|
Post by jazznoise on Aug 17, 2013 4:53:34 GMT -6
I'm sure with some extensive digging, one of us could find the published studies on this... if we had someone amongst us who demanded peer-reviewed scientific proof. There are those who scream "BS" without it. You all know the types. I'm one of the types. There's never been a repeatable experiment to verify this and one of the papers exclusively used bone conduction, which really only plays into our hearing in a meaningful way if we've a bad ossicles or the music is horrendously loud. To the point where'd youd back back to the former. Either way, the test was never recreated. The results correlate, but since the high frequencies can't be isolated it's dubious. It could just be a form of distortion within the test system. Being deaf at 20Khz but able to hear 30khz seems a little counter intuitive anyway. Ain't nothin on the basilar membrane than has a low enough mass to react to 30khz and no other mechanics to suggest another modus operandi. So, yeah, it's balloney to me until we see emperical evidence that supports this theory. Anyway, almost no audio equipment is callibrate to hear that high. A needle could move at 30Khz, but the cutting lathe filter took all that stuff out as otherwise the M.E wouldn't be able to cut as hot. And your desk, mics, compressors and EQ's were all happily cutting chunks out of the ultra sonics.
|
|
|
Post by Ward on Aug 18, 2013 17:27:38 GMT -6
To your point...Part of the loudness war is to drastically cut everything above 12khz - which is also the top end of a 128 bit (or less) MP3. Less true high end means the stronger frequencies from 50hz to 12,500hz get to be pushed harder.
to me, this is a sad state of affairs in the world of sound recording.
|
|
|
Post by ionian on Aug 19, 2013 3:06:32 GMT -6
But what you can't "hear" you may be able to feel. That goes for harmonics and transients. It's definitely been proven. I can't find it now but there was a study done that showed people's brains on a brainscan with all the colors and they played various frequencies above 20 kHz that we can't hear and they showed on the brainscan that different parts of the brain would change color on the scan with activity when it happened showing that even though we can't hear above 20 kHz, the brain is still picking it up and responding to it. As for me, I've been recording 24 bit / 96 k for the longest time. It works for me and I've don't feel a need to change it. Regards, Frank
|
|
|
Post by jazznoise on Aug 19, 2013 9:18:13 GMT -6
It's definitely been proven. I can't find it now but there was a study done that showed people's brains on a brainscan with all the colors and they played various frequencies above 20 kHz that we can't hear and they showed on the brainscan that different parts of the brain would change color on the scan with activity when it happened showing that even though we can't hear above 20 kHz, the brain is still picking it up and responding to it. As for me, I've been recording 24 bit / 96 k for the longest time. It works for me and I've don't feel a need to change it. Regards, Frank Do you've any specifics I could use to search with? There's no papers I can find on this. The few papers I have found involve bone conduction, which is an extremely insensitive element within our hearing apparatus and were totally unrepeatable- so even if abstract, the test is not proof of anything anyway. There are benefits to recording at higher SR's, as we've discussed. But I don't believe ultrasonics are one of them and there is no evidence to suggest it. That said, there are studies on ultrasonics being used to alter brain activity: www.ncbi.nlm.nih.gov/pubmed/21354315But the mechanisms involved have nothing in common with music reproduction equipment. Again, the music would have to be deafeningly loud. My Bloody Valentine talked about this at one time, but with onstage volumes of 120dB I guess that shaking feeling in your skull might be relevant.
|
|
|
Post by tonycamphd on Aug 19, 2013 9:55:51 GMT -6
I have probably a dumb question, but what bit depth would be required, to record the quietest "audible" passage at a minimum of 21 bit resolution?
thanx T
|
|
|
Post by svart on Aug 19, 2013 13:12:54 GMT -6
Tony, that really depends on the preamp settings.. Bit depth is only relative to the amount of signal going in. This would mean that you would have to calibrate a preamp and known level into the converter to determine the level you seek. It would also be a sliding scale depending on your preamp gain.
|
|
|
Post by tonycamphd on Aug 19, 2013 13:58:44 GMT -6
yes svart, i understand, what i mean is, assuming proper gain staging, take an acoustic guitar and play some one strum chords with long sustain, at the initial striking of the strings, your meter will hit 0 and you utilize full bit depth right? when the sound decays just to the point before inaudibility, the bit depth will be reduced way under full resolution right?
