a lot of people in this thread seem to be confused about frequency and sample rate. From my understanding (I did a lot of wiki reading when first investigating hi def audio), 24/192 means 24bit frequency sampling (meaning 8 more bits of only dogs can hear it frequencies)
24 bit refers to the length of the digital word. It does not refer to sample rate or frequency.
Increasing sample rate during analog to digital conversion does not
makes a warmer more analog sound--sound waves are smoother, less pixelated.)
It increases the maximum frequency that can be represented in the digital audio.
By using a higher sample rate, required filtering takes place at higher frequencies. All filters introduce non-linearities. By having this filtering at a much higher frequency than the human ear can register, the effects of these non-linearities is minimized.
In The Master Handbook of Acoustics
, (4th edition), F.Alton Everest writes," In spite of vigorous research activities on all aspects of human hearing, our knowledge is still woefully incomplete."
This is especially true when it comes to our understanding of the mechanisms of human hearing above the highest frequency we can 'hear'; 20,000 hertz.
In The Complete Guide to High-End Audio
, Robert Harley writes,"One mechanism, however, by which today's digital audio degrades musical quality is too low a sampling frequency. Although the upper limit of hearing is 20,000 Hz, ... research has shown that increasing the bandwidth from 20,000 Hz to 40,000 Hz improves musical reproduction. Although we can't hear 40,000 Hz sine waves, removing energy above 20,000 Hz from the signal reduces the music's sense of openness, transient attack, and natural quality."
So,below are some digital audio basics. Hopefully this will clear up some of these misconceptions.
I suggest that anyone interested in home theater and music reproduction read both of these references.
Everest's masterwork contains a wealth of information on sound, how it is perceived and how to improve (or design from the ground up) a room for home theater or music reproduction.
Harley's book has an appendix with a description of digital audio basics.
Sound energy in air is actually air pressure compression and rarefaction. Sounds change as these pressures change. In analog audio sound pressure is transduced (converted from one form of energy to a different form) by microphones or instrument pickups into AC current.
Analog audio to digital audio conversion takes this alternating current electrical signal and converts it to binary data. This is done by measuring (sampling) the voltage (which is a model of the audio's amplitude) of the electrical signal at a very high rate.
This is the basic way in which linear PCM (Pulse Code Modulation) digital audio is converted from an analog electrical signal.
Linear PCM is the data format output by CD, DVD-Audio,
Dolby True HD (after conversion from bitstream), DD Plus (after conversion from bitstream), and DTS MA (after conversion from bitstream).
It doesn't matter where the conversion from the native file format of the medium takes place. Whether Dolby True HD or DTS MA Lossless is converted in the player or by a receiver or separate preamplifier/processor, it is converted to PCM.
PCM is specified by:
word length (unit = digital bits) at sample rate (unit = cycles per second or Hz).
This is the data format of PCM audio.
The word length specifies the maximum number of possible amplitude variations (think volume levels), sampled from the audio's analog electrical signal.
The sample rate specifies the number of these variations which are measured in one second.
The standard for CD is 16 bit word length / 44,100 sample rate. 16 bit word lengths means that CD has a maximum of to 2 to the 16th power ( = 65536) possible amplitude variations.
Sample rate effects the highest possible frequency that can be modeled in analog to digital conversion. The Nyquist Theorem essentially limits the highest possible frequency in the converted digital word to one half of the sample rate. Filtering requirements in practice further limits the highest frequency that can be represented.
So for CD, the theoretically highest frequency audio that can be represented by the digital word, is 1/2 of 44,100 = 22,050 Hertz.
In practice, filtering requirements lowers this maximum frequency to a little more than 20,000 hertz. Since the filtering frequency range is close to the limit of human hearing's frequency response, filtering can cause detrimental effects on the sound output.
The first CD players suffered badly from these effects. Designs soon incorporated 'oversampling' to move the filtering to much higher frequency ranges to limit these effects.
DVD-Audio with 5.1 channels standard is 24 bit / 96 Khz sample rate. So for 24 bit PCM there are 2 to the 24th power (= 16,777,216 possible levels) at a (theoretical) highest frequency of 48,000 hertz.
In stereo (2 channel audio), the DVD-Audio standard provides for 24 bit / 192,000 sample rate audio.
The new (lossless compression) audio codecs (encode / decode) used in Blu Ray discs (Dolby Ture HD and DTS Master Audio Lossless), also output 24/96 (for up to 7.1 audio) or 24 /192 for stereo. In fact, these limits are due to limitations of the disc, not the codec.
DD+ uses a lossy compression scheme. I haven't seen a specification of its PCM output. Dolby's web site does say that DD+ has maximum data bandwidth rate of up to 6.144 Mbit/s (megabits per second).
In comparison, Dolby True HD has a maximum data bandwidth rate of 18 Mbit/s.
I expect Dolby True HD to become ubiquitous. It will eventually be used in broadcast television and radio, and also packaged music media, replacing the CD, SACD and DVD-Audio formats.
This change may come much sooner than expected. Recently DirecTV and Dish announced they will soon provide some 1080p video. Both are touting 'Blu Ray' quality. So True HD is bound to follow.
And music only Blu Ray disc albums are starting to be released. Neil Young and Warner Music Group will be releasing his entire catalog on Blu Ray with True HD soundtracks. The first batch is due in the next month or two.
That's why I am watching for an audio card that decodes and/or bitstreams these new codecs.
Finally, there are also misconceptions about audible differences between higher word lengths and sample rates.
I own a small recording studio. I also have a high end audio system. I can hear huge differences between 16 bit and 24 bit audio recordings I have made. I can also hear subtle, but quite audible differences between 48,000 sample rates and 96,000 sample rates. I can even hear very subtle, but definitely audible differences between 96,000 and 192,000 sample rates.
It is my experience that increasing word length has more impact on sound quality than increasing sample rate. However, there seems to be a limit to the benefits of increasing both of these parameters past 24 bit/ 192 khz. The law of diminishing returns starts to rear its head.
In the interests of full disclosure, I am a 54 year old. Two years ago, tests revealed that I had a measured upper frequency limit around 18,000 hz. That makes my ears quite unusual. After the average person reaches the age of 30, the upper limit is usually around 16,000 hz.
Having said that, my 84 year old father, who has measured hearing loss especially in the mid-band where it makes it somewhat difficult for him to discern speech, can sometimes hear small differences between 16/44.1 and 24/96 audio. So don't dismiss the benefits of high bit rate and sample rate audio as only the purview of dogs.
While my behavior has sometimes been described in terms that could be seen as for the dogs, I assure you I am actually 100% human.
I hope this information will be helpful.