Obviously I can't compare a hybrid SACD's PCM vs DSD layer because they can be mastered differently.
What I mean is compare them in a same bitrate. DSD is at least 2.8224mhz and the bitrate is 4x of Audio CD. I don't think it is useful to compare PCM formats such as 16-bit 176.4khz or 32-bit 88.2khz vs DSD@2.8224mhz because I can't even ABX 16/44 vs 16/48. A more reasonable comparison would be 16-bit 44.1khz vs DSD@705.6khz or lower.
Another question: it is possible that an "intermediate" format such as 4-bit 176.4khz could sound better than 16-bit 44.1khz if a carefully designed noise shaping/filtering algorithm is used?
Did someone make a research on such things?
If you want to compare, convert both to the highest settings your system supports and do an abx.
In general there is little advantage for playback when using higher than 44.1/16.
I did try 8-bit 88.2khz vs 16-bit 44.1khz by using Adobe Audition 1.5 and foobar2000 as my soundcard supports up to 24bit 96khz, but the problem is that I can only use the noise shaping algorithms provided by those software, I don't have the ability to design the algorithms myself so it is very difficult to make a fair comparison.
Build your own noise shaping filters! (http://www.hydrogenaudio.org/forums/index.php?showtopic=47980)
(the program is attached to post #17)
You obviously need to compare the formats in the context of what they are expected to maintain:
From the article of DSD in wikipedia (http://en.wikipedia.org/wiki/DSD-CD):
Because of the nature of sigma-delta converters, one cannot make a direct comparison between DSD and PCM. An approximation is possible, though, and would place DSD in some aspects comparable to a PCM format that has a bit depth of 20 bits and a sampling frequency of 96 kHz.[23] PCM sampled at 24 bits provides a (theoretical) additional 24 dB of dynamic range.
@lvqcl: Thanks so that I can do some experiments
@[JAZ]: Then it seems that DSD is less efficient than PCM (at least in the research mentioned in wikipedia)?
Yes, DSD is essentially dead for various reasons, one of which was how well PCM works.
***sigh***
DSD is a kind of pcm.
Not PWM or PDM?
Yes, DSD is essentially dead for various reasons, one of which was how well PCM works.
It is not really dead anymore, with recent CES I've attended this year, I saw a tons of new DACs and stuffs supporting DSD, DXD and other unicorn audiophile formats.
You can grab one that can capable of playing DSD for as low as 150 USD from Schiit Audio.
Just talking from a business view of the situation, people/companies invest badly all the time, a new or ten new products don't really mean anything. I don't know if it's dead or not but that certainly doesn't prove it's not dead. For example, did you see new 3D TVs as well?
DSD is a kind of pcm.
Please elaborate on that. It's certainly not obvious, for at least all of the online sources I searched so far differentiate between PCM and PDM.
Yes, DSD is essentially dead for various reasons, one of which was how well PCM works.
It is not really dead anymore, with recent CES I've attended this year, I saw a tons of new DACs and stuffs supporting DSD, DXD and other unicorn audiophile formats.
You can grab one that can capable of playing DSD for as low as 150 USD from Schiit Audio.
That's probably because its not much that trouble to add DSD support to a sigma delta DAC chip, and audio manufacturers like to add alphabet soup to the descriptions of their products.
Yes, DSD is essentially dead for various reasons, one of which was how well PCM works.
It is not really dead anymore, with recent CES I've attended this year, I saw a tons of new DACs and stuffs supporting DSD, DXD and other unicorn audiophile formats.
You can grab one that can capable of playing DSD for as low as 150 USD from Schiit Audio.
OT I know, but It seems to me it's not only format names which are unicorn-like, as I can't help thinking the
Schiit folks were really having a laugh/taking the wee at all audiophools when they came up with that horrid name o' theirs - even if it (probably) is their family name.
Since it's much simpler nowadays to 'throw in' DSD using inexpensive soft-/hardware ... why not toss it into a device or software-player (JRiver?) ...
Caylx (Korea?) is soon releasing a $$$ portable device -- like AK120 -- that'll do double DSD. Also the latest Sony portable headphone amp/DAC can do double DSD. Plenty of attn. at RMAF. Search YouTube for RMAF panel disc. on DSD.
Have a cookie...
DSD is a kind of pcm.
Please elaborate on that. It's certainly not obvious, for at least all of the online sources I searched so far differentiate between PCM and PDM.
The presence of a noise shaper turns it into PDM. But essentially it's just a highly oversampled PCM stream with a very low resolution of the pulse magnitude representation.
@lvqcl: Thanks so that I can do some experiments
@[JAZ]: Then it seems that DSD is less efficient than PCM (at least in the research mentioned in wikipedia)?
Because DSD has each bit providing an additive refinement, and PCM has each bit providing a geometric refinement, DSD must be less efficient in some senses.
Were you to add a bit to the DAC resolution, that would be an increase in efficiency vs. doubling the bit rate of temporal refinements.
One could, (and some have) as well added noise shaping (also called A*PCM in Jayant and Noll) or "Sigma-Delta" or "Delta-Sigma" depending on author to standard pCM and have noise-shaped PCM.
But the dacs in use nowdays are all 1-bit to 4-bit dacs using noise shaping and massive oversampling in order to give the proper resolution at low frequencies.
You can see a slide deck at www.aes.org/sections/pnw/ppt.htm on conversion that will help explain this.
You can see a slide deck at www.aes.org/sections/pnw/ppt.htm on conversion that will help explain this.
Thanks for that link - I especially enjoyed the Hayser Lecture presentation.
The final point is that the formats are more or less equivalent. Which means you shouldn't pay through the nose for DSD material.
However, I'd say that excellent R2R DACs (say, old AKM or TI PCM series) are better than some otherwise excellent sigma-delta DACs (such as Cirrus Logic, Wolfson or ESS; sigma-delta AKM is indistinguishable to me from R2R). The difference is minor, but there and ABXable.
PCM is the base format for an R2R DAC, DSD is the base format for a sigma-delta DAC.
So, not all DACs are sigma-delta. Heck, for cost-no-object purposes you can make an R2R DAC at home, given precision resistors.
Most DACs which support both, handle them equally well - but there are few exceptions - mostly using ESS chips. Again, the difference is minor.
(On an unrelated note: ESS Sabre blows, it's tonally off somehow.)
However, I'd say that excellent R2R DACs (say, old AKM or TI PCM series) are better than some otherwise excellent sigma-delta DACs (such as Cirrus Logic, Wolfson or ESS; sigma-delta AKM is indistinguishable to me from R2R). The difference is minor, but there and ABXable.
That's interesting. Can you give details about the ABX test ? From all I've read on HA (and elsewhere) these differences are very small and presumed below audibility thresholds.
In fact, I didn't. I only did a single-blind test there, as I had contact with the helper during the test. (but still didn't see what was connected where)
The test was Presonus FP10 (older AKM) against Lynx Hilo (CS), both used as DAC, line out to Eddie Current Super 7 amplifier, Hifiman HE-500 (modded) - switchbox in my hands, free switching until a decision. Chiptune samples.
That got 11/12 tries.
Generally I was unable to properly rate either, both sound extremely good in DAC capability, but different in timbre. FP10 sounding "thicker" while Lynx sounding "lighter", both very detailed but I'd say Lynx getting the upper hand there. Noise floor difference was masked by EC S7 noise floor, which is slightly audible (-90 dB or so) - but few amplifiers do better than -108 dB anyway which is the FP10s DAC limit. (Tested via loopback into Lynx, which is superior there.)