So what maximum bit depth would be required to get the quieter passages sampled at higher bit depths?
or like you said, is this not useful as bit depth is a representation of dynamic range?, just would like to clarify this in my nugget, or am i just plain koo koo? 8/
thanx T
|
|
|
Post by svart on Aug 19, 2013 15:37:10 GMT -6
Well, lets say you have a 24db converter that you are getting an ENOB of 20 bits. That means 20 bits of usable dynamic range as the A/D sees it. Now.. Here is the mindbender and what I alluded to earlier.. When you design conversion systems, you need to decide what your usable dynamic range *will be*. This determines how much gain or loss you want to put on the front end and determines the sweet spot/median of the range. All in all and avoiding the math equations that determine these things, 24 bits will theoretically give you around 144dB of SNR, or slightly better than the human ear can hear, which is 140dB. That's about 6db per bit. There are other things that lessen this amount. Amplifiers add their own noise to circuits (Noise Figure it's called) and this looks roughly like an additional noise floor increase. Attenuators can attenuate loud signals but you can eventually be limited by the system noise floor or the thermal noise floor (both determined by the parts used and the junk noise flying around on your system) which can also limit your dynamic range. It's a real balancing act.
You can set your range so that you can pick up impossibly low signals but you'll compress earlier on higher level signals. Or, you can set your range so that you pick up very large level signals just fine but your faint sounds will be in the noise floor. Doesn't matter, your range is still 144dB and will always be, but *what* you want to put in that range is completely up to you and your amplifier/attenuator settings.
Now if I understand this question right: "when the sound decays just to the point before inaudibility, the bit depth will be reduced way under full resolution right?"
you are asking if the bit depth (lets say 24bits) is lessened somehow? It's not. 24 bits is always 24 bits, but we lose a few bits to noise, so while the converter might actually be sampling those "lost" bits, it's sampling noise, with your inaudible signal mixed in.
|
|
|
Post by tonycamphd on Aug 19, 2013 15:59:14 GMT -6
great stuff, one more question and i'll give you peace lol! Is their any advantage to 32 bit? From what i loosely understand, because you don't have to worry about digital clipping as much, you can drive your input on your daw closer to 0 and take advantage of more resolution? If this isn't the case, what am i missing?, and why make 32 bit?
|
|
|
Post by wiz on Aug 19, 2013 16:19:48 GMT -6
Here's the thing.
What is the dynamic range of the music you are recording?
That is, the difference in dB between the loudest and softest parts of the song?
Work that out, then you will see that 24 bit has waayyyyyyyy more dynamic range than what you are recording 8)
I will find the link to a great movie about how digital audio works, it will explain everything for you.
You also need to learn about why hitting up around 0dbfs is not the best thing for you to be doing.
If, you were to adjust your preamps for the tone you want out of them with the mic position you are using, and ensure that that signal never peaked above -18dB or so ( which in some cases might require a pad if you are driving the hell out of your preamps.) and you focused on getting the signal chain prior to the A to D covert or operating so that each of the pieces is operating in its sweet spot for the tone you are chasing. Then I can assure you the differences between 48,44.1,96khz 24 bit vs 32 bit recording depth , dither type etc will have far less impact than the fact that you use a sm57 on that rack Tom vs a audio technic atm 25.
Cheers
Wiz
|
|
|
Post by wiz on Aug 19, 2013 16:32:20 GMT -6
|
|
|
Post by wiz on Aug 19, 2013 17:16:05 GMT -6
I just watched part of that video again, I love it. The way its presented is both entertaining and incredibly informative.
