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 9th April 2014
  #2731
Airwindows
 
chrisj's Avatar
I kinda failed until I went over to the headphones… which bypass the amplifiers, and the Lavry Black headphone drivers are MONSTERS, insanely powerful. Thank goodness I didn't try to push the NS10s, they'd have just blown on me. They were what I first heard the 'differences' on, but it wasn't resilient enough, it was one of those 'now and then you really get a sense of it' things. To hear it every time I was compelled to go over to headphone listening.

I have this feeling that no matter what I do or say, you're going to say 'but it would be miraculous and impossible for you to have heard this, therefore there has to be another explanation'. Can you not? You keep saying 'I don't believe'. I would have believed I could hear the harpsichord thing, or the jazz track, but not so much. I still want extended resolution (defined as sample rate and wordlength) to cover the full range of all possible listening and audio-content situations.

And none of us want to be listening to 384K while flapping angel wings. I would just like my painfully-arrived-at data point to be counted, because it's helping me determine a much more reasonable definition of 'high resolution' for me. I'm REALLY CLOSE to 44.1 being enough for any possible situation, but it's just 'inadequate' enough to keep hinting at me that something's missing. To me, pushing things with ABX illustrates that yes, it's right at the edge there, because I can exaggerate things either with EQ or just turning up the volume until I can pick it out every time.

And I shouldn't have said 'more resolution is better', I should have said 'of 16/44.1 vs 96/24, 96/24 is better'. In some ways 192K might be actively worse (one of the few things I agree with 'xiphmont' about). It might hurt midrange quietness (I don't think his ideas about IM distortion are relevant, though, because generally those upper octaves will sit empty and lack of sound doesn't induce IM distortion)
Old 9th April 2014
  #2732
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by ezraz View Post
You are focusing on the X axis and the ticks going from 44100 to 48000 to 96000. That's barely more than doubled.

Focus on the Y axis - the bit depth. The count of those ticks goes from 16000 to 15000000.

Methinks my ears might make out the 14,984,000 more samples on the Y axis.

Also, if the X axis correlates to the changes in frequency, does the Y axis correlate to the changes in timbre? i used to think of bit depth as raw data rate but i now think of it as the Y axis, and the hidden world of soundwave blending.
Sampling and quantization are two different things (though combined in PCM). Quantization -- the y-axis -- fixes the amplitude of a sample. This has nothing to do with timbre, it's just the level of that sample. If properly dithered, quantization is no different than signal + noise (just like in the analog world). If you want less noise, you need more quantization steps (bits). A 16-bit quantizer will use 65536 steps; a 24-bit quantizer will use 16.7 million steps. That may seem like a huge difference, but it's only 48 dB. The noise floor of 16 bit samples is -96 dB down, which is very likely much, much lower than the noise floor of the room in which your speakers are.

Have you ever heard the noise floor of a CD?

Note that we're talking about delivery formats. There are several very good reasons to record and produce music with 24-bit samples.
Old 9th April 2014
  #2733
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by chrisj View Post
I have this feeling that no matter what I do or say, you're going to say 'but it would be miraculous and impossible for you to have heard this, therefore there has to be another explanation'. Can you not? You keep saying 'I don't believe'.
Because you are saying that you can hear ultrasonic frequencies. If it were actually true, I'd say let's go get you tested by an audiologist and then publish the results! You'd be a celebrity in the psychophysics world.

Please know that I completely grant that it is possible; it's just so very unlikely that I'm going with the simpler hypothesis. But this is your thing -- if you can really hear ultrasonics, seriously, go get tested to make it official and then get in touch with the closest university.
Old 9th April 2014
  #2734
Lives for gear
 
3rd Degree's Avatar
 

Quote:
Originally Posted by chrisj View Post
My full ABX session, unedited, from beginning to end. I told you I wouldn't be able to identify the harpsichords, and I couldn't, though I thought I could. I bailed out of several tests because I could tell they'd be too tricky to do 20 times in a row. 5 is 4, with Airwindows SlewOnly applied identically to both 4 examples (this is like a really vicious highpass, for mix checking purposes). At first I didn't think I was gonna be able to do 4, even.
File placement was static.

This was grueling and unpleasant, and gave me my final 'take' on Neil's claims about 192K. I think 24 bit is great but 192K is seriously way in excess of what's needed or wanted, unless we're using noise shaping that cuts down error in the audible band using that samplerate (I have a noise shaper like that, and DSD works that way). Only then is it relevant.

