March Audio Posted January 23, 2023 Posted January 23, 2023 So a few examples of whats happening in Hi res audio files. Very low level musical content above 22kHz, just noise above 30kHz. Same again Noise above 24kHz A 2L recording. Wow there really is something above 22kHz up to about 50kHz! Its such low level though, below -80dB. Then lots of noise shaped noise! Again above 22kHz its very low level in the region of -80dB. Followed by massive DSD noise shaping. So that 178kHz bandwidth was really useful ! Again -80db above 22kHz and just noise above 30kHz 2
davewantsmoore Posted January 23, 2023 Posted January 23, 2023 9 hours ago, andyr said: ... do not seem to be able to explain three simple things about the sound of amplifiers: The non-flat frequency response (at the speakers) explains the differences. That, and/or listeners not sitting in quite the same place each time, and/or "listening for things". 8 hours ago, Addicted to music said: I don’t care what anybody saids, it’s not small -100dB is small. 8 hours ago, Addicted to music said: It has the potential to effect other things. Go on?!..... 1
davewantsmoore Posted January 23, 2023 Posted January 23, 2023 19 hours ago, andyr said: There certainly is a place for them ... but where I perhaps diverge from (what I've observed of) your philosophy is that I don't think the measurements we currently are able to do on, say, an amp - are adequate to tell us whether it will sound better than another amp. So why did you jump to the conclusion that you should get rid of the spike?
March Audio Posted January 24, 2023 Posted January 24, 2023 (edited) Just for giggles I have performed an experiment to give people an idea of what is going on above 22kHz in a hi res recording. This is a 192kHz 2L track. You can see some very low level stuff going on up to 50kHz. Its just noise shaping above this. I then filtered out all signals below 22kHz You cant directly hear this content as its above 22kHz and your range of hearing. So I slowed the replay down by a factor of 4. This shifts all that content above 22kHz down to between 5.5kHz and 24kHz. Into the audible range I then amplified it by 40dB (x100) If you want to hear this content, here are short clips: https://1drv.ms/u/s!AnQ0c7fb_4zLg173fY7IE8p9XWYs?e=sUHJiq High pass filtered above 22kHz https://1drv.ms/u/s!AnQ0c7fb_4zLg2THgbQZ3j27VLQs?e=iADg2w High pass filtered above 22kHz slowed down to 0.25 https://1drv.ms/u/s!AnQ0c7fb_4zLg2M4k6fjbD1177Cv?e=sqnuf3 High Pass filtered above 22kHz, slowed down to 0.25, amplified by 40dB (x100) https://1drv.ms/u/s!AnQ0c7fb_4zLg2B7c-_jcXBwQk_l?e=9dSn4P Edited January 24, 2023 by March Audio 2
aussievintage Posted January 24, 2023 Posted January 24, 2023 51 minutes ago, March Audio said: If you want to hear this content, here are short clips: High pass filtered above 22kHz https://1drv.ms/u/s!AnQ0c7fb_4zLg2THgbQZ3j27VLQs?e=iADg2w High pass filtered above 22kHz slowed down to 0.25 https://1drv.ms/u/s!AnQ0c7fb_4zLg2M4k6fjbD1177Cv?e=sqnuf3 High Pass filtered above 22kHz, slowed down to 0.25, amplified by 40dB (x100) https://1drv.ms/u/s!AnQ0c7fb_4zLg2B7c-_jcXBwQk_l?e=9dSn4P Yep, nothing much up there. 1
March Audio Posted January 24, 2023 Posted January 24, 2023 (edited) 3 minutes ago, aussievintage said: Yep, nothing much up there. Its so quiet it wouldnt be audible even if it were below 22kHz. We have to massively amplify it to hear anything. Then you need to consider what you speakers would also attenuate the signal further. Edited January 24, 2023 by March Audio 1
dcathro Posted January 24, 2023 Posted January 24, 2023 On 23/01/2023 at 11:21 AM, BugPowderDust said: There's no secrets that many of the original CDs (pre 1986) sounded thin and lifeless, because they pretty much just lift and shifted the vinyl master to digital. Again this is the opposite of my experience, which is that the first CDs sound more rich, dynamic and lifelike. On 23/01/2023 at 11:21 AM, BugPowderDust said: Some CDs I have from that period are basically unlistenable even on my $60k DAC. Your ears would have to be painted on to have not found this while delving through early CD releases. How many of these unlistenable CDs do you have? Maybe you need a better DAC . I have about 3000 early CDs, mainly from 1982 to 1984 and although the quality varies, I have rarely heard a remaster that is better. 2
March Audio Posted January 24, 2023 Posted January 24, 2023 (edited) 14 hours ago, dcathro said: I have rarely heard a remaster that is better. I tend to agree in general terms, but that's got little to do with the potential audio quality of the then V now. Most remasters are compressed, equalised and generally inappropriately messed with. Edited January 25, 2023 by March Audio 1