That video, should be required watching before anyone is allowed to talk about sample rate and bit depth on newsgroups LOL....
Pay attention to the bits bout, stair steps, bit depth etc
cheers
Wiz
|
|
|
Post by jazznoise on Aug 19, 2013 17:47:13 GMT -6
Great stuff, Wiz! I do like that he corrects the misconceptions of PCM audio. Staircases are a S+H phenomenon - it's more something you find in an ARP 2600 or an MXR Bluebox than a Digi 192.
Here's something of interest:
this is the input of all modern ADC's. The higher order versions probably seem ridiculous until you remember that higher frequencies will be attenuated even more extensively by an LPF. Assuming you're sampling above 500Khz, this noise is easily removed.
From this you could make an argument that ADC fidelity will improve at lower sampling frequencies - but in practice the Johnson noise from the analogue electronics will largely dominate this issue and the ADC stage will inevitably have been designed with the highest SR in mind.
|
|
|
Post by sozocaps on Aug 19, 2013 18:01:51 GMT -6
Great video...!!! On the subject of a square wave It seems the original square wave would be more pure if we raised the sample rate thus raising the anti-aliasing filter this having more odd harmonics to calculate ?
|
|
|
Post by wiz on Aug 19, 2013 18:07:40 GMT -6
I like that video jazz noise, hows the irony of the clicks and pops in the audio 8)
cheers
Wiz
|
|
|
Post by jazznoise on Aug 19, 2013 18:23:42 GMT -6
I like that video jazz noise, hows the irony of the clicks and pops in the audio 8) cheers Wiz I know! I also think she's quite under-rehearsed. But the amount of white papers and videos they've done are great - Op Amps For Everyone is a great user guide for Operational Amplifiers.
|
|
|
Post by tonycamphd on Aug 19, 2013 20:07:17 GMT -6
From the mouth of Bonnie Baker, you heard it, digital converters "Decimate" analog signals! nuff said lol!
the pops and clicks were indeed funny, but none the less, informative video for sure.
|
|
|
Post by Martin John Butler on Aug 19, 2013 21:19:55 GMT -6
48K/24 bit. I used to track at 44.1, but did a track in 48 for a video, and I liked the sound much better. I realize for a standard CD that requires some dithering, unless I can figure a workaround like cowboy does. Thing is, I just don't care, I like what I hear in 48, so I'll work with it for now. Once I get a new computer, I'd love to try 88.2 or 96/32 like Tony uses.
|
|
|
Post by wiz on Aug 19, 2013 21:43:21 GMT -6
48K/24 bit. I used to track at 44.1, but did a track in 48 for a video, and I liked the sound much better. I realize for a standard CD that requires some dithering, unless I can figure a workaround like cowboy does. Thing is, I just don't care, I like what I hear in 48, so I'll work with it for now. Once I get a new computer, I'd love to try 88.2 or 96/32 like Tony uses. Dither, isnt the word you are looking for, when you change sample rates, say from 48Khz you are recording at, to 44.1Khz for putting to CD, its called Sample Rate Reduction. cheers Wiz
|
|
|
Post by Martin John Butler on Aug 19, 2013 21:47:21 GMT -6
Hey, thanks Wiz, I'm learning as I go here.
I once heard or felt, I'm not sure which, a 5-10hz tone while trying to record. I was doing guitar solos on my bands album, (The Demons, Mercury records), and kept hearing something. I was in one of the worlds finest studios, and the world class producers and engineers treated me like I was some punk, wasting their precious time, when they wanted to go home, when I mentioned something about the sound was bothering me. After 20 minutes of trying to work around it, I stopped the session, and refused to go on until they found out what the sound I was hearing was. The tape tech discovered after 25 minutes of checking the machines, and a lot of rolled eyes, that an ultra low 5-10Hz tone was being printed on all the tracks, which would have made a record needle jump, and could potentially have caused a 100,000 album recall. They shut their wise ass mouths up after that, and I've trusted my ears ever since. I was 21 years old.