I'm still a little dissatisfied with 44.1K but would probably be fine with even 48K, and I'm not worried about 44.1K anymore if it's handled properly. The problem is more causing it to splatter by making the reconstruction have to stress out the DAC beyond its design limits (Gibb effect, intersample peaking which can be shown using SSL's "X-ISM" meter). If you run this meter and the analog clip lights up on intersample peaks, there's your '44.1K' problem right there. The problem naturally goes away at higher sample rates, but at 44.1K your headroom for super highs is MANY DB down from full scale, whether you know it or not.

Furthermore, ABX testing is still insensitive. I can tell whether I will or won't be able to keep up an observation 20 times in a row. When your heart starts pounding and you have to hold your head in precisely the same place and you're listening painfully loud, it's distracting. ABX sucks. All it can tell us is that, human beings can conclusively tell the difference between two 24/96 files otherwise identical except for the filtering of all >20K information.

Turns out this is a lot harder than it appears.

The reason it appears so easy is that, when pleasure listening, we are constantly getting hints of this reality but they are fugitive. They're not in the same place two times running. To pin them down every time requires that you hear and fix your attention exactly the same, over and over, maddeningly.

ABX sucks, but it has its uses. Please cut it out with the 'there are no audible differences' when there are.

It's not so ironic you have the same opinions I do. I didn't get the ABX testing until yesterday but I was able to back up what I said. Now I don't want to put words in your month but I said I could tell the difference between 24 bit to 16 bit, and 16 bit to MP3. I felt it was going to be harder to tell the 16 bit to the MP3 and the test showed that...slightly. I think the sample rate is less important but I have not yet tried it because I already know.

As you said, these test are pretty grueling and I honestly usually missed the first, getting used to sound and the last. It was much easier to pick out at first. I only tested, and learned to test, because someone asked an I thought it was interesting. However, it's not exactly fun.



On my own thoughts. What is it that having a higher sampler rate than 44.1k, 48k, 96k, or up to 196k is bothersome to people? I understand it creates issues for an engineer, I am strictly talking about consumers. Lets say that you don't want to work that high, it's wasted resources above what you feel is appropriate (I still do 24 bit, 44.1k myself). Is the resources you need the problem, the reliability of your system? IMO, by the time this format takes off, that will be a non issue, but still, I am wondering what the negative is from the consumer side.


I really want to know if people feel that 24 bit, if sample rates stayed the same, or reasonable for your studio, if anyone really thinks this is just a bad idea. I keep hearing people saying it doesn't do anything good but if it's bad, why is it bad? I showed myself I could hear the difference so the difference is there for me, if it's not for others with lesser speakers, lesser converters, is it still bad? Lets say the price stays the same, or within 10%, which I would rather buy 16bit/44.1K waves anyway for 10% more.

What is the real negative? Please enlighten me.
Old 9th April 2014
  #2735
Lives for gear
 

Quote:
Originally Posted by nuthinupmysleeve View Post
Please clarify... choice of converters is not relevant in this discussion about sample rates from what I can see, but maybe I don't get your point?
Real world converters tend to perform better at higher sampling rates. So if you want optimal performance up to 20kHz, as a practical matter using 44.1k sampling is a bad idea. This has nothing to do with Nyquist, the mathematics is fine, it is rather what is practical in the real world.

Conversion has a sweet spot, probably something like 20bit / 60kHz for audio. Which is similar to what some early digital designers actually wanted, but couldn't for various reasons.
Old 10th April 2014
  #2736
Airwindows
 
chrisj's Avatar
Quote:
Originally Posted by bogosort View Post
Because you are saying that you can hear ultrasonic frequencies. If it were actually true, I'd say let's go get you tested by an audiologist and then publish the results! You'd be a celebrity in the psychophysics world.

Please know that I completely grant that it is possible; it's just so very unlikely that I'm going with the simpler hypothesis. But this is your thing -- if you can really hear ultrasonics, seriously, go get tested to make it official and then get in touch with the closest university.
Pshaw, what I did was NOT so exciting or unusual. What would be unusual is to be able to clearly hear that stuff at normal, conversational levels. I had to go really far beyond that to pick up ANY difference. Like I said, fletcher-munson curves: it's uninteresting to consider 'can someone hear 5 hz if it is shaking their entire body violently'. It's all a matter of degree.

