Guest Eggcup The Daft Posted January 14, 2016 Posted January 14, 2016 Re 16 bit distortion - I've had this conversation previously. I was first put on to this reading HiFi World - Noel Keywood said a few times that this is the case, and went as far as saying that there was no point performing the measurement at 16bit becajuse it was always at least 2% and it made products look bad(!). Now, and especially after what @@Newman said when I raised this before, I don't buy into his conclusions - but I have no doubt that he is performing the measurements concerned competently, and that his figures are THD and not THD+N. Anyway, after this I went away and read more widely into both spec sheets and other independent measurements, and they do seem to tally with this message. I haven't checked back (writing this at work) but I remember that Newman didn't think the distortion would be audible... but as I understand it, this playback distortion is still higher than the distortion that should arise from format conversions and other issues being discussed in this thread. Keywood also seems to believe that the noise floor is relevant, from his recently published reviews of portable digital players. I don't buy that. This... That's correct. If you re-quantise the audio to 16bits.... The noisefloor will rise to about -96dB. If you are making music ..... it is possible to put something low in the mix, and accidentally lose it's quitest detail into the noise floor ..... and once you've done that, you can raise it's level later (you've lost the stuff which went below the noise) ..... and so, you don't want to work in 16bits. If you are playing back music.... a typical quiet room, might have just audible sounds of 40dB. So if you put the quietest sound available with 16bit (conservatively 96dB below maximum), so it played back at a volume of 40dB (ie. it is JUST audible on it's own) ..... then the maximum volume level will be at 136dB. Clearly that is insane. Makes much more sense, thanks. Though I'd say 40db range in a quiet room is an underestimate, if not by much.
Newman Posted January 14, 2016 Posted January 14, 2016 Don't drag me into this! I'm scratching my head trying to recall my conversation with you.... was it about the claim that "CD quality has high distortion at low volumes"...?
davewantsmoore Posted January 14, 2016 Posted January 14, 2016 more sense Sometimes these things can be quite difficult to explain, without being too word, brief, or accidentally giving people the wrong idea about something :-S
davewantsmoore Posted January 14, 2016 Posted January 14, 2016 Re 16 bit distortion - I've had this conversation previously. I was first put on to this reading HiFi World - Noel Keywood said a few times that this is the case, and went as far as saying that there was no point performing the measurement at 16bit becajuse it was always at least 2% and it made products look bad(!) Never read a hifi mag for anything but entertainment purposes, but He's either been taken out of context ... making a generalisation of a specific case ... or talking rubbish. Perhaps a triple play. At the risk of sounding like 'vague hand-wavey excuses' .... One potential issue is, things have come a long way in the last couple of decades (so it will depend on when he's referring to) .... this (the march of time) has nothing to do with 16 v 24 bit (sampling been around a century) .... but how we build devices like a DAC. He may be talking about DACs with very poor linearity (which affects the THD measures). Certainly modern converters don't behave like this.... anyone with a computer that has a line in and a line out, can test this for themselves (there's no need for 'debate')
Guest Eggcup The Daft Posted January 15, 2016 Posted January 15, 2016 Never read a hifi mag for anything but entertainment purposes, but He's either been taken out of context ... making a generalisation of a specific case ... or talking rubbish. Perhaps a triple play. At the risk of sounding like 'vague hand-wavey excuses' .... One potential issue is, things have come a long way in the last couple of decades (so it will depend on when he's referring to) .... this (the march of time) has nothing to do with 16 v 24 bit (sampling been around a century) .... but how we build devices like a DAC. He may be talking about DACs with very poor linearity (which affects the THD measures). Certainly modern converters don't behave like this.... anyone with a computer that has a line in and a line out, can test this for themselves (there's no need for 'debate') He has made these claims in 2014 and 2015, and about devices using modern chips like the ESS Sabre range. He clearly says "all 16 bit conversion has these levels of distortion". He is unequivocal about it. And whatever we might think about his agenda or conclusions, he has done the measurements and I see no need to doubt his results. I have a line in and line out on my computer (I hope - we had a lightening strike yesterday and at least my DSL modem got fried). How do I do the test for myself? Don't drag me into this! I'm scratching my head trying to recall my conversation with you.... was it about the claim that "CD quality has high distortion at low volumes"...? I think your relevant comment was that the 16 bit playback distortion referred to would not be audible, and that Keywood was referring to it as part of his subjectivist agenda. You were pretty convincing! I'd still prefer to see a 24 bit standard for modern playback, partly because of this matter, even so: I still buy CDs, and even from shops, so it can't worry me that much in practice. I'd prefer not to misrepresent your viewpoint, hence the mention.