I do believe we don't know all there is to know about human perception yet. Do people really believe we don't use ALL of our brain for something? I think we perceive audio in unfathomed ways, and think we're sensitive to frequencies well beyond the hearing range. So all the, "people can't hear above frequency X anyway" kind of talk means didley to me. I'll go with what sounds right to me, until my last breath.
|
|
|
Post by Martin John Butler on Aug 19, 2013 22:02:13 GMT -6
Watch, someone will eventually prove we feel colors.
|
|
|
Post by lolo on Aug 19, 2013 22:52:02 GMT -6
48K/24 bit. I used to track at 44.1, but did a track in 48 for a video, and I liked the sound much better. I realize for a standard CD that requires some dithering, unless I can figure a workaround like cowboy does. Thing is, I just don't care, I like what I hear in 48, so I'll work with it for now. Once I get a new computer, I'd love to try 88.2 or 96/32 like Tony uses. Dither, isnt the word you are looking for, when you change sample rates, say from 48Khz you are recording at, to 44.1Khz for putting to CD, its called Sample Rate Reduction. cheers Wiz Got a bit lost in this thread, way over my head. What sample rates do you mix and record at Wiz?
|
|
|
Post by tonycamphd on Aug 19, 2013 23:20:17 GMT -6
I have to agree with MJB on the larger point, music is pure magic, it captured me a long time ago, and never let go, my belief that understanding electronics better, would ultimately help me make better music is why I took interested toward it, but as great and useful a tool that knowledge is, I think it is a mistake to try to quantify the human listening experience with a measurement from an oscilloscope IMO No matter what anyone tells me, i trust my ears first, and i hear an improvement in my audio at 96/24 over 44.1/24
|
|
|
Post by wiz on Aug 20, 2013 1:00:13 GMT -6
Dither, isnt the word you are looking for, when you change sample rates, say from 48Khz you are recording at, to 44.1Khz for putting to CD, its called Sample Rate Reduction. cheers Wiz Got a bit lost in this thread, way over my head. What sample rates do you mix and record at Wiz? i record and mix at 44.1Khz 24 bit for stuff I record myself. Stuff I mix for others, I mix at whatever the samplerate that comes in is. Again, if it works for you... do it. Even if its not real, measurable but you feel something is better , go for it. If you think it makes you play or sing better go for it. Thats my artist/producer side talking there. Never get in the way of the perfomer. Now, my scientific and wallet side says, when you aint being artistic, check these things out for yourself, learn and go seek the information to understand all you can, then decide for yourself. Try and seperate the wheat from the audio chaff. If for example, and I aint picking on you Martin, if, I was producing Martin, and he felt his really expensive power cord or 96Khz made a difference, I wouldnt say a word and just let him get on with the job. Now, If Martin and I ran a studio together and we were discussing what power cords to buy to run our business, we would have a different discussion 8) and then also, if there was a point where he showed me and I felt the same as he did, well then I would accept that too. cheers Wiz
|
|
|
Post by wiz on Aug 20, 2013 1:02:54 GMT -6
I have to agree with MJB on the larger point, music is pure magic, it captured me a long time ago, and never let go, my belief that understanding electronics better, would ultimately help me make better music is why I took interested toward it, but as great and useful a tool that knowledge is, I think it is a mistake to try to quantify the human listening experience with a measurement from an oscilloscope IMO No matter what anyone tells me, i trust my ears first, and i hear an improvement in my audio at 96/24 over 44.1/24 There is absolutely every chance, that your audio interface functions in such a way at 96 vs 44 that you like it. Absolutely. There is also every chance, that if you took a project to completion in both 96 and 44 kHz sample rates and the final medium was CD, that you wouldnt be able to tell the difference listening on systems other than your studio systems. So what does that mean? It means, if you like it, hear it believe it, use it. Just dont think for a second, that ALL interfaces operate better or indeed differently (sonics results speaking) at 96 over 44.1 cheers Wiz
|
|