I think including 5 hz in high-resolution audio is good too, but happily digital audio will already do that with no trouble at all, so we needn't worry about it. Digital audio CAN also include 24K or indeed 40K with no trouble at all and I absolutely do not understand the passionate revulsion against ever doing this or even talking about it.

I quite sympathise with getting mad over 'only 192K sounds like a sunshiney day!' but meh, advertising. I've had to live with that sort of nonsense over and over, gracefully, even when it directly affected my plugin-writing job. If I have to put up with that stuff when it threatens my livelihood, I've got no sympathy whatever when people are just mad because it threatens… what?

It means nothing to you whether people hear 25K with their ears, or with bone conduction in their skulls or whatever. It's like you're defending the observations of people years ago who were trying to map out human sensory potential, and while this is interesting I just do not see where the sense of offense comes from. I'm particularly concerned when it comes down to ABX testing and the requirement for experiences to be endlessly repeatable and testable, because creative music composition is all about creating SPECIAL moments that jump out and create 'magical' impressions outside our normal experience, which seems like it'd be a natural match for intentionally setting the 'resolution bar' a nice clear margin outside what our sensory experience could possibly be (no matter what circumstances it is, or who we are).

Let's assume that it's obnoxious and unjustified to go 'all digital audio sounds bad unless it is 192K, that's the only kind that is any good'. I share the concern with that statement, as it's wildly overstating the case.

What is wrong with doing, say, 96K or 48K if it's convenient? Does anybody here actually see a moral or intellectual reason for setting the 'resolution bar' used for audio output, to EXACTLY what even the most gifted listener can do on his or her best day?

I don't understand why overkill is wrong. In my plugins, I do internal processing at 64-bit floating point, and I actually noise-shape to the 32-bit CoreAudio floating point buss. I might be able to hear stuff going on at the 16 bit noise floor, I might dither to the 24-bit noise floor on principle and in case that tiny bit of 'vibe' helps somehow (I would never, EVER hear it double blind, the idea is totally insane) but the very idea of 'hearing' what's happening at a 32-bit floating point noise floor is bonkers. But I'm noise shaping it anyway, because it's correct: because I'm throwing out information otherwise. Might add up across a whole mix—maybe.

This is digital. It has a way of sneakily degrading the sound if you let it. WHY do we have to select output resolutions/samplerates only just barely beyond human perception? Why can't we have more of a buffer zone? Bits are cheap, for crying out loud. This should not be such a big deal.

It would be good to shrug off the hyperbole of the wackier Pono claims (96K underwaterness, etc) and just allow it to create an environment where finally we can have an output format that's good and solidly beyond what it absolutely has to be. It would be SO easy. All that has to happen is for Pono to be a big hit, which it is, and then people listening to 48/24 or 96/24 mixes and getting to bathe in the sound unlike our mp3-streaming, overcompressed world.

Loudly demanding that digital audio must run out of steam at 22K where humans mostly do, seems to me like a very unreasonable thing. Heck, if we stretched it up to 40K you could have such nice filter transition bands compared to the insanity of the infinitely-steep brickwall model of the CD. People really have to let go of the technical-limitation-derived decisions of digital audio's past. Processing power and storage space improvements have made that stuff so very obsolete, yet it's being clung to.

We don't NEED to keep it rigidly to human hearing. Bits are cheap.
Old 10th April 2014
  #2737
Lives for gear
 
bogosort's Avatar
Quote:
Originally Posted by chrisj View Post
Pshaw, what I did was NOT so exciting or unusual. What would be unusual is to be able to clearly hear that stuff at normal, conversational levels. I had to go really far beyond that to pick up ANY difference. Like I said, fletcher-munson curves. . .
Except that it is exceedingly unlikely for any adult to hear over 22 kHz at any SPL (provided the air itself doesn't go nonlinear), but you're saying you can. Why are you bringing up the Fletcher-Munson curves? They go asymptotic at 20 kHz; you'd need infinite amplitude to hear beyond it.

Quote:
It means nothing to you whether people hear 25K with their ears, or with bone conduction in their skulls or whatever. It's like you're defending the observations of people years ago who were trying to map out human sensory potential, and while this is interesting I just do not see where the sense of offense comes from.
Do I really appear offended? Funny that. You provided a 20/20 ABX result in which the only difference in the files was ultrasonic. I'm merely trying to understand how this truly extraordinary result came to be.

