The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Launch of Pono Studio Headphones
Old 26th March 2014
  #1891
In regards to the "successful" kickstarter campaign, what is stopping a couple folks interested in this being successful dropping a million to boost that number?

Looks like 80 people have pitched in $5000.00 or more.
Old 26th March 2014
  #1892
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by jkorten View Post
I can hear the difference on my computer's Radio Shack Optimus speakers, and on my Shure SRH1840 headphones. I'm sure I will hear the same on my Harbeth's in the living room, but I haven't tried yet.
Again, what do you think you're hearing?? I went ahead and set up a little experiment so you can see what's in a square wave. Here's the output of my lab grade HP 33120A function generator set to sine at 10 kHz, +4 dBu into my lab grade Stanford Research SR760 analyzer, which has a real-time 100 kHz bandwidth (forgive the crappy cell phone pics):



I show a sine so you can see the quality of the generator: 0.012% THD at 10 kHz. Next up is a square wave, same frequency and amplitude, with a marker on the fundamental:



You can see the THD measurement shoots up to 34.8%; this is to be expected (higher would be even better), as square waves are mostly harmonics. Next I put a marker on the 2nd harmonic (first overtone):



This is a distortion product -- there should be no 20 kHz tone in a 10 kHz square wave. But it's 64 dB down, only 20 dB above the noise floor; no human being can hear this. Next I move the marker to the 3rd harmonic:



The 30 kHz tone is only 6 dB down; most importantly, this guy right here marks the first difference between a 10 kHz sine and 10 kHz square. If you claim to hear a difference, you're telling us that you're hearing this.

In case you cry foul because I used a digital function generator, I also happen to have an all-analog generator, the Rek RAG-101. Here's a 10k (+4 dBu) sine from the Rek:



You can see that it's much lower quality than the HP (an order of magnitude worse THD), but it's pure analog baby! Now let's see what a 10k square wave looks like:



THD is only 1.2% so it's obviously producing much less harmonic energy; worse still, the difference between the distortion products and the true harmonics are only about 20 dB. Clearly this generator can't hang with the HP, but look closely at that picture: after 10 kHz the next closest tone is a 20 kHz sine at -40 dB. Do you really think you (or anyone else) can hear such a thing?

If you are indeed hearing something, it is undoubtedly one or more of your gear that can't handle 30 kHz. A crappy tweeter (such as found in a Radio Shack speaker, perhaps?) might even make a tiny audible buzz trying to reproduce 30k. One way you could test would be to send a pure sine wave through your gear, then just slowly ramp up the frequency and see if you start hearing things.
Old 26th March 2014
  #1893
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by Mixerman View Post
No. It's not a debate that interests me. My only intention was to point out the absurdity in your claim that "tech industry" collusion amounts to some kind of tin hat conspiracy theory, when there is clear evidence to the contrary. I mean, you asked if there was a "secret handshake" as if it's preposterous that tech companies are working together, when it's been established that they are.
That is still not the whole "Tech Industry". There is no Tech Industry club where they all come together to agree on direction, pricing or propaganda or whatever. There was huge competition between all these players and still is. Of course as in every industry, some of the heavy hitters make unspoken agreements or in the case of Steve Jobs, make angry calls to competitors (which then get ignored and Google make their own browser anyway).

Quote:
This puts your "expertise" in question.
I was very much in the Tech Industry in the time that Bob is alluding to. I was dealing directly with some of the people that stood to profit most from "free music" (and some did) and all the big players were trying to sign content contracts from the content providers. If anyone was colluding it was the content providers. Things really didn't happen the way Bob is presenting them.

