from her sleep to say "that sounds amazing!", she did some more listening and pointed out some stuff. So far Here's what I've got. The XDA-1 appears to have a more open soundstage and treble. But the DC-1 can resolve things better. For instance in california dreamin the tambourine was audibly more resolved. You could hear the slight clink of the other bells while the XDA-1 even though it reproduced them...not as well. Also the bass simply came out through the mix.
She feels that when the DC-1 does something...it is authrotiative. Like here it is. While the XDA-1 is a more easy sound. She said when the treble comes you can point directly to the speakers as if they are being played on the speakers. This is a negative in my book as I like the speakers to dissapear but a positive in hers as it's clearer.
But we both agreed. The speakers need to be repositioned with the DC-1 as they are not in the right place.
IGNORE ABOVE: Technical errors invalidate all of this. Greenkiwi:
I have a xenos 3HA and the DC-1 headphone amp couldn't get hold its composure at higher volumes, audibly clipping, straining etc. Also it had less bass impact. But I tried them only shortly after they were plugged in. And I didn't feel like listening very long. I just tried them as I type. And well I was surprised. It sounded better! The very obvious strain and clipping - well it's not as obvious now - also I'm actually willing to listen longer! So like I said initial impressions...I guess
Something I noted is that the DC-1 gets rather hot after being left turned on for some time. This is a good thing! But it wasn't near as hot yesterday. So maybe it has to do with it getting upto operating temperature for some component. I have no idea but if I had to guess it would be the capacitors. Interesting as I didn't expect the headphone amp to improve at all - but it has. I wasn't even going to bother retesting the headphone amp but it deserved a second chance obviously. Now I'm going to try using the analog outputs of the DC-1 to the Xenos and compare. Watch out!
IGNORE ALL THE ABOVE TOO! It's crap. There is nothing wrong with the headphone output unit.
|
|
|
Post by garbulky on Oct 12, 2013 19:26:40 GMT -5
Instant update:
So I compared it with the xenos with the DC-1 driving the Xenos directly. It's hard to say off just brief listening but the xenos was more pleasant to the ear and I would give the tip of the hat to it. The voices were more obvious on the DC-1 but I felt the DC-1 was being forgiving on some of the "harsh" details of this recordings in terms of treble.
Well who went and bought the bass to the DC-1's headphone amp? It's got some! They both performed really well on this heavily compressed song. The Xenos was slightly more natural. But that could just have been the tone controls on the Xenos. At this point it's a hard one to tip the hat to either. It's a preference thing here. Both bought lots of bass.
|
|
|
Post by garbulky on Oct 13, 2013 2:06:07 GMT -5
Gentlemen, I may have to eat my words on the DC-1 - all of it. TECHNICAL ERRORS. Please ignore everything I posted about it till now.
So... INITIAL IMPRESSIONS: it sounds damn good - best I've heard. Period. It's worth it. Leave the unit turned on for two straight days at least.
|
|