I've checked whether frequency response matches, it was linear on both. Volume was matched using a voltmeter on 1kHz sine to within 0.5% - the knobs on FP10 didn't allow a better match.
The "better" part is kinda in air quotes in the above post. The only clearly audibly inferior DAC was the Anedio D2 loaner based on ESS Sabre 9018 - but I didn't bother to blind test it. I haven't heard such grainy treble in a while - something must be messed up in the filter.
Considering all the reviews, I bet Sabre 9018 is just a terrible chip, even in the best of implementations. Unfortunately many DACs use it.
^Since the tube headphone amp mentioned has only one input, I take it your A/B switchbox was mounted before it in the chain and your helper manually reconnecting it, each trial, was your "randomizer". Did you have to hand this switchbox back to your assistant (helper) each time [or was what you held in your hands simply a remote control of it] and what brand/model was the switchbox, or do you have more details on it if was homemade. Thanks.
In fact, I didn't. I only did a single-blind test there, as I had contact with the helper during the test. (but still didn't see what was connected where)
The test was Presonus FP10 (older AKM) against Lynx Hilo (CS), both used as DAC, line out to Eddie Current Super 7 amplifier, Hifiman HE-500 (modded) - switchbox in my hands, free switching until a decision. Chiptune samples.
That got 11/12 tries.
Generally I was unable to properly rate either, both sound extremely good in DAC capability, but different in timbre. FP10 sounding "thicker" while Lynx sounding "lighter", both very detailed but I'd say Lynx getting the upper hand there. Noise floor difference was masked by EC S7 noise floor, which is slightly audible (-90 dB or so) - but few amplifiers do better than -108 dB anyway which is the FP10s DAC limit. (Tested via loopback into Lynx, which is superior there.)
I've checked whether frequency response matches, it was linear on both. Volume was matched using a voltmeter on 1kHz sine to within 0.5% - the knobs on FP10 didn't allow a better match.
The "better" part is kinda in air quotes in the above post. The only clearly audibly inferior DAC was the Anedio D2 loaner based on ESS Sabre 9018 - but I didn't bother to blind test it. I haven't heard such grainy treble in a while - something must be messed up in the filter.
Considering all the reviews, I bet Sabre 9018 is just a terrible chip, even in the best of implementations. Unfortunately many DACs use it.
Not good enough, sorry.
Shouldn't the tube amp adding more own distortion by design as any modern DAc should?
How can someone rate the source at all with a tube amp behind, leave alone assigning it to PCM, DSD or filtering?
Did i simply get the test setup wrong?
Shouldn't the tube amp adding more own distortion by design as any modern DAc should?
Just because its made with tubes doesn't mean that it is necessarily audibly flawed.
But yes it is hard to make a tubed buffer that doesn't add measurable noise and distortion to the output of a good DAC.
But on balance there's the slight matter of the source at one end, and the transducer at the other and they are generally the weakest link.
Finally, the well known limitations of human hearing at its best which it often is not, is generally the ultimate limiting factor.
How can someone rate the source at all with a tube amp behind, leave alone assigning it to PCM, DSD or filtering?
Some call it naivete, some call it arrogance, and some call it wishful thinking. ;-)
Did i simply get the test setup wrong?
Probably you were too kind to it! ;-)
Homemade passive switchbox with a good toggle button and connectors. No monkey business there, low capacitance.
Switchbox is after the DACs - it's not a coax spdif switchbox.
The amp has low harmonic distortion, espectially higher up - but that isn't an issue with DACs anyway.
What might be problematic for DACs is the relatively low input impedance, on the order of 50 kOhm.
--
Yeah, I'd love to blind test everything. Pity it takes a bunch of hours. Note that every switcheroo took me a bunch of minutes to discern and be sure. It's that slight a difference.
Plus the switcher is only two port. I also have not tested any other inputs than coax SPDIF.
Did you have to hand this switchbox back to your assistant (helper) each time...
I'm still looking for an answer to this part of my question. Thanks.
Did you have to hand this switchbox back to your assistant (helper) each time...
I'm still looking for an answer to this part of my question. Thanks.
No, the coax cables were long enough. I didn't even listen while they were being reconnected. The devices were in another room. The DACs were given a bunch of minutes (~2 minutes at least, likely more, I didn't measure) to sync properly. Cables were ~3m long. (But cable length doesn't worsen jitter and is neglible for the connection to the amp at audio frequencies.)
Thanks.
Another question: What was the channel tracking error value of the dual-ganged analog potentiometer used to level match the Presonus FP10 to the other DAC?
Thanks.
Another question: What was the channel tracking error value of the dual-ganged analog potentiometer used to level match the Presonus FP10 to the other DAC?
The actual pot balance is irrelevant - it is a digital volume control actually. At the position involved the outputs were balanced properly in both devices according to the voltmeter. (Which is 0.05% error at the range. I could use better hardware.)
Both devices were ran with "standard" -6 dB and not 0 dB. (I like my digital volume controls done in software, but this test used a better SPDIF interface which does not have a volume control. (loaner Audiophileo 2) I wanted to exclude USB/Firewire handling from influencing this test, as well as reduce the effect of SPDIF jitter if any.)
I've never heard of a digital rotary volume knob which when noisy or scratchy can be remedied by a squirt of contact cleaner, as this FP10 owner attests fixed his volume problem:
http://forum.recordingreview.com/f9/scratc...nus-fp10-42612/ (http://forum.recordingreview.com/f9/scratchy-pot-presonus-fp10-42612/)
Also a different reviewer (Chris from DE, dated 01/16/13) from another site complains of his newer Presonus unit, "The only problem I have is sometimes the main volume pot gets noisy when turning", here:
http://www.musiciansfriend.com/pro-audio/p...rding-interface (http://www.musiciansfriend.com/pro-audio/presonus-firestudio-project-recording-interface)
but his solution is the old rotate the pot back and forth to scrape the contacts clean method: "...after a few rolls, it is clean".
However, I'd say that excellent R2R DACs (say, old AKM or TI PCM series) are better than some otherwise excellent sigma-delta DACs (such as Cirrus Logic, Wolfson or ESS; sigma-delta AKM is indistinguishable to me from R2R). The difference is minor, but there and ABXable.
PCM is the base format for an R2R DAC, DSD is the base format for a sigma-delta DAC.
So, not all DACs are sigma-delta. Heck, for cost-no-object purposes you can make an R2R DAC at home, given precision resistors.
Most DACs which support both, handle them equally well - but there are few exceptions - mostly using ESS chips. Again, the difference is minor.
(On an unrelated note: ESS Sabre blows, it's tonally off somehow.)
Excuse me but even if this ABX experiment has some kind of logic i really have a problem with nearly all stated conclusions here and how you prove that.
This is why we have a TOS that tries to prevent some talk about tender feelings being attributed to different kind of gear.
I do know about the TOS. Talking about "tender feelings" when I cannot name why it sounds different seems fine to me.
That's why I called it something nebulous like "timbre". If I was able to say "one of those sounds like it has less jitter" or "one of these has overshoot, the other does not" then I would, but I do not know what actually causes the audible difference.
I've excluded a few obvious items such as frequency response and noise floor (covered by amp).
It is possible that the 0.5% difference in loudness is actually audible - both of the devices were around 40 sons, since that's how loud I listen - think radio broadcast at decent not too loud volume.
