384khz?

[Mr. C]

Howabout looking at this from a different perspective, how many of you can tell the difference between an amplifier with a bandwidth of the second harmonic and one that is capable of producing over the 5th harmonic whatever it maybe SS or Valves even if your speakers have a maximum bandwidth of 1.5 harmonics?

And still keeping it simple lats say using a good quality red book recording or vinyl

Sample rate is merely a mathematical solution of the original redbook 44.1Khz, the only time you will really benefit is by playing the record file at its true RECORDED rate, by using up-sampling you are merely multiplying the original file format by a pre determined number, for this an SRC chip-set will be used and these usually contain another a PLL (Phase lock loop) of some description which acts as another re-clocking device in the data reclaim chain. The extra bits are just random noise which are added to make up the numbers. Unless that file was recorded at that NATIVE format configuration.

Simple test compare a native 176.4Khz 24 bit (raw) to the same native file in 44.1Khz/16bit format that has been up-sampled.

While I do agree that the marketeer departments of many firms have been going into overdrive with DSD quad speed, DXD, sample rates up to 768Khz etc, what is really relevant and what really matters?

Again very important fact in how good the actual recording sounds is down to the transfer from the original master no matter what the stated sample rate, was the mixing engineer on form that day, did the record company have an influence on HOW the recording sounded, budget constraints and equipment the mastering was performed on.

Over the years at shows more than a few visitors have been surprised at how good a decent red book file can sound.

With regard to hearing ranges although the maximum for human hearing can be as much as 23Khz in your early 20's (providing no Motor Head concert bass bin visits!) with have a s/n ratio of around 128dB give or take a few.

By the time you reach 40 you maybe lucky to have 16Khz and some have as low as 9Hz however think about the harmonic structure of music, if studio's recorded with 20Khz microphones the music you listen to at home would just not sound the same.

Without question we cannot hear into those 2nd, 3rd, 4th and above harmonics, however we do perceive them if they are taken away we unequivocally know they are missing.

Working out what is relevant to the individual is more important that my dac does 768Khz and 32 bits, maybe so but can the rest of the replay chain ACTUALLY demonstrate what that piece of equipment is genuinely producing.

[Qwin]

This kind of explains where I was coming from.
http://www.tested.com/tech/1905-the-...-24-bit-audio/

In my limited experience, I have heard improvements with higher resolution, but think 24/96 is the point to stop, as beyond this the increased demand on hardware (cost) does not justify the small, if any, audible gains.

There used to be a lot of criticism of the CD sampling format, are we suddenly defending it?

[struth]

I was told you only "sense" up to 35kh, sometimes much less. so no need to go any further than the next multiple

[Sherwood]

As interesting as this discussion is, it has become somewhat circular. Can we agree the following or at least establish whether we have a consensus on the following:

a) It is desirable to achieve the maximum possible "fidelity" in any recording process.
b) The end product (i.e. a recorded piece) can only be as good as its weakest link.
c) It is not possible to reproduce something that was not captured by the original recording process no matter how much processing power you throw at it. You can paint and polish a turd but .........
d) The original cd format was constrained by the technology of the time with regard to data storage so that the 20-20k range was achievable rather than ideal
e) That the extreme sampling and upsampling rates being developed are (perhaps) driven by marketing divisions and that the benefits of higher sampling rates, dynamics, and frequency bandwidth can be achieved without having to go anywhere near these stellar rates.
f) That if a company sells an HD music file it should be a genuine HD recording (HD at all stages) and priced honestly (i.e. reflecting actual production costs rather than as a form of market segmentation and profit maximisation).


I think back to the development of the 12" record and the point made in an earlier post about RIAA equalization. When the long playing disc was being developed there was a mechanical problem that presented designers with a trade off between length of recording and fidelity. The width of a groove varies according to the loudness of the music being recorded and the frequency of the signal. Low frequency and loud signals cut a wide groove in the disc which ultimately limited the number of grooves that could be cut and in turn the length of a recording. An electronic solution was found to this mechanical problem by cutting the discs with artificially reduced bass and then restoring that bass in the playback process. That increased the play time of the 12 inch disc to the 20 odd minutes we have today. As good as this technology can sound it can bettered albeit by sacrificing playback time. I have a Morrisey Mullen 12 inch Jazz single by EMI which has just a single track on each side which gives some idea of how the lp is comprised by storage limitations. Short of making a 24" LP I guess that not much could be done with vinyl to overcome this problem. The same is not true of the cd. Better technology and better engineering have made it possible to get more data onto a disc at a lower cost.

I for one, want to hear more (true) HD content so long as I am not being ripped off.

Geoff

BTW: Here is a not very good recording of that EMI record. Worth keeping an eye out for as it makes a very good test record. https://www.youtube.com/watch?v=vUV2mnCCbXA

[Jimbo]

The mathematical and technical aspects and possibilities of digital audio are interesting to kick around but ultimately no matter what is said it comes down to a simple fact. Can you hear the differences? The industry loves to bandy round and promote both hardware and software with a range of specifications for digital audio playback, however often these translate into meaningless gibberish when you actually listen.

As Mr C mentioned in his post mastering is everything in any format. Buying a Hi Rez recording in 24bit / 196 kHz is not a guarantee of anything other than your pockets will be picked for more cash!

