NeoG Posted August 19, 2013 Posted August 19, 2013 It seems to be the case that all I keep on coming up with is that the only difference between a properly designed, dithered 16-bit signal and the same in 24-bit is the SNR. The fun part of that is, that it asserts that everyone that can hear the different between 16-bit and 24-bit source audio in a playback system using a badly designed 16-bit DAC. Besides recording and multiple passes of AD/DA, I can't see where the difference from a 96 to 144dB noise floor would be audible in a standard listening environment.. Even if your room was extremely quiet (25dB SPL background), you'd still need to hit peaks of 121dB to utilise the entire range of 16-bit. So, who denies this assertion?
CraigC Posted August 19, 2013 Posted August 19, 2013 Digital isn't my strength, I don't understand how the increasing the number of bits i.e resolution increases the dynamic range unless the full amplitude of the waveform wasn't sampled in the lower resolution format. Or is he arguing that the higher resolution sample contains less artificial information than a lower resolution sample, the reconstruction of the waveform between the larger space of sample points increases the noisefloor?
davewantsmoore Posted August 19, 2013 Posted August 19, 2013 the only difference between a properly designed, dithered 16-bit signal and the same in 24-bit is the SNR Dithering is an unnecessary complication. The only difference between a 24bit and a 16bit system is the SNR.... which is a complicated way of saying that the 24bit system can store quieter sounds. 1
davewantsmoore Posted August 19, 2013 Posted August 19, 2013 So, who denies this assertion? Its just like you say. *tumbleweed* There is the possibility that certain converters might work better with more bits.... this though, is a "fault" of the converter..... and nothing to do with 24bit being "better" (unless you actually wish to utilise the extra dynamic range, of course).
NeoG Posted August 19, 2013 Posted August 19, 2013 I don't understand how the increasing the number of bits i.e resolution increases the dynamic range unless the full amplitude of the waveform wasn't sampled in the lower resolution format. Seems to be a common misconception that 24-bit increases the data points traversed within the usable dynamic range. All it really does is redefine the minimum level relative to 0dBFS. Audio data louder than -96dBFS will have the same resolution in both 16-bit and 24-bit. The extra 8 bits are only utilised to store amplitudes between -144dBFS and -96dBFS. Is there any relevant data in there for playback? Common wisdom dictates that there isn't. Perveyors of Vinyl are happy with audio data that pretty much ends at -70dB normalised to peaks. It's like adding 8 steps to a 16-step ladder. Instead of putting 24 steps in the same space, they just made the ladder 50% longer. 1
davewantsmoore Posted August 19, 2013 Posted August 19, 2013 Digital isn't my strength, I don't understand how the increasing the number of bits i.e resolution It is very misleading to think of increasing the bits, as increasing the "resolution". This is where much of the misconceptions around stem from it seems. If you take some audio ... which consists of sounds which all fall between the range of 45 to 85dB (arbitrary) ..... If you capture the sound, once with a 16bit and then with a 24bit system .... They have the identical "resolution" .... 24bit does not capture the sound "better".... it simply offer the ability to store (additional) quieter sounds than the 16bit system. In our example (45db to 85db) the potential benefit of storing quieter sounds, is not relevant. 1
davewantsmoore Posted August 19, 2013 Posted August 19, 2013 Seems to be a common misconception that 24-bit increases the data points traversed within the usable dynamic range. Yes, it is.
Volunteer sir sanders zingmore Posted August 19, 2013 Volunteer Posted August 19, 2013 So how do you explain the night and day differences that everyone hears?
Guest fordgtlover Posted August 19, 2013 Posted August 19, 2013 So how do you explain the night and day differences that everyone hears? How do you explain that people go to war to fight over who has the best imaginary friend? 4
NeoG Posted August 19, 2013 Posted August 19, 2013 So how do you explain the night and day differences that everyone hears? A CONTENDER Suggestive shared hallucination and badly designed 16-bit DACs
hochopeper Posted August 19, 2013 Posted August 19, 2013 So how do you explain the night and day differences that everyone hears? Jitter, harmonic distortion, imd, etc Transport, clocks, iv stage, power supply and decoupling, PCB layout, oversampling ... Take your pick and consider that DAC data interface and output modulator design differences may be preferred also.
davewantsmoore Posted August 20, 2013 Posted August 20, 2013 So how do you explain the night and day differences that everyone hears? Difficult to generalise ... without more information. Do you mean the playback of the same material, stored in 16bit and 24bit containers... all else being equal ?! .... or a less tightly controlled test ?!