Quote:
What is wrong with doing, say, 96K or 48K if it's convenient? Does anybody here actually see a moral or intellectual reason for setting the 'resolution bar' used for audio output, to EXACTLY what even the most gifted listener can do on his or her best day?
I'm on record in this very thread for stating that I'd be fine with a 24/48 delivery format; it's the standard for our cousins in the film industry, it covers all the outliers, and it's easier to filter than 44.1k.

But as an audio engineer I find it appalling that some would have us double the delivery bandwidth just because we can. It literally serves no one except the snake oil salesmen. Worse still, we'd have to waste our processing power and storage on useless ultrasonic energy. How can you think this is reasonable?

Quote:
I don't understand why overkill is wrong. In my plugins, I do internal processing at 64-bit floating point, and I actually noise-shape to the 32-bit CoreAudio floating point buss.
Obviously there is an enormous difference between process calculations and final storage. It is not overkill to use double precision in DSP; but even you must agree that it would be beyond foolish to store our audio in 64-bit floats. Right? Overkill isn't always good, eh?

The job of an engineer is to know when it is overkill and when it is not.
Old 10th April 2014
  #2738
j_j
Lives for gear
Quote:
Originally Posted by mellotronic View Post
One way is encoding by 1 & 0's and the other is %100 linear.
What's 100% linear? LP? Not even remotely close. Tape? Maybe slightly better, but with jitter that makes any modern digital system look rock solid.

What do you imagine is going on, anyhow?

Quote:
Most people can hear and "feel" the difference. If you can't then perhaps you would have been better suited to have been a scientist or mathematician?
"most people" have failed miserably in reasonable listening tests. Of course, yes, we can all identify analog from the distortion and noise issues. I hope you're not suggesting that's the improvement?

I really don't get the point of your discussion here, you're simply insulting people, and you're dismissing science entirely. Audio is science. Science brought you every form of audio from tin-can and string to modern 192/24, SACD, DSD, MP3, AAC, AC3, what-have-you.

So whatever do you think you're saying here?

As to "hear" and "feel", well, no, that's not what "most people" can do. Please don't make these absurd, disproven claims as part of an insult to the people who brought you every bit of audio transmission, audio, and playback, it's simply not constructive, and it is, frankly, a ridiculous behavior that does not further discussion.
Old 10th April 2014
  #2739
j_j
Lives for gear
Quote:
Originally Posted by ezraz View Post
if it can be proven it can be proven in black and white. but the second you need something as convoluted as a musical ABX test the data goes to ****e. i doubt your application of them to my ears. and i have that right to point out the flaws.
ABX, done right, is definitive. It is not "convoluted", and the data from it is not "****e".

You can doubt all you want, but rejecting reality is not something that results in a good long-term result, now, is it?
Old 10th April 2014
  #2740
j_j
Lives for gear
Quote:
Originally Posted by chrisj View Post
All it is, ALL it is, is a massive highpass coded in the most primitive way. You get only the slew value between pairs of samples, which translates to 'loads of treble and nothing else'. I did it because I was getting nervous about hearing the difference between 24/96 and filtered 24/96.
You're using a 2-tap pre-emphasis filter? You're cutting most of the audible audio signal out, boosting the rest, and putting it into transducers that you haven't validated?

Sorry, you need to test every bit of your system to make sure you're just not hearing in-band IM.

A lot of "I can hear ultrasonic" tests have flopped on that very issue. I don't know that's happening here, but it's a possibility.

By the way, signal to signal difference is indeed very primitive, you might consider looking into a free filter design program and learn more about how to do filtering.
Old 10th April 2014
  #2741
j_j
Lives for gear
Quote:
Originally Posted by theblue1 View Post
Of But I think it's safe to say that most of us 'science types' had already stipulated that, if you turn the playback level WAY up and focus on low level signals like reverb and other tails, then, for sure, given the right material, it is, indeed, going to be relatively trivial for those with a reasonable amount of ear training to differentiate the CD-quality from the 24/96.
Especially when most of the low frequencies are eliminated and high frequencies are boosted by 6dB.
Old 10th April 2014
  #2742
j_j
Lives for gear
Quote:
Originally Posted by chrisj View Post
But that's obvious. I did it by making the ultrasonic information VERY LOUD. Fletcher-Munson curves, you know? If I hadn't, I wouldn't have heard a difference. I didn't hear a difference on most of the examples: only solo cymbals put out enough 'air' to do it.