Both use high line out levels in consumer range, so there is likely an amplifier in the way.
Note that in case of FP10, said amplifier is different than the one in mic/line ins apparently. (Those are not linear in FR!)
It is also possible that chiptunes cause an audible difference in IMD - they have lots of high frequency contents and square waves.
mzil: I used the back line outs in FP10 - they still have a volume control, but it is digital. Fancy design, same pot controls both digital and analog volume depending on the output in question. I know specifically because the low levels do not cause channel imbalance like with phones out.
The tracking of the pot is important when you're controlling the mic/line ins, main out or phones out. (And the tracking capability is so-so, but in this range it is good - near the center of the pot for -6 dB.)
Lynx Hilo has a digital control only on line outs as well.
--
Next time I'll do a double-blind trial, but without the loaner converter... But not any time soon.
On the topic, I've quick tested DSD64 vs PCM (24bit, 44.1k) on my Lynx Hilo. Failed to notice any difference in ABX at 16 long trials. This is "reportedly" not true with some chipsets, specifically said ESS Sabre 9018.
Volume was matched using a voltmeter on 1kHz sine to within 0.5% - the knobs on FP10 didn't allow a better match.
Using the FP10's knobs to control the level matching, as you wrote in that post, means you were using an
analog potentiometer. This applies not just to the MAIN OUT knob (and the PHONES OUT knob); the MIX knob also works in the analog domain.
From Presonus FP10 User Manual: (http://www.presonus.com/uploads/products/1714/downloads/FP10_Owners_Manual_EN.pdf)
"Mix. The mix knob is like a balance control. It balances the headphone MAIN and CUE outputs between
Inputs 1-8 and playback outputs 1 and 2. This feature is an analog hardware mixer..."
Hmm. Regardless, I have checked and the tracking was way better than volume matching.
Do you recommend I should run a test with DACs on a bench connected to a linear power supply, driven by I2S off an ARM microcomputer?
This could be arranged, but I have much more important things to do for a long time.
mzil: I used the back line outs in FP10 - they still have a volume control, but it is digital. Fancy design, same pot controls both digital and analog volume depending on the output in question. I know specifically because the low levels do not cause channel imbalance like with phones out.
That doesn't prove anything. They are separate knobs and all your observation supports is that the PHONES OUT analog pot happens to seem inferior in regards to low level channel tracking, at least on your sample unit. Another person with the same unit might happen to get a better headphone pot, or worse.
---
I personally feel outboard DAC testing is almost worthless. Even in nearly ideal, double-blind, well controlled conditions, people inevitably hear differences [if they do at all] due to the various electronics, including the op amps used in these units,
not the actual DAC chips themselves. Small impedance mis-matches from the output device to the switcher/preamp causing subtle frequency responses errors and channel tracking error causing small lateral shifts in the soundstage also come into play at times, even if the listeners aren't conscious of it and attribute the changes to other things.
Oh, but what if the difference is in the chip itself too? I'm not sure anyone has tried it.
I'd also like to test opamps double blind some time later - both as a buffer and with gain 4, at +/- 12V from a linear power supply.
The same breadboard setup could do both. The tricky part is randomly switching between the devices - I'll need a good passive crosspoint switch (or a multitarget DPDT) for that.
Hmm, actually, cable-or could work too, as opamps have a huge output impedance anyway.
Not sure if the same applies to cable-or of DACs - they have low enough impedance for that to possibly matter - but it is a constant condition.
If you think an impedance mismatch of way less than 0.1 Ohm is audible (also picofarad order of capacitance mismatch), I'd say you're a proponent of cable audibility and you really should try to ABX that. As I said, I've actually measured the voltages to within 0.5%. Including the switcher in all 4 configurations of cables.
I personally feel outboard DAC testing is almost worthless. Even in nearly ideal, double-blind, well controlled conditions, people inevitably hear differences [if they do at all] due to the various electronics, including the op amps used in these units, not the actual DAC chips themselves. Small impedance mis-matches from the output device to the switcher/preamp causing subtle frequency responses errors and channel tracking error causing small lateral shifts in the soundstage also come into play at times, even if the listeners aren't conscious of it and attribute the changes to other things.
Anybody who thinks good op amps have sonic signatures needs to ABX about 20 of them cascaded - and get thoroughly humiliated.
ADCs and DACs are among the most perfected of all audio components, but op amps can be even better on the test bench.
Preserving +/- 0.1 dB matching across an entire audio system is doable - generally only good level matching is required.
As far as the metering goes - +/- 0.1 dB is about the same as +/- 1%. Flat response in the meter is not an absolute requirement because the goal is matching, not necessarily flat response.
The Uni-T UT61E 5 digit DVM is a reasonably priced tool for this job, and it is very flat over the audio band and beyond.
I wouldn't be so sure about "inaudible" - IMD behavior differences could in theory be audible, especially on complex signals. (Products on the order of -40 dB exist.)
It does not cascade trivially either. The difference, if any, would be very subtle though.
Remember what you're asking for though: 10 high end audio opamps costs some $100 and you'd need more than one brand... You're asking someone to toss $400 + more in precision resistors and capacitors.
Loopback testing would necessarily involve additional DAC/ADCs which could make the test invalid. I happen to have a bunch of spare ADA4898-1s, but that makes only for 5 pairs.
I also have a slightly cheaper Uni-T meter, 3,5 digits.
I wouldn't be so sure about "inaudible" - IMD behavior differences could in theory be audible, especially on complex signals. (Products on the order of -40 dB exist.)
It does not cascade trivially either. The difference, if any, would be very subtle though.
Spurious responses on the order of -40 dB in the normal audible range from a good op amP? Please provide a real world example.
Remember what you're asking for though: 10 high end audio opamps costs some $100
I said nothing about high end op amps. I said good op amps. The tests I was party to involved TL074s.
and you'd need more than one brand...
No, this is not an attempt to prove a global truth. Besides, if 20 TL074 sections are transparent, why would you need more expensive chips?
You're asking someone to toss $400 + more in precision resistors and capacitors.
Test was done with standard grade generic parts, which are actually quite sonically transparent. The test can be done with only resistors in the signal path, and good 1% parts can now be had for pennies each.
Loopback testing would necessarily involve additional DAC/ADCs which could make the test invalid.
This is a discussion about op amps only. But converters are not that much harder. Most of my tests involved converters in pro grade audio interfaces.
I happen to have a bunch of spare ADA4898-1s, but that makes only for 5 pairs.
I also have a slightly cheaper Uni-T meter, 3,5 digits.
I see much more complexity than is actually necessary. If one actually does good ABX test one tends to find that the latest-greatest higher end stuff is unecessary.
I personally feel outboard DAC testing is almost worthless. Even in nearly ideal, double-blind, well controlled conditions, people inevitably hear differences [if they do at all] due to the various electronics, including the op amps used in these units, not the actual DAC chips themselves. Small impedance mis-matches from the output device to the switcher/preamp causing subtle frequency responses errors and channel tracking error causing small lateral shifts in the soundstage also come into play at times, even if the listeners aren't conscious of it and attribute the changes to other things.
Anybody who thinks good op amps have sonic signatures...
That wasn't my point, at all.
I wouldn't be so sure about "inaudible" - IMD behavior differences could in theory be audible, especially on complex signals. (Products on the order of -40 dB exist.)
