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 2nd April 2014
  #2161
Lives for gear
 
nuthinupmysleeve's Avatar
 

Quote:
Originally Posted by bogosort View Post
Fair enough, but a bit of thought should be given to the filter, otherwise you're only testing whether you can hear the phase shift of an 8th-order LPF at 20k. I'd recommend moving the cutoff to at least 22k and experiment whether using a linear phase filter improves things.

BTW, wasn't trying to take a jab at your ears; I actually have no idea if I would hear it either (I should test it). Was just trying to point out that a steep filter at 20k will definitely have in-band artifacts.
Fair enough on all points, and thanks for chiming in.
Old 2nd April 2014
  #2162
Gear Guru
 
Sounds Great's Avatar
 

Quote:
Originally Posted by theblue1 View Post
If all analog systems behaved with reliable linearity, we wouldn't need digital. heh


But the best analog playback system for the masses is the record player, and they are a bit difficult to use in your car or while you are going for a run or a trip down the ski slope.

The silly thing about all of hype is that DSD, which is FAR closer to reproducing the sound of the source, has been around for over a decade as a storage and playback system. Not sure why Neil Young and the rest of the industry never took notice.
Old 2nd April 2014
  #2163
Lives for gear
 
nuthinupmysleeve's Avatar
 

Quote:
Originally Posted by j_j View Post
Make sure the in-band gain is the same, too.

You just read a completely horrid ABX protocol that depends on the actual filter implementation more than anything else.

Given prior experience I have great faith in the ability of people to make bad filter designs.
Have you done an a/b/x with some of these bad filter designs?

Again, I could NOT hear the difference when I removed the upper frequencies using a low pass filter set at 20k. If he also cannot, even with this supposedly HORRIBLE and inferior testing methodology, then suffice it to say that is even a stronger point.

Please do the test as I suggested... using quality LPF in a two track editor, and test it with a/b/x testing yourself. I know what is supposed to be audible, but why not test it yourself before throwing out poo?
Old 2nd April 2014
  #2164
Lives for gear
 
nuthinupmysleeve's Avatar
 

Quote:
Originally Posted by Sounds Great View Post
The silly thing about all of hype is that DSD, which is FAR closer to reproducing the sound of the source, has been around for over a decade as a storage and playback system. Not sure why Neil Young and the rest of the industry never took notice.
They did... it was reportedly what they originally planned. Not sure why they switched horses.
Old 2nd April 2014
  #2165
Lives for gear
 
paul brown's Avatar
Quote:
Originally Posted by nuthinupmysleeve View Post
using quality LPF
what would you consider as being quality? i'm just looking for a starting point.
Old 2nd April 2014
  #2166
Lives for gear
 
nuthinupmysleeve's Avatar
 

Quote:
Originally Posted by paul brown View Post
what would you consider as being quality? i'm just looking for a starting point.
Good question! I have the waves suite and I just did it with a Q1 filter myself.

It might be interesting to see if you can do it using whatever LPF filter is available in your two track editor just for the heck of it. If the difference is as horrible as people say, it should be extremely audible then we can work our way toward higher quality testing.
Old 2nd April 2014
  #2167
Quote:
Originally Posted by Sounds Great View Post
But the best analog playback system for the masses is the record player, and they are a bit difficult to use in your car or while you are going for a run or a trip down the ski slope.

The silly thing about all of hype is that DSD, which is FAR closer to reproducing the sound of the source, has been around for over a decade as a storage and playback system. Not sure why Neil Young and the rest of the industry never took notice.
Well, I find the technical explanations of why the technology would deliver audibly superior sound to contemporary multi-bit oversampling converters to be not persuasive. And the golden ear contingent has always seemed divided on the format.

You know, I used to be able to say that I'd compared SACD (DSD) to CD on a ~$40K stereo and heard no significant difference... but we did a little digging and realized the SACD in question was actually just repackaged CD audio.
Old 2nd April 2014
  #2168
Gear Addict
Quote:
Originally Posted by nuthinupmysleeve View Post
They did... it was reportedly what they originally planned. Not sure why they switched horses.
For those interested, there are portable DSD players (this one should be out soon for less money than Pono) and DSD downloads available.