Wow, where to start?

First, "loud" refers strictly to perception. Loudness is sensation level. SPL is not loudness, SPL is the analytic, measured level.

Second, your difference filter only provides a maximum boost of 6dB at FS/2, no more, and that only if you didn't normalize the filter, too.

What it mostly does is remove most of the actual audio energy, allowing you to turn the level way, way, way up. Now, I can plot the frequency response of your difference filter quite easily, maybe I should plot it here for your own reference.

There, it's uploaded, and attached here.

What do we see?, well, we see that it already has 3dB of gain at 22.05kHz, for instance, and is above zero from thereabouts of 16kHz. On the other hand, in the lower part of the spectrum, it's reducing the energy below 4kHz by a factor of at least 20dB.

What does that mean, since even with percussion, a lot of the energy is there? It means that you've allowed yourself to boost the gain by close to 20dB.

Now you can: 1) listen to the noise floor, which is pre-emphasized like everything else, and boosted by 20dB or so. 2) put all the ultrasonic energy you've got into your electronics and transducers at 26dB higher than normal.

Could you hear a difference? Maybe. Does it mean that you're hearing ultrasonics? I doubt it. You should hear the noise floor in silences, for sure, all by its lonesome, and if you're not, something is wrong.

But, does putting more than a 20dB spectral tilt on an audio signal mean anything sensible?

I don't think so.
Attached Thumbnails
Launch of Pono-ble.jpg  
Old 10th April 2014
  #2743
j_j
Lives for gear
Quote:
Originally Posted by ezraz View Post
You are focusing on the X axis and the ticks going from 44100 to 48000 to 96000. That's barely more than doubled.

Focus on the Y axis - the bit depth. The count of those ticks goes from 16000 to 15000000.

Methinks my ears might make out the 14,984,000 more samples on the Y axis.

Also, if the X axis correlates to the changes in frequency, does the Y axis correlate to the changes in timbre? i used to think of bit depth as raw data rate but i now think of it as the Y axis, and the hidden world of soundwave blending.
Ok, that's not even wrong.

More "samples" doesn't get you anything but wider frequency response, if you need it.

Now, as to resolution, what's this 16000? It's 65536 levels for 16 bit, 16777216 for 24 bit, to start with, and all that does is lower the noise floor. Your internalization of "ticks" really isn't helping your understanding.

Sampling is "sampling rate", and determines maximum bandwidth.

Quantization (amplitude) determines peak level to noise floor, and has nothing to do with bandwidth.
Old 10th April 2014
  #2744
j_j
Lives for gear
Quote:
Originally Posted by chrisj View Post
I have this feeling that no matter what I do or say, you're going to say 'but it would be miraculous and impossible for you to have heard this, therefore there has to be another explanation'.
Would you mind stopping these false, frankly insulting fraudulent versions of what people are telling you and start listening.

I've explained in another post some of what you've done, and I'm rather bothered with this filtered version that you didn't hear the change in noise floor every time, frankly. I think that, alone, constitutes a failure of what should have been a positive control in your experiment. And that's just for starters.

We don't know anything about the linearity of your headphones, etc, either, and it is most certainly necessary to know that, it's something that other tests have tripped over, and fallen hard upon.

You need to validate your test. Do you understand that? It's not as simple as you want it to be. Running tests is NEVER as simple as you want it to be.
Old 10th April 2014
  #2745
j_j
Lives for gear
Quote:
Originally Posted by johnnyc View Post
Real world converters tend to perform better at higher sampling rates. So if you want optimal performance up to 20kHz, as a practical matter using 44.1k sampling is a bad idea.
Um, what? That's exactly opposite the physics of the matter.

Now if you mean that the filtering in front of the convertor is easier to make, ok, I can buy that part.
Old 10th April 2014
  #2746
Gear Head
 
bandpass's Avatar
 

To be fair, Chris applied the filter only in an "ear training" session; the original sample # 4 was then ABXed unmodified, albeit at elevated levels.