It does not cascade trivially either. The difference, if any, would be very subtle though.
Spurious responses on the order of -40 dB in the normal audible range from a good op amP? Please provide a real world example.
Ok, I'll provide a graph, it's not a standard IMD test, uses more tones. (a large chord)
Standard two tone CCIF IMD can be easily -60 dB on good parts. In a great implementation, -80 dB.
That with expensive AD8610s. TL074 is way worse at it.
The resistors don't need to be 0.1%, yes, but they have to be paired. Plus you need a good listen with hard tracks. (again, chiptunes, noise and metal are best IMO)
Anyway, this is off-topic. Where's the on-topic thread?
I wouldn't be so sure about "inaudible" - IMD behavior differences could in theory be audible, especially on complex signals. (Products on the order of -40 dB exist.)
It does not cascade trivially either. The difference, if any, would be very subtle though.
Spurious responses on the order of -40 dB in the normal audible range from a good op amP? Please provide a real world example.
Ok, I'll provide a graph, it's not a standard IMD test, uses more tones. (a large chord)
Standard two tone CCIF IMD can be easily -60 dB on good parts. In a great implementation, -80 dB.
That with expensive AD8610s. TL074 is way worse at it.
The resistors don't need to be 0.1%, yes, but they have to be paired. Plus you need a good listen with hard tracks. (again, chiptunes, noise and metal are best IMO)
Anyway, this is off-topic. Where's the on-topic thread?
How about this for a multitone?
(https://dl.dropboxusercontent.com/u/36129341/multitone%20delta%201010LT%20r-p.png)
This tone's peak amplitude is -1 dB FS It was looped from the output to the input of a popular-priced M-Audio 1010LT 8 channel in, 8 channel out audio interface which uses <$0.50 NE5532 op amp chips for buffers.
http://www.jameco.com/webapp/wcs/stores/se...CFUhp7AodxmMAUA (http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&pa=27385&productId=27385&keyCode=WSF&CID=GOOG&gclid=CKba5-fazrwCFUhp7AodxmMAUA)
All spurious responses appear to be > 115 dB down, which assures total inaudibility.
Note this is not a high end or premium audio interface and is built with standard commercial grade parts. There are at least 2 op amp buffers, a DAC and an mid-grade ADC in the loop.
http://www.m-audio.com/images/global/manua...10LT-Manual.pdf (http://www.m-audio.com/images/global/manuals/Delta1010LT-Manual.pdf)
Arnold you miss the point. You have to do a good listen on hard tracks!
Arnold you miss the point. You have to do a good listen on hard tracks!
If you put a smiley on that I'd be laughing, but in the absence of any evidence of attempted humor I must deal with the above as a criticism.
I suggest that you reread the post I was responding to. It made a specific technical claim involving a measurement which I addressed.
There is also a bit of background to this. In another thread I strongly supported the idea that at this point in the current SOTA we must use listening tests to evaluate very complex audio tools such as perceptual coders.
This is not true of simple electronic devices such as op amps, power amps, ADCs, DACs and even loudspeakers. Since they perform relatively simple forms of signal processing and other than loudspeakers generally perform orders of magnitude better than our hearing thresholds, relatively simple technical tests are very applicable. This fact is represented in widely accepted recommendations for technical gear used for doing critical listening tests such as BS 1116-1. Please see section 7 "Reproduction devices".
If you put a smiley on that I'd be laughing, but in the absence of any evidence of attempted humor I must deal with the above as a criticism.
I thought it is that obvious i don't need a smiley.
AstralStorm does some statements here all the time that sound exactly like the audiophile nonsense posted elsewhere and when asking for evidence the next statement follows again with no solid evidence. "you need a good listen with hard tracks" for example. What exactly does that mean?
I wonder how someone with more than 700 posts over here doesn't realize that Hydrogenaudio is a bit different (luckily)
Sorry Arnold if i made you upset in a way!
Yeah yeah. Blame the messenger. Please continue disregarding anything I'll say regardless of how ridiculously precise I have to be.
There *is* a difference. I don't know what causes it. It is inaudible on simple tracks (IMO), it is audible on hard tracks (single-blind test). Manifests to my ears as difference in timbre (read the definition of the word, I'm using the common one).
To be more exact, the difference sounded like extra "fatness" in the sound, as if everything was bassier or bass was "later". (Except there was no FR or phase response difference large enough to explain it.)
Arnold: Your graph is worthless, because you used (mostly) harmonically related frequencies. That doesn't work for IMD testing. Please check how CCIF IMD is done and how SMPTE IMD is done. The frequencies used in those do not have simple common divisors.
(E.g. CCIF: 19/20. SMPTE: 30/7.) Essentially your test is CCIF times 10. It is pretty hard, but only when tried in top octave. At low frequencies such simple ratios are generally handled well and SMPTE is harder. Then, there are even harder IMD tests with "weirder" ratios.
The problem with such test is that the spurious frequencies will reinforce and cancel each other very predictably since every pair of the frequencies is related with a relatively similar ratios or a close multiple of it.
Plus they're not 115dB down, subtract 30 dB... and you've made it easier on the opamp by reducing volume by that many dB. (I assumed those are dBV, not dBFS. If they're dBFS, please provide voltages.)
Sorry I haven't posted mine yet, various reasons which are not interesting to you I bet. I'll get on it next month. (and unfortunately no earlier)
The said chord is a really "wolf" chord, highly atonal, 5 tones. I'll try to find the file in the library, hopefully I won't have to recreate it.
Yeah yeah. Blame the messenger. Please continue disregarding anything I'll say regardless of how ridiculously precise I have to be.
This is kind of a silly thing to say given that you still haven't posted anything of substance.
Arnold: Your graph is worthless, because you used harmonically related frequencies.
800 and 900 Hz are harmonically related? No, no they're not.
His test contains both harmonic and non-harmonic frequencies, and so it can test for both THD and IMD.
That doesn't work for IMD testing.
You're mistaken. 800+900 = 1700Hz
From his graph you can see that IMD is less than -129 dB. Clearly this is negligible.
----
FWIW, THD is just a degenerate case of IMD. If THD is 6 orders of magnitude below relevance, probably IMD is negligible too!
It might be, but again, 8/9 is a relatively simple ratio and IMD is reduced at low frequencies. Try that with 863 and 917 Hz for example... Preferably mutually prime numbers. You should get a decent hash, but still not high (I'd expect -80 dB or better from a good implementation at low frequencies like these. At 2V RMS. This is usually decently well masked in real music. In higher frequencies, the masking is not as important and the amplifiers do worse.)
There are a few low prime ratios in there. (Note that full scale is -30 dB (which unit?). This is less than -95 dB best case to be nitpicky. Actually since when one can record -129 dB anyway (that's below thermal noise floor unless > 3x averaged), this graph is all kinds of fishy to me.
19/20 kHz CCIF IMD is invisible there and that does not mesh with measurements elsewhere.
Also the M-Audio 1010LT is actually excellent at IMD (at least SMPTE, I couldn't find CCIF but should be also very good) and not representative of most devices. It is better than Benchmark DAC1 and 2 for instance. Note that nowhere I said you need exotic components. NE5532's main weakness is low PSRR (esp. vulnerable to ground modulation) and I bet the M-Audio engineers thought very hard on solving that issue. (It is one of the top audio opamps, though due to this problem not the easiest to use.) CMRR and other variables are on par with top current opamps as long as NE5532 is reproduced in modern processes and with decent QC.