Cheers,
Eddie
Old 2nd April 2014
  #2169
Lives for gear
 
paul brown's Avatar
in ocenaudio i just took an mp3 copy of something i recorded at 44.1. i used the LPF built in to the editor to remove the top 2 kHz. i made a copy of the filtered and unfiltered track and then put them in the ABXTester. i got 100% ten times. am i hearing the previously mentioned issues with LPF filter quality? i'm surprised at my sensitivity at such frequencies given my age and as i was listening to an mp3 (albeit something i am intimate with and with good dynamic range/not limited).

now to try with some wavs!
Attached Thumbnails
Launch of Pono-screen-shot-2014-04-02-10.52.01-pm.png   Launch of Pono-screen-shot-2014-04-02-10.52.19-pm.png  
Old 2nd April 2014
  #2170
Lives for gear
 
paul brown's Avatar
now to try with some higher rates and a higher cut-off. ABX testing is a lot of fun!
Old 2nd April 2014
  #2171
Lives for gear
 
nuthinupmysleeve's Avatar
 

Quote:
Originally Posted by paul brown View Post
in ocenaudio i just took an mp3 copy of something i recorded at 44.1. i used the LPF built in to the editor to remove the top 2 kHz. i made a copy of the filtered and unfiltered track and then put them in the ABXTester. i got 100% ten times. am i hearing the previously mentioned issues with LPF filter quality? i'm surprised at my sensitivity at such frequencies given my age and as i was listening to an mp3 (albeit something i am intimate with and with good dynamic range/not limited).

now to try with some wavs!
Very interesting, thank you for testing! What 2-track editor did you use?
Old 2nd April 2014
  #2172
Lives for gear
 
paul brown's Avatar
so i took the first track from Amber Rubarth - Sessions From The 17th Ward {Binaural} (2012) [Chesky 24-96] and repeated the same process this time cutting off everything above 22190 Hz. loaded both versions copied from ocenaudio and got 100% again. it is subtle but there is a difference with headphones out of my macbook headphone output.

the files are FLAC but then saved as wav so that i could use the ABXTester. i'm wondering about the quality of the LPF. you can see the settings in the snapshot. am i doing something wrong?
Attached Thumbnails
Launch of Pono-screen-shot-2014-04-02-11.38.43-pm.png   Launch of Pono-screen-shot-2014-04-02-11.39.31-pm.png   Launch of Pono-screen-shot-2014-04-02-11.45.06-pm.png  
Old 2nd April 2014
  #2173
Lives for gear
 
paul brown's Avatar
Quote:
Originally Posted by nuthinupmysleeve View Post
Very interesting, thank you for testing! What 2-track editor did you use?
Ocenaudio
Old 2nd April 2014
  #2174
Lives for gear
 
paul brown's Avatar
should i do the test more than ten times? is ten out of ten enough to not be a lucky run of guesses? should i try another LPF to eliminate the unknown variable as to its quality. i'm still interested in what is considered a good quality LPF?
Old 2nd April 2014
  #2175
Lives for gear
 
nuthinupmysleeve's Avatar
 

Quote:
Originally Posted by paul brown View Post
should i do the test more than ten times? is ten out of ten enough to not be a lucky run of guesses? should i try another LPF to eliminate the unknown variable as to its quality. i'm still interested in what is considered a good quality LPF?
10 out of 10 is good enough for me!

I would try a more gentle slope and see if that makes a difference.
Old 2nd April 2014
  #2176
Lives for gear
 
nuthinupmysleeve's Avatar
 

Quote:
Originally Posted by paul brown View Post
so i took the first track from Amber Rubarth - Sessions From The 17th Ward {Binaural} (2012) [Chesky 24-96] and repeated the same process this time cutting off everything above 22190 Hz. loaded both versions copied from ocenaudio and got 100% again. it is subtle but there is a difference with headphones out of my macbook headphone output.