I think we're good here. We suspected that Neil's claims were, shall we say, somewhat exaggerated, and we've found that this bears out in practice. At the very least, it's frickin hard to hear a difference: the "under water" stuff is mendacious. Thanks to all, esp. Chris for almost killing himself in the process of getting there!
Old 10th April 2014
  #2747
j_j
Lives for gear
Quote:
Originally Posted by bandpass View Post
To be fair, Chris applied the filter only in an "ear training" session; the original sample # 4 was then ABXed unmodified, albeit at elevated levels.

I think we're good here. We suspected that Neil's claims were, shall we say, somewhat exaggerated, and we've found that this bears out in practice. At the very least, it's frickin hard to hear a difference: the "under water" stuff is mendacious. Thanks to all, esp. Chris for almost killing himself in the process of getting there!
I'm still very puzzled why, with 20dB of boost, why noise floor didn't just leap out.

Also, when you can hear an effect only in one set of equipment, it's wise to first check that equipment. I've had some very interesting experiences that way...
Old 10th April 2014
  #2748
Gear Head
 
bandpass's Avatar
 

Quote:
Originally Posted by j_j View Post
I'm still very puzzled why, with 20dB of boost, why noise floor didn't just leap out.
Dither? The original and filtered samples are in Chris's earlier post, if you want to check.
Old 10th April 2014
  #2749
Airwindows
 
chrisj's Avatar
Yeah, the tone of this really collapsed. It's too bad, and I do not feel it's the fault of the 'Pono guys'.

Happily, the thing that will save us is exactly the fraudulent hypey exaggerations the Pono marketing guys have put forth.

They'll make a player that can play 192/24 (yay technology). What we NEED is only a player that can do 48/24, if that, to cover real-world critical listening of almost any sort (96/24 absolute max, probably only 48/24 would do in practice if handled right)

Pono is already overfunded to the tune of millions of dollars, will come out with content in excess of 48/24, people will listen to that and hear that it's better than mp3s (now THAT is really a trivial ABX unlike 96 vs 44.1). They may never know that 48K is as good as 192K treble-wise (I do wonder whether there are timing concerns: but the sampling theorem also covers all PHASE conditions of all frequencies, so timing should be also covered) but you know what? They'll probably be up for listening to any damn thing including CDs themselves, made into FLAC (and you'll fit a lot more of those onto a Pono player than the HD tracks at 192K)

If there are implementation issues about 'turning abstract theorems into working audio equipment' (a VERY valid concern too lightly glossed over by theorists) the hardware being designed to handle 192K should overdesign things enough that the real audible band will truly be awesome.

If 44.1/16 is really so acceptable, guess what? People's Ponos are going to be LOADED with CD audio, and it's completely okay if people consider them 'slightly less resolution' than the HD because they ARE slightly less, while still being good enough nearly all of the time. And Pono will make them sound great. When was the last time you listened to CD audio through a serious playback chain? I was startled at how good it was when I ran my NAD 515BEE CD player's digital out directly into my Lavry Black. Same deal with Pono, people will be able to hear their CDs 'as intended'. Playback off consumer laptops and boomboxes is not the same as use of a serious analog stage.

People will be doing all this listening regardless of what the 'theorists' say. The money rolling into Pono is a story on its own and creates its own momentum and people WILL check out the 'different' sound, which is not that different in the abstract, and which damn well is going to sound different to people used to cheap consumer audio. Today, true 'high end' sonics IS completely accessible for a few hundred bucks thanks to technology esp. digital technology's escalation. You wouldn't have even been able to get the off-the-shelf chips Pono uses, twenty years ago. It wasn't available at any price. Now all you need to do is take a bit of trouble to redesign the cheap-ass analog stages and the DIGITAL side is ready to support a really high level of performance, mass produced and no more expensive than any other chip. Your bottleneck is the analog stage. Pono is already dealing with that.

Go ahead and vent your outrage if these things outrage you. It won't matter. Pono will change the face of audio when people finally hear, in practice, what digital audio was SUPPOSED to be all along. (and what it could, theoretically, have been all this time)

Old 10th April 2014
  #2750
mixmixmix
Guest
Pono will not change anything. Another pointless format which will quickly fade into oblivion. But it serves as a great catalist for Gearslutz threads. Thank you, Neil.
Old 10th April 2014
  #2751
Gear Guru
 
Kenny Gioia's Avatar
 

Quote:
Originally Posted by chrisj View Post
If 44.1/16 is really so acceptable, guess what? People's Ponos are going to be LOADED with CD audio, and it's completely okay if people consider them 'slightly less resolution' than the HD because they ARE slightly less, while still being good enough nearly all of the time. And Pono will make them sound great.
Who is this listener that is going to be "sold" on Pono's B.S. marketing and fill their players with CD audio?