Alistair
Old 26th March 2014
  #1894
Lives for gear
Quote:
Originally Posted by Arthur Stone View Post

Is it relevant? I think so: for example, earlier in the thread Kenny and joeq were wondering why it isn't possible to realistically replicate the sound of an instrument...if that isn't possible then why not. Either there is something we psychologically invest into the listening experience (by knowing whether it's a real instrument or recording); or, there is information missing in the recording (or mic or playback system). That intrigues me - is it possible that the ultrasound present with the real instrument but not in the recorded playback has some effect? I'd like to find out.
I, too, am intrigued by this. But just throwing 192k at it seems really juvenile to me. Do some research, prove that you understand (not you personally, the folks advocating a new standard) what constitutes a more realistic experience. My gut feeling is that it has little to do with ultrasonics and more to do with reflections and phase effects. And it could be solved and implemented, but just increasing sampling to 192k is irrelevant!
Old 26th March 2014
  #1895
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by Arksun View Post
Unlike others I dont doubt that you're hearing a difference (although maybe I should as you haven't blind tested yourself to rule out expectation bias)
I believe him too but like you I disagree with the explanation.

Alistair
Old 26th March 2014
  #1896
Lives for gear
Quote:
Originally Posted by jkorten View Post
Well I am alone, alas. I can press the two buttons until I forget which is which but I need a third person. It is easy to hear the difference as the square wave sounds "richer".

At 44.1kHz 16 bit - absolutely not! The square wave becomes a sinewave - so it will sound identical to the sine wave of the signal generator. I have no doubt about that.

My claim was that I can hear the difference between a sine and square wave played through an analog system (a signal generator will even drive a pair of headphones directly for this test).

So... I ask in return, have you tried to hear the difference between an analog square wave and an analog sine wave at 10kHz?

I will do my best to recruit a second subject for an ABX testing per your request, tonight.
So, you've got a playback system that will reproduce 30 kHz at a significant level? Because if it stays in the box (because just one component, like your speakers, can't reproduce it) then it's not adding anything to your experience. And if it IS adding something thru a device that can't reproduce the frequency, then what you're talking about is intermodulation distortion. So, your contention would be that increased intermodulation distortion from an inaudible source is a good thing for sound reproduction?
Old 26th March 2014
  #1897
Gear Guru
 
UnderTow's Avatar
Btw, I could hear a difference between the 10 KHz sine and square waves in the YouTube video that I think joeq posted. Obviously that was after digitisation at 44.1 KHz.

I'll check again tomorrow to make sure I am not making things up.

Alistair
Old 27th March 2014
  #1898
Quote:
Originally Posted by drbob1 View Post
Quote:
Originally Posted by Arthur Stone

Is it relevant? I think so: for example, earlier in the thread Kenny and joeq were wondering why it isn't possible to realistically replicate the sound of an instrument...if that isn't possible then why not. Either there is something we psychologically invest into the listening experience (by knowing whether it's a real instrument or recording); or, there is information missing in the recording (or mic or playback system). That intrigues me - is it possible that the ultrasound present with the real instrument but not in the recorded playback has some effect? I'd like to find out.
I, too, am intrigued by this. But just throwing 192k at it seems really juvenile to me. Do some research, prove that you understand (not you personally, the folks advocating a new standard) what constitutes a more realistic experience. My gut feeling is that it has little to do with ultrasonics and more to do with reflections and phase effects. And it could be solved and implemented, but just increasing sampling to 192k is irrelevant!
I believe problems representing real sounds, particularly unamplified ('acoustic') instruments, in real environments is very much related to two overlapping issues and then totally bollixed by a third. (I've seen over 170 live, unamplified orchestral performances, and, during the too-often played chestnuts, I sometimes ponder the difficulties of capturing the amazing sound of 80 or 100 well-trained humans playing their wind, string, horn, and percussion instruments.)

The three issues are these as I see them:
  • issues of stereophonic/binaural capture
  • issues of preferred playback output (headphone/speaker)
  • issues related to typical human behavior patterns
Let's stipulate a cooked up, 'best-case' (from a very narrow perspective) scenario:
- A paraplegic with excellent hearing and a love of music has a binaural dummy head mic system designed to carefully match his head and upper body shape and then he uses his vast billions (best-case scenario, yeah?) to subsidize a series of recordings of his favorite works by his favorite orchestras in his favorite hall, from his favorite position. (And, of course, they are transcribed via a state-of-art transcription system. We'll stay agnostic on just what that would be. ) He then listens back over custom, mega-high end headphones (that shut out all extraneous noise but are light as a feather, important for our friend). He's going to have the conjectural best possible recreation of what he would normally hear.

