If you want body shaking bass and don't mind looking an idiot in public there is this What is the SUBPAC? - SUBPAC.com
Well there are many possible reasons for that. I did do a load of handset quality testing on phones some years back and the number of codec options is mindboggling. For example, iphone 6 and beyond finally supports HD calling so when my daughters call me on my clanky blackberry which also supports it I am thrown by actually having HF available.
but onto landlines, this will be down to the network operator. Landline terminated calls are limited to 16kbs ADPCM as that is the standard. It's possible that, in optimising for higher quality something was lost and a cludge was needed in the transcoder. Iphones are notorious for requiring a whole bunch of network patches before they will work. Anyone other than apple would have been asked to sling their hooks, but the networks are not able to do that with the mighty shiny. The fact that a mobile voice call works at all is still a minor miracle given the sheer complexity behind it!
None of which has anything to do with the DAC in the phone of course.
but onto landlines, this will be down to the network operator. Landline terminated calls are limited to 16kbs ADPCM as that is the standard. It's possible that, in optimising for higher quality something was lost and a cludge was needed in the transcoder. Iphones are notorious for requiring a whole bunch of network patches before they will work. Anyone other than apple would have been asked to sling their hooks, but the networks are not able to do that with the mighty shiny. The fact that a mobile voice call works at all is still a minor miracle given the sheer complexity behind it!
None of which has anything to do with the DAC in the phone of course.
Well, the total noise sum of the vinyl medium is certainly not dead flat white and can effectively provide dithering for 16bit recoding but not for 8bit
Here is an example.
Initial vinyl rip at 192KHz/24bit.
For the 16bit and the 8bit files, dither is of triangular distribution and noise shape is the most effective from the three noise shapes in Steinberg Wavelab.
Don’t dismiss listening to the 8 bit files. They are the most exposing of the audible effect of dithering and noise shaping. The visual overview of the waveform skirt will horrify you, therefore do the listening before looking.
(1) https://www.dropbox.com/s/jjxgljotxuhgtrp/%281%29%20192k-24b.wav?dl=0
(2) https://www.dropbox.com/s/yongxz7cf0xyavx/%282%29%2096k-24b.wav?dl=0
(3) https://www.dropbox.com/s/c38v847j0byxz10/%283%29%2044.1k-24b.wav?dl=0
(4) https://www.dropbox.com/s/5iqgea9qb3vbcud/%284%29%2044.1k-16b-non%20dithered.wav?dl=0
(5) https://www.dropbox.com/s/f23zvh7c3ci1kad/%285%29%2044.1k-16b-dithered%20Noise%20type%202%20Noise%20shaping%20None.wav?dl=0
(6) https://www.dropbox.com/s/qjrzxe5g0f57085/%286%29%2044.1k-16b-dithered%20Noise%20type%202%20Noise%20shaping%203.wav?dl=0
(7) https://www.dropbox.com/s/z7jvflghmelqekg/%287%29%2044.1k-8b-non%20dithered.wav?dl=0
(8) https://www.dropbox.com/s/c4yw10yu4exij0k/%288%29%2044.1k-8b-dithered%20Noise%20type%202%20Noise%20shaping%20None.wav?dl=0
(9) https://www.dropbox.com/s/qvff6j5ll8ujldm/%289%29%2044.1k-8b-dithered%20Noise%20type%202%20Noise%20shaping%203.wav?dl=0
George
Yes, you may be right. I think it may be better than some of the old DACs people still have around. The stock iphone earbuds are not good, and and a lot of the audio content available has been lossy compressed, but otherwise maybe okay.
George, Thank you for that. Sometimes people may refer to CDs or some DSP as grainy or brittle. They are typically referring to much more subtle versions of the sound quality effects evident in the 8-bit files.
If we could go on to find out some way to demonstrate the distortion or effect people tend to refer to as "smeared," that might be also helpful. Then we could have some shared understanding of that particular mental perceptual experience as well.
I think the common perception is grainy with an unpleasant high end. Somehow the word "smeared" doesn't suggest grainy. I would like to identify something that is consistently "smeared" even if its a loudspeaker. The word suggests all the HF is present but it is stretched in time. However except for speakers that doesn't happen. The other possibility is that the HF content is shifted from the LF content. This would be the "dispersion" that MQA talks about. I still have never actually seen evidence in a digital system. maybe I'm guilty of "listening with my eyes".
Dielectric absorption, latency, L&R converted to serial data and back again. Time domain stretching abounds.