Or are you saying that it will be CD audio labeled as 96k or better?

Correct me if I'm wrong, smart phones will play back CD audio.

Quote:
Originally Posted by chrisj View Post
Go ahead and vent your outrage if these things outrage you. It won't matter. Pono will change the face of audio when people finally hear, in practice, what digital audio was SUPPOSED to be all along. (and what it could, theoretically, have been all this time)

People are finally going to hear what CD audio sounds like and you're applauding this device?

Even though the marketing says that CD audio sounds like balls?

And this device, that was designed to fix that problem, is now going to let you hear what you think sucks sound great?

My brain just hurt writing that sentence.
Old 10th April 2014
  #2752
Gear Addict
 

Quote:
Originally Posted by chrisj View Post
What is wrong with doing, say, 96K or 48K if it's convenient?
What I took away from this article (http://people.xiph.org/~xiphmont/demo/neil-young.html) and this video (https://www.xiph.org/video/vid2.shtml) is that the main argument against higher sampling rates for a consumer format is that it opens up the possibility of intermodulation distortion in the audible range from ultrasonic frequencies... band-limiting just eliminates the entire issue altogether. Seems pretty logical to me.

IMO the difference between MP3/AAC and WAV/AIFF/FLAC is pretty huge and I'd love to see things move towards lossless files as standard... 24/96 or 24/192 as a consumer format is just ridiculous IMO, it's a very marketable idea but ultimately those articles above set out that there's science behind why 16/44.1 is perfectly good enough as a format.

If there's room for portable music players to be improved, there's probably more to be gained from improving converters and/or the analogue stages
Old 10th April 2014
  #2753
Quote:
Originally Posted by mixmixmix View Post
Pono will not change anything. Another pointless format which will quickly fade into oblivion. But it serves as a great catalist for Gearslutz threads. Thank you, Neil.
But-- a great consolation it will be an oblivion with beautiful clarity and definition, rather than a murky and fuzzy one.
Old 10th April 2014
  #2754
Gear Guru
 
Kenny Gioia's Avatar
 

Quote:
Originally Posted by joelpatterson View Post
But-- a great consolation it will be an oblivion with beautiful clarity and definition, rather than a murky and fuzzy one.


I'm personally excited to hear the music I produced like I never heard it before.

The way it was meant to sound. Not the way I recorded it.

I have a Pono boner. Or a Poner.
Old 10th April 2014
  #2755
Quote:
Originally Posted by Kenny Gioia View Post
I have a Pono boner. Or a Poner.
Giggle.
Old 10th April 2014
  #2756
Airwindows
 
chrisj's Avatar
Quote:
Originally Posted by Kenny Gioia View Post


I'm personally excited to hear the music I produced like I never heard it before.

The way it was meant to sound. Not the way I recorded it.

I have a Pono boner. Or a Poner.
Don't laugh too loud, digital has mysterious ways.

I have a playback app, "Decibel", which forces the audio hardware to be at the correct samplerate for the track playing, and feeds it audio data at a lower level, controlling the DAC buffer and not sharing it with other apps. When I play back a mix on that, it sounds better than it did played out of Logic with all the plugins chugging away and the DAW busily keeping track of everything. I'm guessing this is down to the Lavry having to work less hard to reclock the data coming in, though Logic ain't bad at keeping things going. It's just that the dedicated clock-hogging playback gives you a little bit more of everything.

You COULD end up hearing the music you recorded, sounding better than it did while you were recording/mixing it. This would be impossible in analog, but in digital as long as you are 100% ITB, any timing errors introduced by the clock while you're just MIXING will cause your monitoring to be less good, while the actual file is still fine (you can bounce it high speed or do numerous things, it's not dependent on the audio quality of your monitoring playback). Then if you play the result on a Pono or some other high-performance digital audio player, you hear what was actually there.

When actually tracking you'll want the clock of your actual ADC to be good and steady, but once the audio is data, it's safe: if your recording chain jitters up the clock after that point, you could be monitoring the result and have it sound bad, but only while monitoring. Later playback would give you the 'good' original recording, from the 'good' ADC.

This is actually a strength of the digital studio. I discovered it working with some really crappy digital signal chains, but it holds for things like Pono, too. Any really good playback will do it.