That said, it might not be IMD being responsible for what I hear, but something entirely different. IMD seems the only likely candidate of what we know about. In fact, it'd be great if a mod would split the thread.
It might be, but again, 8/9 is a relatively simple ratio
This is irrelevant. Each of those tones has a finite bandwidth because of the transform limit. Say each one has a width of 0.5Hz. They would also include ratios such as 800.1/899.9 = 8001/8999. This is not a simple ratio. As you can see, it makes no difference.
Preferably mutually prime numbers.
8999 is a prime number.
You should get a decent hash, but still not high
Indeed, <-129 dB is not very high.
Arnold: Your graph is worthless, because you used (mostly) harmonically related frequencies. That doesn't work for IMD testing.
Anybody who can accurately read the frequency scale of that chart knows that the above stattement is incorrect.
The frequencies were arbitrarily chosen - 20, 25, 30, 40, 50, 60, 70, 80. 90. 100, 150, 200, 250, 300 and so on. The last octave is 10 KHz, 11 KHz, 12 KHz, 13 KHz. 15 KHz etc. a few are harmonic, most are not.
Please cite an authoritative independent source that confirms that exceptional claim. I apologize for sending you on a wild goose chase because the physical law that you cite simply doesn't exist.
Please check how CCIF IMD is done and how SMPTE IMD is done. The frequencies used in those do not have simple common divisors.
(E.g. CCIF: 19/20. SMPTE: 30/7.) Essentially your test is CCIF times 10. It is pretty hard, but only when tried in top octave. At low frequencies such simple ratios are generally handled well and SMPTE is harder. Then,
If you could cite a credible authority, I would rerun the test with those test tones. I know for sure that the physical law you claim simply does not exist because I've run a wide variety of 2 tone, 3 tone and multitone tests on a wide variety of equipment types and models. I've also done the math.
Here is an example of Agilent, a highly respected test equipment manufacturer recommending the use of harmonic frequencies in their technical literature:
http://cp.literature.agilent.com/litweb/pdf/5991-2396EN.pdf (http://cp.literature.agilent.com/litweb/pdf/5991-2396EN.pdf) (Figure 1)
One reasonable guide for picking multitone frequencies is to pick frequencies that line up with the FFT bins so that a multitone with flat frequency response is more likely to present itself that way when analyzed by a FFT.
Please check that graph again. The frequencies in that particular multitone are arbitrarily chosen - 15, 20, 25, 30, 40, 50, 60, 70, 80. 90. 100, 150, 200, 250, 300 and so on. The last octave is 10 KHz, 11 KHz, 12 KHz, 13 KHz. etc. a few happen to be harmonic, but most are not.
Multitones of any kind are easy enough to create - another possible sequence of frequencies are the standard 1/3 octave frequencies. Doesn't matter. Another possible sequence involves the frequencies that divide into the sample rate with integer quotients and no remainders. Audio Precision recommends a logrithmic sequence. For IM testing it doesn't matter that much.
The ratio of the test frequencies does not matter within reason. The spurious responses created depend on the orders present in the nonlinearity, but they are always created. If you use harmonic tones some spurious responses due to harmonic distortion land on top of spurious responses due to IM distortion, but there are plenty of circumstances where that does not happen.
http://en.wikipedia.org/wiki/Intermodulation (http://en.wikipedia.org/wiki/Intermodulation)
http://www.prismsound.com/test_measure/sup.../multi-tone.php (http://www.prismsound.com/test_measure/support_subs/apps/multi-tone.php)
http://www.klippel.de/measurements/nonline...distortion.html (http://www.klippel.de/measurements/nonlinear-distortion/multi-tone-distortion.html) (extra good references cited)
http://www.ap.com/kb/show/60 (http://www.ap.com/kb/show/60)
Higher frequencies tend to create more IM products because amplifiers generally have less open loop gain at high frequencies so there is less corrective inverse feedback that is available to linearize them. This is especially true of unity-gain compensated operational amplifiers, but is also generally true.
Arnold: Your graph is worthless, because you used (mostly) harmonically related frequencies.
So, if I use a multitone that consists of lines at Fo(2n+1) where Fo is a base frequency, oh, say 200Hz, n is an integer 0 to the result reaching fs/2, I can't see any IMD in that?
You sure about that?
I'm rather sure you're not so sure.
Yeah yeah. Blame the messenger. Please continue disregarding anything I'll say regardless of how ridiculously precise I have to be.
Arnold: Your graph is worthless, because you used (mostly) harmonically related frequencies. That doesn't work for IMD testing.
I find it ironic that complaints about "Blame the messenger" come in a post that also name-calls a well-prepared piece of evidence as "worthless", and provides no reliable evidence to support that exceptional claim.
Arnold: Your graph is worthless, because you used (mostly) harmonically related frequencies.
So, if I use a multitone that consists of lines at Fo(2n+1) where Fo is a base frequency, oh, say 200Hz, n is an integer 0 to the result reaching fs/2, I can't see any IMD in that?
You sure about that?
I'm rather sure you're not so sure.
Perhaps our correspondent does not realize that legacy IM tests such as SMPTE and CCIF were designed to be easy to implement with the relatively primitive test equipment of the day (SMPTE 1941, and CCIF/ITU 1929).
In addition the SMPTE test is contrived to work well with a relatively dirty medium - movie film optical sound.
According to http://www.rane.com/note145.html (http://www.rane.com/note145.html)
"Required Conditions. Many variations exist for this test (CCIF). Therefore, the manufacturer needs to clearly spell out the two frequencies used, and their level. The ratio is understood to be 1:1."
IOW the frequencies commonly used for this test (19 KHz & 20 KHz)are not sacrosanct.
"Required Conditions. Many variations exist for this test (CCIF). Therefore, the manufacturer needs to clearly spell out the two frequencies used, and their level. The ratio is understood to be 1:1."
IOW the frequencies commonly used for this test (19 KHz & 20 KHz)are not sacrosanct.
And, of course, with modern FFT analysis one does not, and should not, limit one's self to 2 frequencies.
"Required Conditions. Many variations exist for this test (CCIF). Therefore, the manufacturer needs to clearly spell out the two frequencies used, and their level. The ratio is understood to be 1:1."
IOW the frequencies commonly used for this test (19 KHz & 20 KHz)are not sacrosanct.
And, of course, with modern FFT analysis one does not, and should not, limit one's self to 2 frequencies.
Agreed. As Audio Precision points out a well-designed multitone can affirm excellent frequency response, low nonlinear distortion, and low noise with just one measurement.
http://www.ap.com/kb/show/60 (http://www.ap.com/kb/show/60)
The main problem is that you used so many tones, you're obscuring actual intermodulation products. The FFT window used is not selective enough, should use a rapid falloff flat top window, not Hamming.
Also you should perhaps try a less state of the art device. Specifically one that has stated SMPTE IMD at -110 dBV @ 1V RMS, expected CCIF IMD at -90 to -100 dBV @ 1V.
(Of course I'd love to know how they managed to pull it off - must be really careful board layout and power filtering.)