Newman Posted January 15, 2016 Posted January 15, 2016 Certainly happy to repeat that Keywood is misleading and agenda-driven. I did a quick search of my references to NK, and found a post from me to you in which I refer to the Noel Keywood Nonsense Factory. Just as well you weren't the recipient of my other, older post in which I refer to NK as "a loon with bad judgement and extreme bias". But the more topical detail I found was that yes, it was during a discussion of the anti-digital claim that digital audio has high distortion at low volumes. Specifically, we referred to NK in a discussion on how CD players all measure at near 0.2 % (not 2 %) distortion with a -60 dB signal.
Guest Eggcup The Daft Posted January 15, 2016 Posted January 15, 2016 Certainly happy to repeat that Keywood is misleading and agenda-driven. I did a quick search of my references to NK, and found a post from me to you in which I refer to the Noel Keywood Nonsense Factory. Just as well you weren't the recipient of my other, older post in which I refer to NK as "a loon with bad judgement and extreme bias". But the more topical detail I found was that yes, it was during a discussion of the anti-digital claim that digital audio has high distortion at low volumes. Specifically, we referred to NK in a discussion on how CD players all measure at near 0.2 % (not 2 %) distortion with a -60 dB signal. Thanks
LHC Posted January 15, 2016 Posted January 15, 2016 Interesting. Noel Keywood is one of the more respected reviewer out there, I didn't think any of his articles are over the top in claims.
weirving Posted January 15, 2016 Posted January 15, 2016 So if a song is 10mb MP3, what would that same track be as an ATRAC (ballpark it if you know)File size for a CD track ripped to 256 kb/s MP3 versus same track recorded on minidisc using ATRAC3 at same bitrate would be nearly the same. To further muddy the waters, there is such a thing as ATRAC3 Lossless, which creates much larger file sizes, but being lossless - gives sound quality equivalent to FLAC, which is essentially CD quality, since no information is thrown away, such as with MP3, AAC and other "lossy" compression schemes. Sent from my SCH-R530U using Tapatalk
davewantsmoore Posted January 15, 2016 Posted January 15, 2016 He has made these claims in 2014 and 2015, and about devices using modern chips like the ESS Sabre range. He clearly says "all 16 bit conversion has these levels of distortion". He is unequivocal about it. And whatever we might think about his agenda or conclusions, he has done the measurements and I see no need to doubt his results. Cool.... so any and every measurement of a modern DAC chip (or just ESS, or whatever) will show this effect. Yes?
davewantsmoore Posted January 15, 2016 Posted January 15, 2016 (edited) Interesting. Noel Keywood is one of the more respected reviewer out there I didn't even know who he was ... Edited January 15, 2016 by davewantsmoore
Guest Eggcup The Daft Posted January 15, 2016 Posted January 15, 2016 Cool.... so any and every measurement of a modern DAC chip (or just ESS, or whatever) will show this effect. Yes? Yes. But at 0.2% not 2%. That is a HUGE mistake I made.
davewantsmoore Posted January 15, 2016 Posted January 15, 2016 (edited) . Edited January 15, 2016 by davewantsmoore
davewantsmoore Posted January 15, 2016 Posted January 15, 2016 Yes. But at 0.2% not 2%. That is a HUGE mistake I made. 0.2% THD (ie. at -54dB) .... approximates the noise floor / THD+N measurements of a 16bit system .... you can see in the picture if you imagine a -60dB signal (instead of the -90dB signal) ... that the THD+N is ~ 54dB lower. In in a quiet studio (not a normal room).... if we set the level of that distortion to be just above (or equal to) the noise floor of the room (ie. inaudible) .... the maximum sound in the system (0dBFS) is at 125dB SPL (!!!)
Newman Posted January 16, 2016 Posted January 16, 2016 Yes, that's what we went though in the old thread.