the files are FLAC but then saved as wav so that i could use the ABXTester. i'm wondering about the quality of the LPF. you can see the settings in the snapshot. am i doing something wrong?
Oh, one other thing... are you sure your macbook can work at 96k natively? I think they can, but it's also possible it's doing real time downsampling. Look at what it says in utilities/ audio midi setup.
Old 2nd April 2014
  #2177
Lives for gear
 
paul brown's Avatar
i'm trying to understand what influence the passband attenuation has on the filter in my testing. can someone check my settings in the screenshot and maybe guide me if i am making a mistake. in the meantime i'll do some study on it and play around with the settings.
Old 2nd April 2014
  #2178
Lives for gear
 
paul brown's Avatar
Quote:
Originally Posted by nuthinupmysleeve View Post
Oh, one other thing... are you sure your macbook can work at 96k natively? I think they can, but it's also possible it's doing real time downsampling. Look at what it says in utilities/ audio midi setup.
96/24 is its highest output rate for the mac headphone out.
Old 2nd April 2014
  #2179
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by paul brown View Post
should i do the test more than ten times? is ten out of ten enough to not be a lucky run of guesses? should i try another LPF to eliminate the unknown variable as to its quality. i'm still interested in what is considered a good quality LPF?
Yes, the idea is to make the probability that you're guessing lower than some threshold. This is a deep subject, but if you can get 13 out of 15 you can be reasonably sure you're not guessing. A repeated test at another time with similar results can provide more weight. Just don't discount the ones you didn't "pass"!

As for filters appropriate for this test, that's another deep subject. Unless you're a filter expert, you're probably going to be hearing the filter effects rather than the bandwidth differences. But if you want to experiment, try a linear phase filter. Also, as JJ mentioned, you have to make sure the in-band gain is the same for both files.
Old 2nd April 2014
  #2180
Lives for gear
 
paul brown's Avatar
Quote:
Originally Posted by bogosort View Post
But if you want to experiment, try a linear phase filter. Also, as JJ mentioned, you have to make sure the in-band gain is the same for both files.
i'm waiting for him to show me the flaws and maybe some advice about the in-band gain. any particular linear phase filter you would recommend?
Old 2nd April 2014
  #2181
j_j
Lives for gear
Quote:
Originally Posted by paul brown View Post
i'm waiting for him to show me the flaws and maybe some advice about the in-band gain. any particular linear phase filter you would recommend?
Why should I show you 'the flaws', you are clearly stipulating that you don't know exactly what the in-band gain is.

Why would you want to use a symmetric (i.e. contant delay) FIR? Why not an apodizing FIR?

What do you want to measure? Picking some arbitrary filtering algorithm is more testing the filter design and implementation than anything else.

It says nothing, per se, about sampling rate.

So what do you want to test. You need to decide that first. Filter, sampling rate, what?
Old 2nd April 2014
  #2182
j_j
Lives for gear
Quote:
Originally Posted by nuthinupmysleeve View Post
Have you done an a/b/x with some of these bad filter designs?
Some of them, yes.

Some of them are easily measurable to be above known thresholds.

What was your point beyond bringing up 'throwing poo'.

Or was that your real point, trying to be insulting and professionally insulting?
Old 2nd April 2014
  #2183
j_j
Lives for gear
Here, if you want constant delay, use this in octave to get your filter design:

bb=remez(319,[0 20/44.1 22/44.1 1],[1 1 0 0]);

Hope that helps.

Note, that's presuming an 88k sampling rate, and filtering to 44k bandwidth.

If you want to use 96k, use

bb=remez(319,[0 20/48 24/48 1],[1 1 0 0]);
Old 2nd April 2014
  #2184
Lives for gear
 
paul brown's Avatar
Quote:
Originally Posted by j_j View Post
Why should I show you 'the flaws', you are clearly stipulating that you don't know exactly what the in-band gain is.

Why would you want to use a symmetric (i.e. contant delay) FIR? Why not an apodizing FIR?

What do you want to measure? Picking some arbitrary filtering algorithm is more testing the filter design and implementation than anything else.

It says nothing, per se, about sampling rate.