See some sample measurements (the nwavguy's are better):
http://nwavguy.blogspot.com/2012/04/odac-released.html (http://nwavguy.blogspot.com/2012/04/odac-released.html)
http://scientistsaudio.blogspot.com/2012/0...kii-part-1.html (http://scientistsaudio.blogspot.com/2012/06/leckerton-uha-6smkii-part-1.html)
That is -106 dBV @ 1V RMS for ODAC, -100 dBV @ 1V RMS for Leckerton from another source I could find. Despite that, it is discernible in a test from Leckerton UHA-6s mkII when used as DAC, same amplifier (FP10). I've done that test way earlier, also single blind, USB source from the same machine. (8/8) So it's likely not IMD. THD dominates IMD in these two devices, but is still very low at -90-something dB.
Could be power supply filtering? But how?
Could be the channel imbalance in that test? But not in this last one...
--
We still need to run listening tests - we do not know *everything* about electronic sound devices yet. Of course, first run the known electrical tests to ensure the devices are not broken.
This includes: jitter (for DACs), IMD, THD, noise, clipping, aliasing and frequency response.
In this test, I have not excluded any of the above except noise and FR, though all are assumed good. I do not have a good enough hardware to measure such low distortion. (Scratch that, I do have it since very recently. Will try to do that next month.)
--
The idea is to explain the difference. Might be the single blind trial process, who knows? I'll have to run double-blind, this is hard with physical devices.
Or maybe I got really lucky and hit 11/12 by chance.
The main problem is that you used so many tones, you're obscuring actual intermodulation products. The FFT window used is not selective enough, should use a rapid falloff flat top window, not Hamming.
It appears that some take me for an naive person that would not check such things out before basing a web site (www.pcavtech.com) or even a post on an important conference such as HA on it. IMO it is similarly unwise act like reliable authorities such as Audio Precision would make the same mistake.
Here is the pure multitone:
(https://dl.dropboxusercontent.com/u/36129341/multitone%20pure.png)
Here is the multitone with some nonlinear distortion added:
(https://dl.dropboxusercontent.com/u/36129341/multitone%20impure.png)
Obviously, there are more than enough actual intermodulation products to be seen and analyzed. While the FFT software I am currently using does not support flat topped windowing I am familiar with such things, and their major benefit is that the tops of the peaks of the multitone spectra would be more even given that the frequency choices were arbitrary. If one chooses frequencies that correspond exactly to FFT bucket center frequencies this minor situation goes away with any reasonable windowing technique. Of course one has to have some basic understanding of windowing to anticipate this.
That FFT argument is pure WTF. The choice of windows determines how precisely you can resolve two different frequencies. If, as we have established, no IMD frequency exists at all the FFT window makes no damn difference.
Its not even wrong, it makes me wonder if you understand what you're even trying to talk about at all.
That FFT argument is pure WTF.
For those of us who don't understand all of this, whose FFT argument are you referring to, AstralStorm's or Arnie's?
That FFT argument is pure WTF.
For those of us who don't understand all of this, whose FFT argument are you referring to, AstralStorm's or Arnie's?
Sorry, I mean AS's argument that the choice of window function would somehow create energy at frequencies where there is none currently. Changing the window might turn one peak into 2, or two into 1, but it still has to conserve energy. It can't pull it from thin air.
Sorry, I mean AS's argument that the choice of window function would somehow create energy at frequencies where there is none currently. Changing the window might turn one peak into 2, or two into 1, but it still has to conserve energy. It can't pull it from thin air.
Thanks - that explanation makes sense to me.
That FFT argument is pure WTF. The choice of windows determines how precisely you can resolve two different frequencies. If, as we have established, no IMD frequency exists at all the FFT window makes no damn difference.
Its not even wrong, it makes me wonder if you understand what you're even trying to talk about at all.
Methinks that many people are impressed with NWAVGUY's fine site, and some without understanding the background think that everybody should do things his way or be in error.
It's either that FFT thing or you're facing relatively high jitter in the signal source, therefore the high sidebands after integration.
The whole idea is to have as close to a specific frequency and not have the frequency move around, as the IMD products will also move and reduced by time averaging, biasing the result down.
(Much like spread spectrum is used to reduce noise peaks.) But then, these products, if any, would also get wider.
The window function does not create energy, but the sidebands do obscure any low energy in the signal.
The right window for this, again, is a flat top window with high order sidelobe drop (there do exist 7th order ones) and decent first sidelobe level. There's a tradeoff involved between main lobe width and first sidelobe volume of course.
--
I have another theory: the tube amplifier introduces distortion that interacts with the low level distortion present in the amplifier output, making it more audible and thus discernable.
Plausible?
You're mistaken. 800+900 = 1700Hz
From his graph you can see that IMD is less than -129 dB. Clearly this is negligible.
The whole idea is to have as close to a specific frequency and not have the frequency move around, as the IMD products will also move and reduced by time averaging, biasing the result down.
(Much like spread spectrum is used to reduce noise peaks.) But then, these products, if any, would also get wider.
How many bins would you have to average to get audible IMD down to -129dB? 100,000? 1,000,000? More?
Seriously, think before you post.
How many bins would you have to average to get audible IMD down to -129dB? 100,000? 1,000,000? More?
Seriously, think before you post.
A bit unstable signal generator clock. (SNR of 1010LT D/A is around 102 dB, so you have many times time averaging already.)
2x averaging drops noise floor by 1-3 dB depending on how white the noise is, drops IMD with a stable clock less. With unstable clock, you can treat IMD as a colored noise.
To go from 102 dB to 128 dB, you need few hundred times averaging, maybe even 1000x. With an unstable signal source that would drop IMD a lot.
(Also, what is with -30 dB up top? Usually -3 dB or -1 dB is used.)
How many bins would you have to average to get audible IMD down to -129dB? 100,000? 1,000,000? More?
Seriously, think before you post.
A bit unstable signal generator clock. (SNR of 1010LT D/A is around 102 dB, so you have many times time averaging already.)
2x averaging drops noise floor by 1-3 dB depending on how white the noise is, drops IMD with a stable clock less. With unstable clock, you can treat IMD as a colored noise.
To go from 102 dB to 128 dB, you need few hundred times averaging, maybe even 1000x. With an unstable signal source that would drop IMD a lot.
(Also, what is with -30 dB up top? Usually -3 dB or -1 dB is used.)
The above seems to paint the picture of a frightful lack of familiarity with FFTs for someone who apparently wants to want to lecture HA about them.
The noise floor for a random noise as portrayed on a FFT is partially based on bin size. The SNR that is used on equipment specs is usually based on a 20-20K -3 dB band or A weighting or such. The FFTs I have been posting are based on a 65k point FFT as they self-document. The bin width is around 1 Hz and depends on the number of points and the sampling frequency. The amount of noise captured by a measurement is proportional to the square root of the bandwidth of the measurement. So a measurement taken over a 1 Hz band captures about 140 times less noise or about 43 dB less noise than one taken over a 20-20KHz band. Therefore the noise floor for these charts can be about 43 dB less than the specified SNR of the Delta 1010LT audio interface.
Coherent (steady, pure) signals are different. They typically are what they are as the carrier by definition primarily fits into a single bin. In general, jitter just makes new sidebands that show up in new bins, it doesn't take much energy out of the carrier.
(http://www.hovirinta.fi/audio/psykoakustiikka/aiheet/direct4.gif)
Knowledge the threshold of hearing and of spreading functions associated with spectral masking strongly suggests that any spurious response that is 100 dB or more down will unconditionally fall under either the masking curve or the curve associated with the threshold of hearing. IOW, it won't be heard, period.
How many bins would you have to average to get audible IMD down to -129dB? 100,000? 1,000,000? More?