dcathro Posted January 25, 2023 Posted January 25, 2023 18 hours ago, March Audio said: I tend to agree in general terms, but that's got little to do with the potential audio quality of the then V now. Most remasters are compressed, equalised and generally inappropriately messed with. There are a lot of variables at play. I have multiple variations - different labels or country of release- of many of my favourite albums and they can sound quite different. Who knows what the sources were, what EQ or compression choices were made, what A to D converters were used, etc. One advantage that the early releases had was the age/state of the source tapes. 3
wayneh Posted April 23, 2023 Posted April 23, 2023 All I know is since I have started downloading my music (hi res flac), I haven't listened to a single cd since. To me 16bit sounded as though the speakers had a blanket over them. The problem is I would put a blu ray movie on and then put a cd on. 99% of the time I would turn the cd off. Now when I put some hi res music on after a movie it usually gets turned off when I start to fall asleep. So for me hi res (music or movies) is better than 16bit.
andyr Posted April 23, 2023 Posted April 23, 2023 17 minutes ago, wayneh said: All I know is since I have started downloading my music (hi res flac), I haven't listened to a single cd since. To me 16bit sounded as though the speakers had a blanket over them. The problem is I would put a blu ray movie on and then put a cd on. 99% of the time I would turn the cd off. Now when I put some hi res music on after a movie it usually gets turned off when I start to fall asleep. So for me hi res (music or movies) is better than 16bit. So, to follow on from this ... can one make the music recorded on a CD sound better - by ripping it to disk and converting it to 24bit/96kHz?
muon* Posted April 23, 2023 Posted April 23, 2023 (edited) Whether 16/44.1 or 24/192 IMO it is how it is reproduced and the master it comes from matters much more, I find very little between them if both are done well, and I favor redbook 16/44.1 personally. Correction, with my DAC my 16/44.1 CDs are reproduced with 14 bits Each to their own though, the above is just what I like and my personal view. Edited April 23, 2023 by muon* 2
davewantsmoore Posted May 27, 2023 Posted May 27, 2023 On 23/04/2023 at 3:03 PM, wayneh said: All I know is since I have started downloading my music (hi res flac), I haven't listened to a single cd since. To me 16bit sounded as though the speakers had a blanket over them. Take one of your 24bit files you have downloaded, and convert it to 16bit. Then audition the two files blinded. On 23/04/2023 at 3:24 PM, andyr said: So, to follow on from this ... can one make the music recorded on a CD sound better - by ripping it to disk and converting it to 24bit/96kHz? No, because doing so does not add any quiter sounds (what more bits helps with storing), nor does it add any higher frequencies (what more smapling rate allows you to store). If you did this ... and you heard any difference (what so ever). Then. its possible your playbck equipment works with higher performance (due to its internal quirks) when playing higher sampling rates, or multiples of 48khz rates, or similar.... this is actually realively rare (especially with audible differences, as opposed to just measurable differences) , but can happen. 2
andyr Posted May 27, 2023 Posted May 27, 2023 1 minute ago, davewantsmoore said: No, because doing so does not add any quieter sounds (what more bits helps with storing), nor does it add any higher frequencies (what more sampling rate allows you to store). Thanks Dave, for the clear explanation.
davewantsmoore Posted May 27, 2023 Posted May 27, 2023 1 hour ago, davewantsmoore said: Take one of your 24bit files you have downloaded, and convert it to 16bit. To be clear, what this will do is show you specifically the difference between 16 and 24 bit (ie. you will be sure that there is noting else different about the content than 16 vs 24 bit) .... it will be inaudible (unless there was a quirk or a problem). That doesn't mean all difference between "the CD" and "the download @ 24bit" is always imagined .... just that something else is reponsible for the difference than the "24bit-ness". 2
Cloth Ears Posted May 28, 2023 Posted May 28, 2023 On 27/05/2023 at 6:01 PM, davewantsmoore said: To be clear, what this will do is show you specifically the difference between 16 and 24 bit (ie. you will be sure that there is noting else different about the content than 16 vs 24 bit) .... it will be inaudible (unless there was a quirk or a problem). That doesn't mean all difference between "the CD" and "the download @ 24bit" is always imagined .... just that something else is reponsible for the difference than the "24bit-ness". Making sure that the "convert it to 16bit" is done with reasonable software. Not all conversions are equal - or is it, some are more equal than others?