Hi George, thanks for going to the effort, noble effort, however.....
I fully agree that the 24bit 192k and 96k files sound clearer, more open, more friendly, more real etc than the other files no question.
Both these files cause me to question the dust/scratch behaviour of your phono preamp stage perhaps, or is it the arm, or is it the ADC...both files have a particular record dust scratch impulse/reverb/echo/stretch characteristic, with file 2 longer/worse/different.
The differences of the first 10 seconds dust noises impulses on the first two tracks are subtly striking, and track 3 more so, but of course this has the extra processing of SRC.
Anyway, a confounder here in this experiment is the resampling down to 44k and this is where extra junk gets caused, extra obscuring/discordant junk that changes the experiment.
Studio guys I speak with, talk nowadays about tracking recording and virtual mixing/processing at 88.2k for CD releases.....no numerically awkward down SRC causing extra subjectively wrong artifacts.
All of your 44k files do indeed readily sound different, the non dithered examples actually sounding better to my ear, including the 8bit non dithered wave files despite the savage nicely white hiss.
The reason is change in impulse response attack character and change in post impulse echos/reverbs/tails and consequent subjetive reduction in instantaneous dynamics and subjective empahasis of tails.
The ear can willingly 'hear through' static (white) noise, but when the white noise foundation/reference keeps changing and the intermods keep changing the ear/brain fatigues and goes into ignore fine detail mode and pretty quickly, this is why I am happier with correctly white noise.
Thanks George, this reinforces my previous findings that I prefer no dither/no noise shaping, and that I readily hear different noise shapings like Panasonic MASH, Sony SBM etc.
Dithering/noise shaping is a devious magic trick, it gives better measured (sine wave) specs at the cost of shifting impulses in time and consequent comb mixing in time and amplitude responses.
All modern AV production is done at 48k(x) and is technically/measured better sounding than 44k.
Is this the death knell of 44k into the future ?.
Tesla had good reasons for specifying 60Hz mains, is there a similar situation here regarding audio sampling rates ?.
Dan.
No wave editor files were viewed.
I fully agree that the 24bit 192k and 96k files sound clearer, more open, more friendly, more real etc than the other files no question.
Both these files cause me to question the dust/scratch behaviour of your phono preamp stage perhaps, or is it the arm, or is it the ADC...both files have a particular record dust scratch impulse/reverb/echo/stretch characteristic, with file 2 longer/worse/different.
The differences of the first 10 seconds dust noises impulses on the first two tracks are subtly striking, and track 3 more so, but of course this has the extra processing of SRC.
Anyway, a confounder here in this experiment is the resampling down to 44k and this is where extra junk gets caused, extra obscuring/discordant junk that changes the experiment.
Studio guys I speak with, talk nowadays about tracking recording and virtual mixing/processing at 88.2k for CD releases.....no numerically awkward down SRC causing extra subjectively wrong artifacts.
All of your 44k files do indeed readily sound different, the non dithered examples actually sounding better to my ear, including the 8bit non dithered wave files despite the savage nicely white hiss.
The reason is change in impulse response attack character and change in post impulse echos/reverbs/tails and consequent subjetive reduction in instantaneous dynamics and subjective empahasis of tails.
The ear can willingly 'hear through' static (white) noise, but when the white noise foundation/reference keeps changing and the intermods keep changing the ear/brain fatigues and goes into ignore fine detail mode and pretty quickly, this is why I am happier with correctly white noise.
Thanks George, this reinforces my previous findings that I prefer no dither/no noise shaping, and that I readily hear different noise shapings like Panasonic MASH, Sony SBM etc.
Dithering/noise shaping is a devious magic trick, it gives better measured (sine wave) specs at the cost of shifting impulses in time and consequent comb mixing in time and amplitude responses.
All modern AV production is done at 48k(x) and is technically/measured better sounding than 44k.
Is this the death knell of 44k into the future ?.
Tesla had good reasons for specifying 60Hz mains, is there a similar situation here regarding audio sampling rates ?.
Dan.
No wave editor files were viewed.
Last edited:
Just because the mental experience of hearing a sound is perceived as "smeared," that doesn't necessarily mean there is some time delay, although there very well might be. For one consideration, it doesn't sound anything like a guitar effects delay pedal, not to me anyway. It's not exactly a delay sound, it's more of a smeared sound. Doesn't sound like latency either, again, to me.
Therefore, thinking of things that might cause time delays in some way or other as possible causes remains speculative. However, such potential causes might be fruitful areas for investigation into a cause.