Bilbo Posted January 16, 2016 Posted January 16, 2016 Probably a dumb question but - if you generate a 5khz sine wave and add/modulate it with say a 28khz sine wave the resulting more complex waveform would look like a rippled 5khz shape. If this then passes through a A/D then D/A conversion do you get exactly the same wave shape output?
davewantsmoore Posted January 17, 2016 Posted January 17, 2016 Yes, if the sampling rate is larger than 56kHz (2x 28kHz) If you are sampling at a lower rate. eg. 48kHz .... then you need to filter out all frequencies above half the sampling rate. (eg. everything above 24kHz) .... and so you just have 5kHz.
Bilbo Posted January 17, 2016 Posted January 17, 2016 OK - so the 5kHz waveform looks like a sine wave again. So would that lead to loss of harmonic content above 22kHz? If this is so, then perhaps people can hear subtle differences in Hi Rez because a greater part of the highly complex waveforms' upper harmonics above 22kHz are better preserved with higher sample rates. Perhaps that's what allows people to detect the type or maker of a violin or piano. I don't actually know but it is possible isn't it?
davewantsmoore Posted January 17, 2016 Posted January 17, 2016 That's a logical possibility.... but when investigated, not shown to be the case.
Newman Posted January 17, 2016 Posted January 17, 2016 OK - so the 5kHz waveform looks like a sine wave again. So would that lead to loss of harmonic content above 22kHz? If this is so, then perhaps people can hear subtle differences in Hi Rez because a greater part of the highly complex waveforms' upper harmonics above 22kHz are better preserved with higher sample rates. Perhaps that's what allows people to detect the type or maker of a violin or piano. I don't actually know but it is possible isn't it? You do realise the two cases are identical in content below 20 kHz? Different violins etc sound different because they are different below 20 kHz.
Bilbo Posted January 17, 2016 Posted January 17, 2016 You do realise the two cases are identical in content below 20 kHz? Different violins etc sound different because they are different below 20 kHz. Now I'm confused. I thought following "davewantsmoore" response, the 5kHz wave would lose the 28hHz modulation and return to a sine wave. Have I got that wrong? I am however aware that it is the lower harmonics that provide the distinctive sound of an instrument. The relative 'volume' of the harmonics create the distinction. I guess I'm open to the idea that even these tiny upper harmonics could possibly change the perception of a sound.
LHC Posted January 22, 2016 Posted January 22, 2016 Ok back on topic. The key claim of those associated with the MQA technology is that human can hear time resolution down to 7us (microseconds). This time resolution cannot be achieved by CD sampling rate, so therefore necessitating a sampling rate of 192kHz. This is their claim. They also acknowledged that human can't hear audio frequencies above 21kHz, but they argue our auditory senses in the time domain are far more sensitive. I have been reading this Yamaha website which provide very detail explanation of these claims. This is a very good resource of information. http://www.yamahaproaudio.com/global/en/training_support/selftraining/audio_quality/chapter4/02_audio_universe/ In a nutshell it goes like this. Whilst human can't generally hear audio frequencies above 21kHz, this is only true for continuous audio signals, e.g. like a laboratory test tone. The kicker is that most audio signals are not continuous. The white paper went on to explain that our hearing actually has a theoretical upper limit of hearing 18 MHz, translating to a temporal resolution of 0.055 microseconds! So what is the real world limit? Scientifically controlled blind testing by Dr Kunchur show that we can indeed hear down to 6us or even lower. I would agree more testing of this nature would be useful to validate these claims, and people associated with MQA have claimed to done testing backing up their claims too. So what about real world application in audio reproduction? Well Yamaha claims that in a professional studio, one can indeed approach that limit of 6us, and therefore 192kHz is definitely needed there. In domestic application they claim the limitation of our speakers would mean we could at best hear down to 10us, so a 96kHz sampling rate would be a good compromise. See their table here (http://www.yamahaproaudio.com/global/en/training_support/selftraining/audio_quality/chapter5/09_temporal_resolution/).
Bilbo Posted January 22, 2016 Posted January 22, 2016 What's "off topic" about my question? I was hoping someone in the know would clarify it for me.
davewantsmoore Posted January 22, 2016 Posted January 22, 2016 What's "off topic" about my question? I was hoping someone in the know would clarify it for me. Nothing. I guess I'm open to the idea that even these tiny upper harmonics could possibly change the perception of a sound. So are (were!?) many others. As you can see, it is a very important thing to understand. People investigated it (for many many decades), and have found it not to be the case.
Recommended Posts