But... those of us fortunate to be more able-bodied than our exemplary friend above can move our heads and bodies. Even if we are 'stuck' in our seat, we will typically shift body position and head angle frequently in ways both large and small. Everyone's different, of course, but you get the idea.

When we shift our head, even a few degrees, in a complex listening environment (or even a simple one), it changes what we hear in ways that are significant to this topic -- even when we don't consciously notice them -- because they are being used across time by the unconscious parts of our perceptual system that evolved in part to help us spatially map our immediate area. Like a dog tilting his head back and forth to change visual perspective, we humans tend to form our psychoacoustic impressions across the time domain by collating sound from different angles and sonic 'perspectives.'

Obviously, all that is lost in a binaural dummy head/headphone setup. (Which was why our best case scenario used a paraplegic as its subject.)

You get a certain relatively high accuracy of reproduction, but like an early sound movie acted out in front of a necessarily stationary camera, the 'perspective' and subsequent surrender to illusion is static.

And even less persuasive and more distracting when one is, himself, moving through a complex 3D world environment -- the very realism of the stereo sound imaging can prove problematic and distracting from demanding issues or threats in the immediate environment or intersecting one's path through it.


But, finally, there's the 'problem' of speakers and the widespread preference for headphone-free listening when possible... there may be some 'bud babies' who would rather listen to earbuds or headphones than a nice speaker system -- but it ain't me. When I was a kid, my KossPhones were the only way I could hear halfway reasonable audio. But now that I'm a grown up (an old one at that) and, perhaps more importantly, good audio is a whole, whole lot cheaper, I really do want good speakers in a reasonable room for my main system. I don't want to be stuck in cans, even hi fi wireless ones (and I have heard some crappy consumer wireless ones, that's for sure).

And once your listeners have speakers, we no longer have the luxury of binaural dummy head recording, so there's no standard miking approach that can be optimized for a 'standard' PB system; you end up with recordings recorded in a confusion of capture devices and approaches and being played back on a chaotic universe comprised of a wide array of PB system types in a huge field of necessarily unique environments.

... and you end up about here. Not quite the wild west but...
Old 27th March 2014
  #1899
as was promised earlier, tomorrow I will be recording a piano, a chello, and a violin at 192/24. Shortly thereafter I will bounce the tracks in three formats: a 192/24 track, a downsampled-to-44.1/16 track, and the same 44.1/16 track upsampled back to 192/24 for us all to test in our environments.

i'm the first to admit that the allure of 192/24 led me to believe that it could capture more information and therefore more nuance from the original sound. i think this might be the test that will convince me one way or the other.

i can only help that the end result is useful for you all out there.

one thing i do want to mention, confirmation bias goes both ways. it not only blinds "believers" into only seeing their world-view, but it also blinds "non-believers" in the same manner.

confirmation bias prevents us all from being open-minded; but even the open-minded have their bias
Old 27th March 2014
  #1900
Gear Guru
 
Kenny Gioia's Avatar
 

Quote:
Originally Posted by UnderTow View Post
On a completely different note, some argued that the whole Kickstart thing is just a marketing thing rather than the Pono team actually needing the money to launch the player, shop and desktop application. Do any agree? And if so, wouldn't it be much to early to create so much hype? Wouldn't most people have forgotten by the time October comes?

Alistair
I argued it and Neil admitted it.
Old 27th March 2014
  #1901
Lives for gear
 

Quote:
Originally Posted by orenradio View Post
one thing i do want to mention, confirmation bias goes both ways. it not only blinds "believers" into only seeing their world-view, but it also blinds "non-believers" in the same manner.

confirmation bias prevents us all from being open-minded; but even the open-minded have their bias
Red herring.

The scientific method recognizes that all humans experience bias. Testing protocols and evidentiary chains are meticulously constructed/designed to reduce (or eliminate) bias from the results.