Also, its still not clear if "smeared" can be demonstrated in any way as clearly as "grainy" can be demonstrated using 8-bit audio files.
Therefore, thinking of things that might cause time delays in some way or other as possible causes remains speculative. However, such potential causes might be fruitful areas for investigation into a cause.
Also, its still not clear if "smeared" can be demonstrated in any way as clearly as "grainy" can be demonstrated using 8-bit audio files.
The 'smeared' sound is due to recursive flitering in DSP causing time shifting of error terms which then get further recursed until ultimate decay.
Change in signal impulse/decay drives change in system impulse/decay current dependent 1/f noise which causes noisy change in operating points causing noisy changes in PIM and IMD distortions....readily subjectively discriminated.
I would describe the increased decay/echos of the noise shaped files as 'time smearing', all caused in DSP.
Dan.
I find it interesting that people are talking about 'smearing' in digital with respect to an electromagnetic system that is being hit with enormous impulses. If there are no clipping events in the electronics then the tankslapper the cantilever is going through should swamp anything else.
The Holman paper shows a very interesting graph of frequency vs response along with the mechanical mode of the cartridge operation across the audio band. Up to about 1kHz, output is due to groove amplitude, 1kHz to about 3kHz (IIRC) the output is velocity driven, and above that, its acceleration driven. Intuitively this all makes sense if you think about the poles and zeroes (integration and differentiation) in the RIAA response.
Can someone explain to me what 'smearing' sounds like? Loss of high's? sibilance?
Can someone explain to me what 'smearing' sounds like? Loss of high's? sibilance?
One possible reason that valves sound good is very high input impedances. When trying to get good sound from Phones and Laptop style devices I feel this can help. If a heaphone op amp it will almost certainly be in true class A if a high impedance. I very often use sources like this when wanting to show something. I have Laurie Anderson on my cheap Nokia that sounds stunning, Oh Superman.
Why list a number of things which do not do anything harmful?morinix said:
DA has a very minor effect on LF frequency response, mainly affecting subsonics. Certainly much smaller than any speaker/room issues.
Latency refers to the time delay before something happens. It will be the same for all frequencies, so no harm done.
Conversion from parallel to serial data and back again is an example of a cause of latency. Not harmful.
Wait a minute: you were trying to write poetry but I foolishly thought it was prose?
I think it is an easy quest. One can have higher input resistance than is common without too much hiss. Worth a try as even the humble LM324 is class A into high Z ( check with a scope, it's very bad when class B ) . This is the area where it costs nothing to give it the benefit of the doubt. All I was saying was don't discount this as something you might try. This idea is specifically were we would connect to power op amp usually used for headphones.
Dan and Mark thank you for your comments.
I hope all of you have better hearing capabilities than mine.
I can’t hear the effects of the sample rate conversion (the ‘echo’ that Dan says he hears). Nor do I hear the added dither and/or noise shaping over the music content on the 16bit files.
But I do hear the effect on the 8 bit files (success).
The inherent noise from the vinyl recording is not enough to overcome the digitization distortion at 8 bits.
This is obvious at the low level music on the 8bit undithered file (file 7). Notice particularly the section 8sec-12sec, where the digitization distortion is heard as if the cartridge loses contact with the groove side walls (I was monitoring the section with the bit meter and this distortion occurred with the signal level filling the 7th bit).
The dither of course cures this (file 8).
And the high level of added noise with the noise shape 3 (file 9) makes (to my ears) the perceived noise much less noticeable.
I am about to repeat the same exercise, this time with an HD recording as the source.
Attachments refer to the prevously linked vinyl rip files
George
I hope all of you have better hearing capabilities than mine.
I can’t hear the effects of the sample rate conversion (the ‘echo’ that Dan says he hears). Nor do I hear the added dither and/or noise shaping over the music content on the 16bit files.
But I do hear the effect on the 8 bit files (success
). The inherent noise from the vinyl recording is not enough to overcome the digitization distortion at 8 bits.
This is obvious at the low level music on the 8bit undithered file (file 7). Notice particularly the section 8sec-12sec, where the digitization distortion is heard as if the cartridge loses contact with the groove side walls (I was monitoring the section with the bit meter and this distortion occurred with the signal level filling the 7th bit).
The dither of course cures this (file 8).
And the high level of added noise with the noise shape 3 (file 9) makes (to my ears) the perceived noise much less noticeable.
I am about to repeat the same exercise, this time with an HD recording as the source.
Attachments refer to the prevously linked vinyl rip files
George
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.