How's your poner now?
Old 10th April 2014
  #2757
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by chrisj View Post
This was grueling and unpleasant, and gave me my final 'take' on Neil's claims about 192K.
I really don't see how this test has any relation to anything except the filter used. I haven't listened to the files but I have opened them in Audition. Just by looking at the frequency analyser, you can see visual differences in your files compared to the originals all the way down to below 10 Khz when switching between the two. That means that the filter filter affects the signal significantly all the way down to below 10Khz. No wonder you can hear a difference!

As a comparison, I converted the original files with SoX down to 48 Khz and back up to 96 Khz and when switching the frequency analyser graphs between the originals and converted files, there is no visual difference at all below the cut-off point of about 22.8Khz.

All you have demonstrated is that the filter used is not good enough. Try converting the files to 48 Khz using SoX in very high quality setting and then back up to 96 Khz. Then try the same ABX test again to see if you can hear a difference with the originals.

Alistair
Old 10th April 2014
  #2758
Gear Guru
 

Quote:
Originally Posted by chrisj View Post
If 44.1/16 is really so acceptable, guess what? People's Ponos are going to be LOADED with CD audio, and it's completely okay if people consider them 'slightly less resolution' than the HD because they ARE slightly less, while still being good enough nearly all of the time.
"slightly less"?? "good enough"? You and I know they are slightly less and good enough. But people are being told to buy a Pono on the basis that CD=mp3=crap. These promotional videos constantly, disingenuously LUMP mp3s and CDs together into the same category. These differences (where the biggest distinctions occur) are minimized. Both are still underwater.

According to the Pono Chart, Even 96k is underwater!! My player RIGHT NOW can be loaded with 16/44.1 content. In fact, there IS uncompressed content in there right now. According to Pono, it is not 'good enough' When you reach 192kHz, you are merely floating on the surface and not until you experience 384kHz are you flying on Angel's Wings/toting dual machine guns.

Quote:
And Pono will make them sound great. When was the last time you listened to CD audio through a serious playback chain?
yesterday

Quote:
Playback off consumer laptops and boomboxes is not the same as use of a serious analog stage.
I just despair of your conviction that owning a Pono will motivate these people to plug their 'serious analog stage' into a real amp and speakers. There is absolutely no reason why this must follow. None whatsoever.

We all know they will listen to it in their car and on their Beats headphones - AT BEST. More likely on some Skull Candy earbuds or one of those Dock pieces of ****.

Quote:
and which damn well is going to sound different to people used to cheap consumer audio.
not if they plug their Pono into cheap consumer audio. How can you even dream that they won't? Isn't this the most unrealistic kind of wishful thinking? Pushing for better mixes, pushing back on the loudness wars, pushing for better speakers, even pushing for the concept of "an hour a day" of Sit-Down Listening are more meaningful efforts. And THEY are Pie In The Sky, too!

Believing that Pono will contribute to any of these is a pipe dream. A new player will not change the brain wiring of the GenXers and the Millennials who can't sit still to save their lives. These are the people who pay $100 for a ticket to a broadway show and text through the whole thing. These are the people who TEXT while driving. A Pono is not going to MAKE them 'focus on the music'.


Quote:
Your bottleneck is the analog stage.
I agree. Does the Pono have +4 Balanced XLR outs?

Quote:
Pono is already dealing with that.
How? By using the chips from that company that got caught red-handed in that Oppo scam? It's a 5" box that fits (uncomfortably) in your pocket... did they cram a Krell amp and a pair of B&W 802's in there? THAT's your analog stage!

Quote:
Pono will change the face of audio when people finally hear, in practice, what digital audio was SUPPOSED to be all along
I think Pono will change the face of nothing. Pure Pollyanna-ism. People can hear that same thing NOW. Even the high res part - they can hear it NOW. Why haven't things "changed" already?
Old 10th April 2014
  #2759
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by chrisj View Post
"it is relatively trivial for good listeners to distinguish 100% between CD quality and 24/96 if they try"
This has still not been demonstrated. The assertion is still only that. It isn't a fact.

Alistair
Old 10th April 2014
  #2760
j_j
Lives for gear
Quote:
Originally Posted by bandpass View Post
Dither? The original and filtered samples are in Chris's earlier post, if you want to check.
Dither does not eliminate the noise floor. Not sure what you are thinking there.
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