Seriously, think before you post.
A bit unstable signal generator clock. (SNR of 1010LT D/A is around 102 dB, so you have many times time averaging already.)
Ok, I see that the problem here is that I was assuming you understood how to read an FFT. Do you understand Arnold's (very gracious) explanation of what that plot is showing? If not, you should ask questions until you do.
(Also, what is with -30 dB up top? Usually -3 dB or -1 dB is used.)
Once you understand Arnold's reply, you will be able to answer this question yourself.
(Hint: the FFT as implemented here conserves power)
So yes, it is averaging a lot (because the FFT is long window - I originally thought a shorter window with major additional averaging specifically because of the sidebands - but that wouldn't reduce the level of the tones either, my mistake). 65k FFT is likely around a second of averaging.
65k FFT should be awfully selective in frequency domain, therefore the only reason for sidebands is that the generator clocks are just not stable enough. Therefore any measurement of THD with them will likewise get smeared out and reduced in level... I do hope they're unstable in the same way in all frequencies (say, digital jitter) - otherwise the IMD will indeed average out as it will show multiple different interference frequencies changing around, looking exactly like colored noise, nicely averaging out due to the long window.
I agree about the masking - do note though that this graph pertains to masking with simple signals and did not evaluate timbre on complex signals. Generally there's another effect in play - an interference changes perceived pitch, even if the tone is masked, as long as there is enough of it in total. The ear will "synthesize" a virtual pitch based on the ratio of interferences.
That said, I think the effect is minor as long as the pure tones are audible, especially for an untrained listener. And perhaps even negligible for a trained listener. We'd have to test it, I'm not aware of any research on masking of inharmonic distortion. This specific effect has not been exactly studied, others have.
Also the IMD is typically further away and also below the tones that could mask it - it's similar to higher order harmonic distortion in this way.
Some source on perception of nonlinearities: http://hephaestusaudio.com/media/2008/11/d...rtion_aes_i.pdf (http://hephaestusaudio.com/media/2008/11/distortion_aes_i.pdf)
This discusses harmonic distortion mostly...
http://www.gedlee.com/downloads/The%20Perc...0Distortion.pdf (http://www.gedlee.com/downloads/The%20Perception%20of%20Distortion.pdf) <- very good tests.
The distortion is more complex than the numbers would lead us to believe. Again more the reason for listening tests everywhere.
--
The real question still stands. What is the cause for the difference I heard? Is that "tube amp making source distortion more audible due to its own distortion" plausible? Do I get to retry with double-blinding?
So yes, it is averaging a lot (because the FFT is long window - I originally thought a shorter window with major additional averaging specifically because of the sidebands - but that wouldn't reduce the level of the tones either, my mistake). 65k FFT is likely around a second of averaging.
I think the problem here is that people were assuming you had some kind of background in engineering/electronics/signals and I see now that you do not. So what exactly is your background? We're probably just talking past you with a lot of engineering in this thread.
The real question still stands. What is the cause for the difference I heard? Is that "tube amp making source distortion more audible due to its own distortion" plausible? Do I get to retry with double-blinding?
bad testing, bad equipment, bad luck, etc
This is the fun part of hardware testing. Something gives you an unexplained result. Now you get to dig into it and try to figure out what it is.
Pah, I just need to read the actual images closely. It has the bin size and windowing function spelled out...
Huge wall of texts would've been avoided if I actually *looked* at the images closely.
Blame me for replying in the morning/late in the evening. Now I feel stupid.
So the only remaining thing is to use a more stable signal source and null out the sines for readability.
--
Yeah, I will go digging for the cause. Starting next months with more advanced measurements of all parts of the chain. Either I'll have to lift the new Agilent - likely impossible, someone will notice and stab me - or carry expensive equpiment on my back to that lab and hope it's nothing power supply related.
Pah, I just need to read the actual images closely. It has the bin size and windowing function spelled out...
Well obviously that will help, but its not the only problem here. You need to understand the images too. All the data in the world won't help you if you don't understand what it means...
Yeah, I will go digging for the cause. Starting next months with more advanced measurements of all parts of the chain. Either I'll have to lift the new Agilent - likely impossible, someone will notice and stab me - or carry expensive equpiment on my back to that lab and hope it's nothing power supply related.
I would start much more simple than that. A very basic RMAA test on an ordinary PC will give you a good idea if something is wrong with your hardware. No need to use complex equipment yet.
So what is all the fuss about here, different hardware measuring differently and a questionable listening test that confirmed it?
Rather than dive into obscure faux-technical excuses (jitter is the ultimate boogeyman), perhaps we can focus on the latter. Samples demonstrating audibility would be in order as well.
I would start much more simple than that. A very basic RMAA test on an ordinary PC will give you a good idea if something is wrong with your hardware. No need to use complex equipment yet.
No, it won't. RMAA is quite incomplete in many ways. I'd also need a soundcard with input that is as good or better than DUT.
I'm not sure I can trust the Lynx for this yet, since I haven't measured its input. (non-loopback) FP10 has known broken inputs, nonlinear in phase and FR.
(I can make RMAA handle absolute levels by calibrating the level beforehand with the voltmeter.)
http://nwavguy.blogspot.com/2011/02/rightm...lyzer-rmaa.html (http://nwavguy.blogspot.com/2011/02/rightmark-audio-analyzer-rmaa.html) - some concerns.
A very long time ago I've ran a test with 3 different sets of volume levels on an Aureon 7.1 Space. The results were wildly different, esp. crosstalk.
I'll mention the more known examples of what I listened to during the test:
Faster than Light Soundtrack - Engi (battle); background aharmonic tone - sounds more/less atonal
PPPPPP - Pressure Cooker; noise timbre
Kraftwerk - Tour de France - Titanium; metallic attack timbre, esp. in presence of thicker bass near the end
The differences are subtle. I can isolate some of the problem spots.
Of note, these tracks also overcome LAME even at a.p.e. Vorbis manages at q6+.
As I've mentioned, I used Hifiman HE-500 w/ pad and grille mods, equalized to perceptually flat with tones and modified double-precision Calf parametric equalizer - the headphone is quite low distortion. Not sure yet how much distortion the tube amp (Eddie Current Super 7) has, but supposedly low as well...
--
Oh, there's a factor that could be vastly different - crosstalk...
So yes, it is averaging a lot (because the FFT is long window - I originally thought a shorter window with major additional averaging specifically because of the sidebands - but that wouldn't reduce the level of the tones either, my mistake). 65k FFT is likely around a second of averaging.
Actually, the length of the averaging period is based on both the sample rate and the length of the FFT. Since this is a 65k FFT and the sample rate is 96 Khz, the period is about 2/3 of a second.
65k FFT should be awfully selective in frequency domain,
Depends what you compare it to. Sample sizes up to 1 million points are available with some FFT-based tools.
therefore the only reason for sidebands is that the generator clocks are just not stable enough.
Completely ignores that there is such a thing as nonlinear distortion in both the amplitude and frequency domains. The later is what we call jitter which can be caused by unstable clocks, but the former is generally far more pervasive and signficiant in the real world.
Generate two or more tones, add nonlinearity in the amplitude domain, mix well, and out pop sidebands.
Nicely summarized here:
http://cp.literature.agilent.com/litweb/pdf/5954-9130.pdf (http://cp.literature.agilent.com/litweb/pdf/5954-9130.pdf)
Therefore any measurement of THD with them will likewise get smeared out and reduced in level...