By contrast, claims which fail to adhere to the scientific method typically make little attempt to acknowledge or eliminate bias.

Extraordinary claims require extraordinary proof, especially regarding bias.
Old 27th March 2014
  #1902
Lives for gear
 
James Lehmann's Avatar
 

Quote:
Originally Posted by orenradio View Post
as was promised earlier, tomorrow I will be recording a piano, a chello, and a violin at 192/24. Shortly thereafter I will bounce the tracks in three formats: a 192/24 track, a downsampled-to-44.1/16 track, and the same 44.1/16 track upsampled back to 192/24 for us all to test in our environments.
Would love to participate in this but I don't have any 192k-ready gear... guess I'll just have to buy a Pono!
Old 27th March 2014
  #1903
Gear Head
 
bandpass's Avatar
 

Quote:
Originally Posted by bogosort View Post
You can see that it's much lower quality than the HP (an order of magnitude worse THD), but it's pure analog baby!
Also, look at the phase-noise of the analog signal (the width of the peaks)—much worse than with the digital.
Old 27th March 2014
  #1904
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by orenradio View Post
as was promised earlier, tomorrow I will be recording a piano, a chello, and a violin at 192/24. Shortly thereafter I will bounce the tracks in three formats: a 192/24 track, a downsampled-to-44.1/16 track, and the same 44.1/16 track upsampled back to 192/24 for us all to test in our environments.
Let's keep things simple or at least test them one at a time: Stick to 24 bit and just compare the sample rates. Also, make sure you use a quality SRC, NOT the one built into your DAW. I suggest the iZotope SRC or the SoX one (It is free).

Alistair
Old 27th March 2014
  #1905
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by Kenny Gioia View Post
I argued it and Neil admitted it.
Where did he admit it?

Thanks,

Alistair
Old 27th March 2014
  #1906
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by UnderTow View Post
Btw, I could hear a difference between the 10 KHz sine and square waves in the YouTube video that I think joeq posted. Obviously that was after digitisation at 44.1 KHz.

I'll check again tomorrow to make sure I am not making things up.
Yup. I hear a difference. I'm quite sure I hear a difference for the 12K waves too. I could be fooling myself of course and obviously I don't hear 36Khz so there is something else going on. Maybe it is just the level difference that fools the ear!

Alistair
Old 27th March 2014
  #1907
Gear Head
 
bandpass's Avatar
 

Quote:
Originally Posted by UnderTow View Post
Let's keep things simple or at least test them one at a time: Stick to 24 bit and just compare the sample rates. Also, make sure you use a quality SRC, NOT the one built into your DAW. I suggest the iZotope SRC or the SoX one (It is free).

Alistair
Here's a shell script to create a set of test files on a linux box (using a sample from lessloss.com; any 24/96 sample can of course be used instead):

Code:
#!/bin/sh -x
set -e
sox=/usr/bin/sox     # Where to find sox

# Fetch the original sample; fade the end and fix the level:
len=30               # Sample length in seconds
url="http://dl.dropbox.com/u/61558493/1395910802_586/1034/BWV_1034_e-moll_Part_I_Adagio_ma_non_tanto_(24_bit_96_kHz).wav"
wget -qO - "$url" | $sox - a.wav fade h 0 $len 1 norm -1

# File descriptions:
set -- \
"a: Original 24/96" \
"b: Down to 24/44.1" \
"c: Down to 16/44.1" \
"d: Down to 16/96" \
"e: From 24/44.1 back to 24/96" \
"f: From 16/44.1 back to 24/96" \
"g: From 16/96 back to 24/96" \
"h: 24/44.1 - original" \
"i: 16/44.1 - original" \
"j: 16/96 - original"

# Create the rest of the files according to descriptions above:
$sox a.wav       b.wav rate 44100
$sox b.wav -b 16 c.wav            dither -S # mild shaping
$sox a.wav -b 16 d.wav            dither -S
$sox b.wav       e.wav rate 96k
$sox c.wav -b 24 f.wav rate 96k
$sox d.wav -b 24 g.wav
$sox -mv-1 [ae].wav h.wav
$sox -mv-1 [af].wav i.wav
$sox -mv-1 [ag].wav j.wav