[Macca]

As interesting as this discussion is, it has become somewhat circular. Can we agree the following or at least establish whether we have a consensus on the following:

a) It is desirable to achieve the maximum possible "fidelity" in any recording process.
b) The end product (i.e. a recorded piece) can only be as good as its weakest link.
c) It is not possible to reproduce something that was not captured by the original recording process no matter how much processing power you throw at it. You can paint and polish a turd but .........
d) The original cd format was constrained by the technology of the time with regard to data storage so that the 20-20k range was achievable rather than ideal
e) That the extreme sampling and upsampling rates being developed are (perhaps) driven by marketing divisions and that the benefits of higher sampling rates, dynamics, and frequency bandwidth can be achieved without having to go anywhere near these stellar rates.
f) That if a company sells an HD music file it should be a genuine HD recording (HD at all stages) and priced honestly (i.e. reflecting actual production costs rather than as a form of market segmentation and profit maximisation).

I'd agree all that, although d) is debateable: as discussed previously red book cd is more than adequate unless you have a large room in a very quiet place, devastatingly good (and compatible) system and unblemished hearing.

[Light Dependant Resistor]

The mathematical and technical aspects and possibilities of digital audio are interesting to kick around but ultimately no matter what is said it comes down to a simple fact. Can you hear the differences? The industry loves to bandy round and promote both hardware and software with a range of specifications for digital audio playback, however often these translate into meaningless gibberish when you actually listen.

As Mr C mentioned in his post mastering is everything in any format. Buying a Hi Rez recording in 24bit / 196 kHz is not a guarantee of anything other than your pockets will be picked for more cash!

But to compound the difficulty of hearing differences as I love a simple solution too to be suggested, are sorry to say, source impedances, cable types, attenuation methods the focus of my life for the last 7 years , let alone speakers. IMO power amps are the least problematic component.

As an example today I heard a Mcintosh valve amp into Spendors costing $20.000... it really was rubbish compared to my Quad 306's and JR149's all of $1200.00

So the whole chain has to be somehow be identified as Go Doubtful or No Go specification so a typical audiophile feels happy stepping out the door
or maybe the next room if online to find his or her best choice of component. The McIntosh being a valve design simply needed a high impedance speaker
like high impedance LS3/5a also made by Spencer and Dorothy, but this had not dawned on those trying to sell it.

[Qwin]

Sherwood - from my point of view there is nothing to disagree with in your summing up, its a fair take on the situation.

[Rothchild]

I'd agree all that, although d) is debateable: as discussed previously red book cd is more than adequate unless you have a large room in a very quiet place, devastatingly good (and compatible) system and unblemished hearing.

Well put Geoff, I agree with Martin on point d. 20-20,000 isn't so much representative of a technical limit, rather a practical assessment of what's actually necessary (as per the discussion about where the 'sensible point' between 44.1 and 96kHz sample rates actually lies).

I would also add that e is correct when discussing playback systems, high sample rates have real-world practical benefits to a recording / mixing engineer (eg lower recording latency as the buffers are cleared more often) but are much less relevant in a 2 channel playback system (as per point d)

[Light Dependant Resistor]

As interesting as this discussion is, it has become somewhat circular. Can we agree the following or at least establish whether we have a consensus on the following:

a) It is desirable to achieve the maximum possible "fidelity" in any recording process.
b) The end product (i.e. a recorded piece) can only be as good as its weakest link.
c) It is not possible to reproduce something that was not captured by the original recording process no matter how much processing power you throw at it. You can paint and polish a turd but .........
d) The original cd format was constrained by the technology of the time with regard to data storage so that the 20-20k range was achievable rather than ideal
e) That the extreme sampling and upsampling rates being developed are (perhaps) driven by marketing divisions and that the benefits of higher sampling rates, dynamics, and frequency bandwidth can be achieved without having to go anywhere near these stellar rates.
f) That if a company sells an HD music file it should be a genuine HD recording (HD at all stages) and priced honestly (i.e. reflecting actual production costs rather than as a form of market segmentation and profit maximisation).


I think back to the development of the 12" record and the point made in an earlier post about RIAA equalization. When the long playing disc was being developed there was a mechanical problem that presented designers with a trade off between length of recording and fidelity. The width of a groove varies according to the loudness of the music being recorded and the frequency of the signal. Low frequency and loud signals cut a wide groove in the disc which ultimately limited the number of grooves that could be cut and in turn the length of a recording. An electronic solution was found to this mechanical problem by cutting the discs with artificially reduced bass and then restoring that bass in the playback process. That increased the play time of the 12 inch disc to the 20 odd minutes we have today. As good as this technology can sound it can bettered albeit by sacrificing playback time. I have a Morrisey Mullen 12 inch Jazz single by EMI which has just a single track on each side which gives some idea of how the lp is comprised by storage limitations. Short of making a 24" LP I guess that not much could be done with vinyl to overcome this problem. The same is not true of the cd. Better technology and better engineering have made it possible to get more data onto a disc at a lower cost.

I for one, want to hear more (true) HD content so long as I am not being ripped off.

Geoff
BTW: Here is a not very good recording of that EMI record. Worth keeping an eye out for as it makes a very good test record. https://www.youtube.com/watch?v=vUV2mnCCbXA

Point (d) refers to but has not identified the Nyquist Shannon Sampling Theorem constraints- these dictate the upper audio frequency of the CD medium.
Pioneer Legato Link as example was a brave attempt at stepping beyond the Nyquist Shannon constraint of 20k actually 22,500 . .the last 2500 being
reserved for filtering. Another entire subject suggests about what type and shape of filter is provided in CD players. Legato Link players are very good by the way.

Point (c) emphasis on the word process as stated please, rather than recording. the recording process may contain companding methods, dither shaping and transient capture
where the recording device integrated circuits are placed to read ahead of the information being captured.... just a few refinements that occur during the recording process that
then follow through to really record the event.