fetischizm Posted August 21, 2013 Posted August 21, 2013 (edited) yep, this is exactly the same as the confusion about higher sample-rate allowing more detail to be rendered in the audible frequencies, instead of just allowing the creation of higher frequencies. that being said I will go for 24 or 32 bit every time (or even higher if possible) as I use DSP and digital volume. dither is cool and all, but if its easily avoided with current technology, i'd rather do that. Edited August 21, 2013 by fetischizm
NeoG Posted August 21, 2013 Posted August 21, 2013 (edited) yep, this is exactly the same as the confusion about higher sample-rate allowing more detail to be rendered in the audible frequencies, instead of just allowing the creation of higher frequencies. There is something to be said for rates higher than 44.1k.. 96k allows you to leave the entire human audio band in tact and use a less intrusive filter at the fringe frequencies. I don't really hear it myself but I concede this is logically sound for avoiding poor quality filters. Edited August 21, 2013 by NeoG
davewantsmoore Posted August 21, 2013 Posted August 21, 2013 (edited) There is something to be said for rates higher than 44.1k.. 96k allows you to leave the entire human audio band in tact and use a less intrusive filter at the fringe frequencies. I don't really hear it myself but I concede this is logically sound for avoiding poor quality filters. Oversampling takes care of that ..... EDIT: NB - Oversampling doesn't have to be done in your device. You can still use a "NOS" DAC ... and pre-oversample (not in realtime) the data, before sending it to the DAC. Like Phasesure NOS1 / xxHighend ..... or like I do with Audirvana (oversample 4x before playback begins using izotope) Edited August 21, 2013 by davewantsmoore 1
metal beat Posted August 21, 2013 Posted August 21, 2013 It is very misleading to think of increasing the bits, as increasing the "resolution". This is where much of the misconceptions around stem from it seems. If you take some audio ... which consists of sounds which all fall between the range of 45 to 85dB (arbitrary) ..... If you capture the sound, once with a 16bit and then with a 24bit system .... They have the identical "resolution" .... 24bit does not capture the sound "better".... it simply offer the ability to store (additional) quieter sounds than the 16bit system. In our example (45db to 85db) the potential benefit of storing quieter sounds, is not relevant. In layman's terms, does 24 bit equate to more " headroom " than 16 bit ?
davewantsmoore Posted September 2, 2013 Posted September 2, 2013 In layman's terms, does 24 bit equate to more " headroom " than 16 bit ? Yes... (assuming your definition of headroom) 16bit... allows you to represent from 0 down to -96dB 24bit .... allows you to represent between 0 and -144dB In a 24bit system.... the values from a 16bit system are not stored more accurately.... they are exactly the same. The 24bit system allows you to store quieter sounds than the 16bit system (specifically between -96dB and -144dB) This much "headroom" is not able to be practically used in a playback system (ignoring crazy circumstances) ... but can be useful in recording, editing/mixing audio.
fetischizm Posted September 2, 2013 Posted September 2, 2013 regarding 'crazy circumstances' thats not really correct, digital volume control is being used more and more for consumer audio and its particularly important for those using digital crossover. of course dithering out to 24bit or >32bit has essentially the same effect, but if the 24bit file is available, and your system is 24bit ready, may as well use it natively IMO
davewantsmoore Posted September 2, 2013 Posted September 2, 2013 regarding 'crazy circumstances' thats not really correct, digital volume control is being used more and more for consumer audio Sure... but the data can be simply padded out to higher bit depths. (I was trying to keep it simple) 1
Nada Posted September 3, 2013 Posted September 3, 2013 (edited) In my little experience the main advanatge of 24bit capable DACs is the numbers sound better to punters, so being gullible Ive generally purchased 24bit DACs. The TDA1541 is a great sounding chip but at only 16bits even Audial has needed to move to modern bigger bit chips....how about 32 bits? Thats even more impressive then 24bit I reckon. Other DAC manufactures like the Metrum have made 24 bit claims for DA chips that might accept 24bits but cant process more then 16bits. Personally Im waiting for a 64bit DAC to come out before unloading my saving........ Edited September 3, 2013 by Nada
Nada Posted September 3, 2013 Posted September 3, 2013 (edited) .....maybe I should wait for a 128bit DAC? Edited September 3, 2013 by Nada
davewantsmoore Posted September 3, 2013 Posted September 3, 2013 .....maybe I should wait for a 128bit DAC? Be interested to know what amplifier you plan to you use with a 32 bit DAC ... which had either the SNR or power to resolve that range. Oh. Sarcarm?! Sorry. :lol:
Recommended Posts