# Create spectrograms to check all is well:
for f in [a-j].wav; do $sox $f -n spectrogram -ho $f.png -t "$1"; shift; done
Old 27th March 2014
  #1908
Gear Nut
 

Quote:
Originally Posted by bogosort View Post
Again, what do you think you're hearing?? I went ahead and set up a little experiment so you can see what's in a square wave. Here's the output of my lab grade HP 33120A function generator set to sine at 10 kHz, +4 dBu into my lab grade Stanford Research SR760 analyzer, which has a real-time 100 kHz bandwidth (forgive the crappy cell phone pics):


I show a sine so you can see the quality of the generator: 0.012% THD at 10 kHz. Next up is a square wave, same frequency and amplitude, with a marker on the fundamental:


You can see the THD measurement shoots up to 34.8%; this is to be expected (higher would be even better), as square waves are mostly harmonics. Next I put a marker on the 2nd harmonic (first overtone):


This is a distortion product -- there should be no 20 kHz tone in a 10 kHz square wave. But it's 64 dB down, only 20 dB above the noise floor; no human being can hear this. Next I move the marker to the 3rd harmonic:


The 30 kHz tone is only 6 dB down; most importantly, this guy right here marks the first difference between a 10 kHz sine and 10 kHz square. If you claim to hear a difference, you're telling us that you're hearing this.

In case you cry foul because I used a digital function generator, I also happen to have an all-analog generator, the Rek RAG-101. Here's a 10k (+4 dBu) sine from the Rek:


You can see that it's much lower quality than the HP (an order of magnitude worse THD), but it's pure analog baby! Now let's see what a 10k square wave looks like:


THD is only 1.2% so it's obviously producing much less harmonic energy; worse still, the difference between the distortion products and the true harmonics are only about 20 dB. Clearly this generator can't hang with the HP, but look closely at that picture: after 10 kHz the next closest tone is a 20 kHz sine at -40 dB. Do you really think you (or anyone else) can hear such a thing?

If you are indeed hearing something, it is undoubtedly one or more of your gear that can't handle 30 kHz. A crappy tweeter (such as found in a Radio Shack speaker, perhaps?) might even make a tiny audible buzz trying to reproduce 30k. One way you could test would be to send a pure sine wave through your gear, then just slowly ramp up the frequency and see if you start hearing things.
Nice equipment! I always wanted my own spectrum analyzer! I have to do all of mine through coding. :(

So it would be interesting to post the spectral analysis of the playback of a 192kHz sampled sine vs square wave at 10kHz and we see what other components are there (sum/difference frequencies with sample rate etc).

It is strange to me the amount of energy put into debunking the use of a higher sample rate. I used to do tape recording in college to work my way through graduate school. Only came back to recording in the digital age after lots of audiophilia in the interim. I gravitated to a high sample rate DAC because the sound of the 192kHz files was so much more lifelike. Just listen to a jazz album like "Magnetic" by Terrance Blanchard, or Hiromi's "Move" and the very first thing you will notice is the sound of the cymbals. WAY more natural sounding.

I record myself on piano (my own audience) to improve my playing, and the difference between 44.1 and 192 on piano is also striking to me. Strings sound more like steel and less like rubber.

So for me the proof was in the pudding long before I started thinking about why.

But I if you think it is IM distortion - well this is probably also occuring on our eardrums as well, and maybe that's why it sounds more natural? I used to play a fantastic clarinet duet written by Leslie Bassett that relied on the two tones of the clarinet causing a "difference" tone (that exists only on the eardrum) that sometimes became so dominant it almost hurt. Fantastic piece.

What I hear with the 10K square vs sine is more rich sound. Like clarinet vs flute.

I'll record at 192k today sin and square and post, people with 192k DACs can hear for themselves too.
Old 27th March 2014
  #1909
Gear Nut
 

digitized waveforms

So I'm getting out my old Revox A77 and dumping all this digital stuff. Whoa boy! What a mess!