davewantsmoore Posted May 29, 2023 Posted May 29, 2023 1 hour ago, Cloth Ears said: Not all conversions are equal No they are not .... but, you also will be super duper really truly unlikely to tell the difference, as they are >>80dB below the loudest sound in the recording. 1
MLXXX Posted May 29, 2023 Posted May 29, 2023 4 hours ago, Cloth Ears said: Making sure that the "convert it to 16bit" is done with reasonable software. Not all conversions are equal - or is it, some are more equal than others? I think most would default to noise-shaped dither. And at normal playback levels such dither should be inaudible. 1
Grizaudio Posted May 29, 2023 Posted May 29, 2023 Not sure if this has been posted already. Well worth a read. https://archimago.blogspot.com/2023/05/results-internet-blind-test-of-24-bit.html?m=1 1
LHC Posted May 29, 2023 Posted May 29, 2023 46 minutes ago, Grizaudio said: Not sure if this has been posted already. Well worth a read. https://archimago.blogspot.com/2023/05/results-internet-blind-test-of-24-bit.html?m=1 Great find. I think this result is consistent with the 2016 Joshua Reiss meta analysis result. 1
MLXXX Posted May 29, 2023 Posted May 29, 2023 (edited) 10 hours ago, LHC said: Great find. I think this result is consistent with the 2016 Joshua Reiss meta analysis result. As I recall, that result was marketed by certain proponents of hi-res as vindication of a need for hi-res, though it was viewed by many others as the direct opposite of that. (The correlations found upon amalgamating the results from a large number of disparate prior studies were quite minor in size, despite technically reaching a statistical threshold. ) Anyway, the Archimago's Musings report blog released earlier this month includes paras I'm quoting below. Similar to the 2016 meta analylsis outcome, there is an absence of evidence supporting a real need for hi-res as a distribution format:- In this blind test, the music used is quite different (modern electronica / pop), and the procedure has changed (much simpler with 1 track, and an option for "no difference" heard) compared to our classical music 24-bit vs. 16-bit test done in 2014 using 2L demo tracks. Nonetheless, results remain similar in that in the big picture, there was basically no audible preference for the higher resolution 24-bit version. While intellectually we can still desire the "best" 24-bit resolution file we can get for our favourite albums, this test adds to the body of evidence that suggests the absence of audible benefits with "high-res" audio. Other tests like Mark Waldrep's comparing not just 24-bits but also increased samplerates (2020) have not been able to ascertain meaningful audible superiority for hi-res (also see the meta-analysis from Reis in 2016, not impressive results as discussed previously). Edited May 30, 2023 by MLXXX 1
davewantsmoore Posted May 29, 2023 Posted May 29, 2023 9 hours ago, LHC said: Great find. I think this result is consistent with the 2016 Joshua Reiss meta analysis result. IIRC the meta analysis basically said "very few studies looked at quantisation" (so we don't know either way) .... where as the Archmagio result (to my view anyways) says it is clearly not audible. Most people answered it was not audible, and of the small number who thought it was clearly audible, the majority got it "wrong" (the preffered the 16bit version). 30 minutes ago, MLXXX said: As I recall, that result was marketed by certain proponents of hi-res as vindication of a need for of hi-res Hehe really!? LOL. 1
muon* Posted May 30, 2023 Posted May 30, 2023 39 minutes ago, davewantsmoore said: IIRC the meta analysis basically said "very few studies looked at quantisation" (so we don't know either way) .... where as the Archmagio result (to my view anyways) says it is clearly not audible. Most people answered it was not audible, and of the small number who thought it was clearly audible, the majority got it "wrong" (the preffered the 16bit version). Hehe really!? LOL. Did they get it wrong? Or did they just prefer 16bit? I read it as just what they preferred rather than trying to pick the 24bit tracks.
MLXXX Posted May 30, 2023 Posted May 30, 2023 50 minutes ago, davewantsmoore said: IIRC the meta analysis basically said "very few studies looked at quantisation" (so we don't know either way) You are referring to the fact that most studies looked at the outcome from reducing both the sample rate and the bit depth, and not just reducing the bit-depth in isolation?
Recommended Posts