ARCHIVED FORUM -- March 2012 to February 2022READ ONLY FORUM
This is the second Archived Forum which was active between 1st March 2012 and 23rd February 2022
To 'highres' or not to 'highres' - that is the question....
http://www.avhub.com.au/editors-blog/hi-fi/when-is-8216hi-res8217-not-8216hi-res8217-397055
MM
There is a tv - and there is a BV
The only positive thing is that it's professional ript.
Sorry for my ignorance - what does 'professional ript' mean?
It's a term they use in the 'pirate download scene'.😂😇
I'm afraid that Mr. Borrowman that article is (1) slightly-to-considerably confused on a number of points, (2) makes at least one assumption that is incorrect, and (3) makes some incorrect statements due to lack of care in wording. The result is a rather low signal-to-noise ratio...
Cheers
-g
Take your time, Geoff ;-)
We (at least I) want to know more - what are these (1), (2), (3) points?
According to Wikipedia, Johnny Ray died of liver failure in 1990 - so this much is true. After that point in the editorial, we're skating on thin ice.
It is easy to state that most hi-res tracks are not actually hi-res, but it's very difficult (or at least extremely time-consuming) to prove. It has been proven that at least one track available for download as a 24-bit file was actually a sample-for-sample identical match to the 16-bit version that was previously available.
In addition, this caused some discussion regarding the moral validity of charging more for an identical product in a bigger package...
The author says "...in order to be classified as a ‘high-res’ file the music performed by the musicians that is contained in that file MUST have been originally recorded digitally with a machine generating a 24-bit word every 48-thousandth of a second."
This is simply incorrect. The Digital Entertainment Group (DEG), the Consumer Electronics Association (CEA), and The Recording Academy have collaboratively agreed on a definition for high-resolution audio as follows:
Lossless audio that is capable of reproducing the full range of sound from recordings that have been mastered from better than CD quality music sources."
The four descriptors are:
MQ-P: from a PCM master source 48kHz/20-bit or higher – typically 96/24 or 192/24 content
MQ-A: from an analogue master source
MQ-C: from a CD master source (44.1kHz/16-bit)
and MQ-D: from a DSD/DSF master source – typically 2.8MHz or 5.6MHz content
So, as you can see, 48 kHz/24 bit PCM is in the list - but so are other formats and sampling rates.
The author also claims that analogue recordings do not qualify as being high-resolution, which is incorrect as is evident in the MQ-A descriptor. This would therefore (potentially) include multi-track recordings made on analogue magnetic tape and analogue mixing consoles.
As a small point, the author repeatedly talks about "DACs" - but I believe that he means "ADC's" instead. The digital-to-analogue converter in a mixing console or recording system has nothing to do with the resolution of the recording it is capable of making.
However, it's also not enough to say that "if you don't have a 24-bit DAC, you don't have a 24-bit recording". For example: back when I was just starting to learn techniques for digital audio editing and mastering (which would have been around 1991 or so), we had 16-bit ADC's, but a 24-bit editing system. So, we would take the signal from a pair of microphones to the mic preamp's, and the output of that device to two 16-bit DAT recorders, running at different input gains (say 12 dB apart). For most of the recording, we would use the recording from the "hotter" recording, dropped by 12 dB in the editing system. However, on a peak, this recording would overload, so we would cut to the signal from the quieter recording (without attenuation) for the peak. This meant that we were effectively achieving an 18 bit recording (the 2 extra bits correspond to the 12 dB difference). Of course, we didn't really have an 18-bit recording, since, at any given moment, there were only 16 valid bits being used to represent the signal. However, the total range (from the noise floor of the "normal" recording to the peak of the peaks) was 18 bits.
Another, much simpler example is a case where a 16-bit recording is modified - for example, with equalisation. Assuming that your equalisation block in the signal processing has more than 16 bits, then a 16-bit input will result in a greater-than-16-bit output. This is not artificially increasing the bit depth - the additional bit depth provides a better representation of the filtered 16-bit signal.
The author writes "upsampling never improves the fidelity of the original recording." This is incorrect. It is based on the assumption that nothing but upsampling is done. However, if any signal processing is done at a the new sampling rate, the effects on the audio fidelity (depending on your definition of "fidelity") can either be an improvement, a detriment, or it may have no effect at all. This, however, is dependent on exactly what processing was done, and how it was implemented.
This statement also assumes that the DAC that will eventually be used to convert the signal to analogue has an identical behaviour at a different sampling rate. This is very unlikely.
The analogy of the steak on the bigger plate is also incorrect. This is a good analogy for the 16-bit Dylan recording distributed as a 24-bit version where the individual samples were simply zero-padded. However, in cases where the signal was re-sampled, there will likely be some effect on the output "fidelity" - either better or worse. Whether or not that change is audible is a secondary question that is dependent on many factors.
Finally, the author states "You’re also likely listening to music that’s so compressed it uses only 8-bits of what’s available, but that’s yet another story." This is just wrong. The implication here is that a bitrate is directly correlated with a perceived audio quality - which is a connection that many people like to make as a last-resort accusation against lossy CODEC's. Although a lossy compression does have an effect on audio quality, this does not mean you're only geting "8-bits of what's available".
But, Johnny Ray is almost certainly dead...
-geoff
Geoff Martin: Finally, the author states "You’re also likely listening to music that’s so compressed it uses only 8-bits of what’s available, but that’s yet another story." This is just wrong. The implication here is that a bitrate is directly correlated with a perceived audio quality - which is a connection that many people like to make as a last-resort accusation against lossy CODEC's. Although a lossy compression does have an effect on audio quality, this does not mean you're only geting "8-bits of what's available".
I take this one back (but I can't EDIT in Safari on my Mac, so I'll leave the error and correct it here).
Re-reading this statement, I guess that the author is talking about dynamic range compression rather than lossy CODEC's. However, it's still an error, since even through a tune might be compressed to have far less than 48 dB of dynamic range (8bits * 6 dB per bit), the bit depth, in part, determines the noise floor which is (assuming the dithering was done properly) independent of the signal. So, you might have "only" 48 dB of dynamic range in your music, but you still want the noise floor to be a far below that as possible. How far is necessary is dependent on many things (such as the bandwidth and spectrum of the signal vs. the bandwidth and spectrum of the noise). The narrower the bandwidth of the signal, the more audible a wide-band noise floor. (A good example of this is to listen to the beginning of "A Thousand Years" by Sting. It starts with a very low-frequency rumble that does nothing to mask the noise floor in the high frequency bands - especially if you're listening on loudspeakers that have little-to-no output at low frequencies...)
P.S. why can't I edit postings in Safari on my Mac? It feels a little like doing a crossword puzzle in pen... "Computer says 'no....' "
Geoff Martin: P.S. why can't I edit postings in Safari on my Mac? It feels a little like doing a crossword puzzle in pen... "Computer says 'no....' "
This was discussed a couple of times here.
Noone really knows - or noneone has explained why.
It started a while ago with some updates of Safari (and the OS).
Probably the way this site is 'written'/compiled does not harmonize with how the browser works.
It is very annoying, but - on the positive side - it forces you (means us) to think before we write ;-))
A workaround should be to use another browser (which I haven't tried yet), or - if you are on an iPad/iPhone - the Beoworld app, that Philippe Robin made a while ago. Another way is to write in another app/another program, and copy/paste that afterwards to Beoworld.
So - back to thinking, before I write a comment on your posts ;-)))
@Geoff
The beomaster 5 scales hi-res files 24/96 down to 24/44,1
Can you hear any difference when you play these files on a BSMoment