So I tried recording 10Khz as a sine and a square wave at 192kHz sample rate and - SURPRISE! There is a subtone at 2kHz that is LOUD and is probably a difference tone from the sample rate. UGH.

(I confirmed that with headphones listening only to the signal generator I did not hear this at all - only the rich overtones of the 10kHz square wave were present.)

So then I decided... hmmm, well maybe this recording chain isn't optimized for non-typical audio situations. Let me record a 2.5kHz square wave for playback. Same issue - you can load up the files in audacity and to "analyze -> plot spectrum" and see the other tones in the square wave file that aren't supposed to be there.

Then I thought - hmmm, I wonder what 44.1kHz does with this 2.5kHz waveform... and you know what the tone shifted higher and was there in all it's glory at a different frequency!

AAAKKK!

I thought at least we could compare 192k vs 44.1k sample rate files of the square wave - but no deal. The subtones give away the sample rate right away.

The recording chain btw - Escort EFG2210 signal generator into Mytek 192ADC captured with Audacity as flac files. They are presented for our mutual concern and horror here. Please let me know if I forgot to do something here that has led to this horrific result. Oh and btw I didn't try changing levels between square and sine to match RMS recording levels, the generator maintains peak to peak voltage.



sine-2_5khz_44_1ksr.wav

square-2_5khz_44_1ksr.wav

sine-2_5khz_192ksr.wav

square-2_5khz_192ksr.wav
Old 27th March 2014
  #1910
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by jkorten View Post
But I if you think it is IM distortion - well this is probably also occuring on our eardrums as well, and maybe that's why it sounds more natural?
I haven't tested it myself, but Paul Frindle has tested for this very thing. He concluded that our hearing is linear outside our range of hearing and thus produces no difference tones. You can try to replicate his results, of course.

Quote:
What I hear with the 10K square vs sine is more rich sound. Like clarinet vs flute.
It's gotta be aliasing or IMD. Interesting side note: clarinet is one of the instruments shown not to produce any ultrasonic energy. All that richness is completely characterized within a 20 kHz bandwidth. (Of course, I'm arguing that this is perceptually true for all instruments.)

Quote:
I'll record at 192k today sin and square and post, people with 192k DACs can hear for themselves too.
Look forward to the results.
Old 27th March 2014
  #1911
Gear Nut
 

10kHz

Wouldn't let me upload last time. Forgot to shorten length to 8 seconds.

square-10khz_192ksr.wav

sine-10khz_192ksr.wav


Trying not to pour myself a scotch at 10AM here...

Obviously what is needed is a sample rate FAR HIGHER than 192kHz. Like 5X higher to lift the beat frequencies up to and beyond 10kHz.

Ugh.
Old 27th March 2014
  #1912
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by jkorten View Post
The recording chain btw - Escort EFG2210 signal generator into Mytek 192ADC captured with Audacity as flac files. They are presented for our mutual concern and horror here. Please let me know if I forgot to do something here that has led to this horrific result. Oh and btw I didn't try changing levels between square and sine to match RMS recording levels, the generator maintains peak to peak voltage.
Whoa! That 2.5k sine into 44.1 is fugly! At first I thought it was your Escort producing a shyte signal, but the 192k capture is obviously much better. Something is horribly wrong with the 44.1 capture. It almost looks like the 2.5k sine is aliasing, but that doesn't make any sense.
Old 27th March 2014
  #1913
Lives for gear
 
bogosort's Avatar
For comparison, here is my 2.5k sine into 44.1. Signal path: HP 33120A -> patchbay -> RME Multiface II (44.1 kHz) -> Audacity.
Attached Files

HP_sine-2_5k_44_1.wav (1.17 MB, 170 views)

Old 27th March 2014
  #1914
Gear Nut
 

Quote:
Originally Posted by bogosort View Post
For comparison, here is my 2.5k sine into 44.1. Signal path: HP 33120A -> patchbay -> RME Multiface II (44.1 kHz) -> Audacity.
Can you match the level to mine and re-digitize? -36db I will sample mine at a large amplitude like yours.