So what do you want to test. You need to decide that first. Filter, sampling rate, what?
thanks for pointing out the flaws in my knowledge. it is a steep learning curve considering the variables i am trying to get my head around. as the discussion is about hearing differences in sample rates, then that is what i am interested in. i had the feeling based on some of your earlier comments that i would run into difficulties with the LPF. i decided to have a go at an ABX following advice on the thread on how to set it up. back to an earlier question, would i be better feeding a 192 recording from my DAC to my ADC and changing the sample rate at that point then compare the captures? anyway, thanks for some new terms that i was unaware of. i'll go look them up.
Old 3rd April 2014
  #2185
Lives for gear
 
paul brown's Avatar
i will say j_j, if i had your obvious level of knowledge i'd be helping people, not posting octave designs to someone who obviously is on a completely different level to you in terms of experience and understanding. if you don't want to help me (i realise you are under absolutely no obligation), then that saddens me but i'll find the answers another way. if you ever need to know anything about horticulture and botany (my areas of expertise), i'll happily walk you through the principles!
Old 3rd April 2014
  #2186
Lives for gear
 

Quote:
Originally Posted by Sounds Great View Post
... The silly thing about all of hype is that DSD, which is FAR closer to reproducing the sound of the source, has been around for over a decade as a storage and playback system. Not sure why Neil Young and the rest of the industry never took notice.
On your first point, there has been at least one paper pointing out why DSD is technically flawed at its heart.
On your second point, Neil has made much of the fact Pono want to bring you "the original master", or as close as possible to it. In the overwhelming majority of releases, especially historical, they are in PCM format.
Old 3rd April 2014
  #2187
j_j
Lives for gear
Quote:
Originally Posted by paul brown View Post
i will say j_j, if i had your obvious level of knowledge i'd be helping people, not posting octave designs to someone who obviously is on a completely different level to you in terms of experience and understanding. if you don't want to help me (i realise you are under absolutely no obligation), then that saddens me but i'll find the answers another way. if you ever need to know anything about horticulture and botany (my areas of expertise), i'll happily walk you through the principles!
Well, my experience on this board has been, um, interesting.

I'm sorry if the octave script isn't helping you, but it's the most concise way to describe a solution.

As to testing your setup, since I can see how to post a .wav now, I will put up a signal you can use to test your filter when I get time to make it. (sorry, reality also calls)

ETA: Octave is free. Getting it, the signal processing and audio packages (also free) and just playing with it, is, by the way, a good way to learn things. You could also probably model spreading of California poppy in a seattle garden (but that's easy, it just takes over if you don't resort to chemical warfare...) too.
Old 3rd April 2014
  #2188
j_j
Lives for gear
Quote:
Originally Posted by Don Hills View Post
On your first point, there has been at least one paper pointing out why DSD is technically flawed at its heart.
I would say that DSD is not so much "flawed" as "highly inefficient".

As long as you filter out the out-of-band noise, it's not a bad thing, in fact most DAC's nowdays use the same kind of oversampling, give or take.

There are dithering issues, but even Stanley agrees that they are at fairly low levels.
Old 3rd April 2014
  #2189
j_j
Lives for gear
Ok, here you go.

Apparently I can't upload a matlab script, but if you stuff this signal through your filter, compensate precisely for the filter delay, and take the difference, you should see very, very small numbers, if any non-zero, except for the first and last few samples of the difference signal.

If you see that, your filter is good enough.

The file has an ugly sounding signal that is basically bandlimited white noise, between 50 and 20000 Hz.

So you will see most anything the filter is doing (once you compensate for the delay) in the difference signal.
Attached Files

ugh.wav (2.00 MB, 106 views)

Old 3rd April 2014
  #2190
Lives for gear
 
nuthinupmysleeve's Avatar
 

Quote:
Originally Posted by j_j View Post
Some of them, yes.

Some of them are easily measurable to be above known thresholds.

What was your point beyond bringing up 'throwing poo'.

Or was that your real point, trying to be insulting and professionally insulting?
Seriously, I apologize if it came off that way... it was meant to be funny.

I know it's difficult to convey humor on a board like this, but I thought that was kinda funny...
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