Not in this world. Real world audio electronics only has enough jitter sufficient to create sidebands that are 80 or more dB below the tone that was modulated by the jitter.
The harmonics in good audio electronics are at least 40 dB down, so the sidebands due to jitter are at least 120 dB down. They might be concealed by the noise floor. Generally speaking distinguishing between two signals that are different by 80 dB is easy enough to do.
I would start much more simple than that. A very basic RMAA test on an ordinary PC will give you a good idea if something is wrong with your hardware. No need to use complex equipment yet.
No, it won't. RMAA is quite incomplete in many ways. I'd also need a soundcard with input that is as good or better than DUT.
You just need a sound card that does not have audible IMD. There are literally billions of these in existence.
http://nwavguy.blogspot.com/2011/02/rightm...lyzer-rmaa.html (http://nwavguy.blogspot.com/2011/02/rightmark-audio-analyzer-rmaa.html) - some concerns.
Do a search, that article is in response to my criticism of nwavguy's instrument fetishism, an odd quirk you seem to be imitating, if not understanding.
Anyway, a simple RMAA test is the first thing you should be doing. One begins with simple, broad tests, and then moves to more specific. Certainly you should have used it before a listening test, let alone anything made by Agilent (which would probably be useless here anyway).
I would start much more simple than that. A very basic RMAA test on an ordinary PC will give you a good idea if something is wrong with your hardware. No need to use complex equipment yet.
No, it won't.
That is one heck of a blanket statement!
RMAA is quite incomplete in many ways.
So is everything short of a Prism, AP or full blown Spectra based measurement system. However, the task at hand is not the grand inquisition or a name-dropping contest.
I'd also need a soundcard with input that is as good or better than DUT.
Who can say this after chastising me for using a too-good sound card that was something with a whopping 8 analog channels in, 8 analog channels out, symmetrical I/O performace and $100 on eBay?
I'm not sure I can trust the Lynx for this yet, since I haven't measured its input.
If you are talking about the LynxTWO, its most noteworthy oddity is that its nonlinear distortion is impressively minimized at FS -10 dB or so but near FS, its just another ca. 100 dB audio interface.
(I can make RMAA handle absolute levels by calibrating the level beforehand with the voltmeter.)
One can make RMAA very omnivorous with respect to levels by doing the recording of the test tones with something other than RMAA.
I don't have a soundcard that is separate from the DUT. Not Lynx Two, but Lynx Hilo.
I could assume it has a great input, which it likely does, but then, I cannot be sure until I measure it, non-loopback.
I'm wary since FP10 has a broken input. (in terms of FR) (Which happened to be found via an extremely cheap old scope.)
$100 for M-Audio Delta 1010LT on ebay? Wow. That thing used to cost $1000+. What a deal. Now, a new one costs $300.
Exactly what I will do, is to take the recordings of the tests and check them with a separate tool indeed.
--
Regardinig the sidebands, I'd expect something looking closer to: https://www.macomtech.com/static/PDFs/Techn...cterization.pdf (https://www.macomtech.com/static/PDFs/TechnicaArticles/RFcharacterization.pdf)
With clear separation between test tones. And those graphs are for RF, not audio, where it's harder to have a stable signal source.
The sidebands here are way excessive. (Blackmann window has slow rolloff, but at 65k FFT that shouldn't matter much.)
All of ARTA, RMAA, TrueRTA and REW do not feature such huge sidebands in THD or IMD measurements. The source peaks should be tight and clear.
I need an actual explanation for them unless I fully accept the measurement. Saying "oh, it's some kind of distorion but not IMD" does not cut it. It's either misset FFT (unlikely given the information) or something else broken, likely the signal generator.
$100 for M-Audio Delta 1010LT on ebay? Wow. That thing used to cost $1000+. What a deal. Now, a new one costs $300.
Not really. The 1010LT is the one with a RC socket RS-232 dongle instead of a outboard 1RU interface. Street price was never higher than $200
The one with the outboard rack-mounted box was the Delta 1010 (no LT). It was a few dB better and had wall-to-wall TRS balanced I/O. Never sold for more than $500 and I have one of those, too.
Regardinig the sidebands, I'd expect something looking closer to: https://www.macomtech.com/static/PDFs/Techn...cterization.pdf (https://www.macomtech.com/static/PDFs/Techn...cterization.pdf)
With clear separation between test tones. And those graphs are for RF, not audio, where it's harder to have a stable signal source.
The sidebands here are way excessive. (Blackmann window has slow rolloff, but at 65k FFT that shouldn't matter much.)
All of ARTA, RMAA, TrueRTA and REW do not feature such huge sidebands in THD or IMD measurements. The source peaks should be tight and clear.
You are in a self-made quandry. You demanded multitones, and you got multones. Of course tests with just 2 tones rather than about 30 give clearer indications. You should be careful about what you ask for - you may get them!
The public faulting of my tests has made me very unwilling to bow and scrape as you seem to be demanding, Take what you got - you've exposed your level of knowledge in this topic and I'm done with you!
OT I know, but It seems to me it's not only format names which are unicorn-like, as I can't help thinking the Schiit folks were really having a laugh/taking the wee at all audiophools when they came up with that horrid name o' theirs - even if it (probably) is their family name.
History of Schiit Audio by the co-founder:
http://www.head-fi.org/t/701900/schiit-hap...obable-start-up (http://www.head-fi.org/t/701900/schiit-happened-the-story-of-the-worlds-most-improbable-start-up)
How the name came about:
http://www.head-fi.org/t/701900/schiit-hap...5#post_10286931 (http://www.head-fi.org/t/701900/schiit-happened-the-story-of-the-worlds-most-improbable-start-up/315#post_10286931)
You are in a self-made quandry. You demanded multitones, and you got multones. Of course tests with just 2 tones rather than about 30 give clearer indications. You should be careful about what you ask for - you may get them!
I didn't ask for 20 tones and for the sidebands. I've personally ran a 7 tone test with no such issues and a "cheap" digital scope to boot. No idea why you have these. Pity that scope can only do about -100 dB before the noise floor hits. Anyway, next time (a long time) I'll be back with some measurements.
--
Also, I do love Schiit story. They do make excellent and cheap amplifiers. (Though not the very best, but close, top 10.) Apparently our measurements have blind spots too - we need to figure out better ones and again, correlate with listening tests.
You are in a self-made quandry. You demanded multitones, and you got multones. Of course tests with just 2 tones rather than about 30 give clearer indications. You should be careful about what you ask for - you may get them!
I didn't ask for 20 tones and for the sidebands. I've personally ran a 7 tone test with no such issues and a "cheap" digital scope to boot.
From an audio perspective cheap digital scopes are useless. They are typically based on 8 bit conversion which is good enough for the purpose of looking at waveforms, but not so much for quality audio. The better ones go up to 12 bits, but analytical software at its best can only follow the example of the human ear, and we already know that 12 bits is marginal to inadequate for high quality listening.
No idea why you have these. Pity that scope can only do about -100 dB before the noise floor hits.
As I demonstrated discrete signals well below -100 dB are clearly visible with that hardware.
Anyway, next time (a long time) I'll be back with some measurements.
Pardon me while I continue breathing steadily and calmly! ;-)
I seem to recall something about claims of artifacts that were 40 dB down. One thing that even an illusory -100 dB noise floor is that it would show them clearly if they existed.