Agree the spectrum shows all kinds of whacky stuff on mine. I think we're looking at quantization issues when the signal is this small. I purposely recorded the signal at a low amplitude thinking that high frequencies are naturally not that loud in music.
Old 27th March 2014
  #1915
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by jkorten View Post
Wouldn't let me upload last time. Forgot to shorten length to 8 seconds.

Attachment 391517

Attachment 391518


Trying not to pour myself a scotch at 10AM here...

Obviously what is needed is a sample rate FAR HIGHER than 192kHz. Like 5X higher to lift the beat frequencies up to and beyond 10kHz.

Ugh.
Something is wrong with your signal path man. Maybe it is time for the scotch.

Attached are 10k square waves from my HP recorded into audacity at 44.1 and 96 (highest my RME goes). You can see the 44.1 is a perfect sine wave and the 96 version has only two overtones. Obviously RME makes quality conversion.
Attached Files

HP_square-10k-44_1.wav (1.03 MB, 106 views)

HP_square-10k-96.wav (2.13 MB, 176 views)

Old 27th March 2014
  #1916
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by jkorten View Post
Can you match the level to mine and re-digitize? -36db I will sample mine at a large amplitude like yours.

Agree the spectrum shows all kinds of whacky stuff on mine. I think we're looking at quantization issues when the signal is this small. I purposely recorded the signal at a low amplitude thinking that high frequencies are naturally not that loud in music.
I set levels at +4 dBu (1.23 Vrms). Re-capture with that amplitude and see if they look like my 10k square waves. If not, scotch time!
Old 27th March 2014
  #1917
Gear Nut
 

Quote:
Originally Posted by bogosort View Post
I set levels at +4 dBu (1.23 Vrms). Re-capture with that amplitude and see if they look like my 10k square waves. If not, scotch time!
Wow. Straight back to Mytek this A/D goes.

Something is totally whacked with it. You are right the signal generator does fine at 192, but still the subtones on the square wave mean something bad is going on here.

Scotch time, if I didn't have so much to do today....
Old 27th March 2014
  #1918
Cheers to jkorten for putting in the time and effort to do this personal testing. As many of us have discovered, setting up a proper test is not always a slam dunk, from many perspectives. A lot can go wrong implementing 'experimental designs' if you haven't been doing it for some time. That's why it's best to go into tests set up by oneself or others with extra care and wide-open eyes, carefully thinking through the logic of the design to make sure it's set up in a way that the results can only reasonably be interpreted to one conclusion and that are not undercut by ambiguity or supposition. Science, it's dirty work, but someone's gotta do it.

Actually, looks like he's come to some conclusions since I started to post this.

jkorten -- I'd recommend taking a couple of deep breaths, do some other work, maybe sleep on it, and reinvestigate your process. Your device could have a problem -- and I'll admit I haven't been following the details of that -- but it's probably more probable that someone who doesn't do this sort of experimental testing frequently could have inadvertently let some error/problem sneak into the test process.
Old 27th March 2014
  #1919
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by jkorten View Post
Wow. Straight back to Mytek this A/D goes.

Something is totally whacked with it. You are right the signal generator does fine at 192, but still the subtones on the square wave mean something bad is going on here.

Scotch time, if I didn't have so much to do today....
Agreed. And look on the bright side: at least you found the weak spot in your chain. I wonder how many other people are hearing "night and day" differences between 44.1 and higher rates simply because their converters suck at 44.1.

But you were right: you were definitely hearing something!
Old 27th March 2014
  #1920
Gear Maniac
 
BasHermus's Avatar
 

It sounds as if the square wave was not properly low-passed before being converted to digital.

In that case, this is what should be expected. Even with a 2.5 kHz signal (given that a square wave is the sum of an infinite number of higher order harmonic sine-waves of the root of the wave (in this case 2.5 kHz)).

This would also mean that you get lower pitched and louder distortions in a 44.1 kHz recording compared to a 192 kHz one.
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump