|
|
|
02-06-2013, 08:19 AM
|
#2241
|
Human being with feelings
Join Date: Jul 2010
Posts: 387
|
And we have our new flmason.
|
|
|
02-06-2013, 08:22 PM
|
#2242
|
Human being with feelings
Join Date: Aug 2006
Posts: 2,019
|
Condensing a lot of stuff...
Quote:
Originally Posted by northern
...Somebody did put hes "nerd glasses" on and thought "now we are going to change key here and scale within the key-change..." and people still love the song...
|
Yeah, I don't think you are correct at all, here. The guy you are citing as putting "his nerd glasses on" is James Jamerson, who was notorious for not being able to speak English clearly, for ignoring the written music and literally throwing away the lead-sheet, for playing his best basslines drunk to the point where he had hard time knowing where he was, etc.
"What's Going On" is like a default pick for the best, or at least in the top 10 bass performances of all time, and it was played by a guy who was too drunk to even sit in a chair. If being dragged on your back into Berry Gordy's dirt-floor basement is what you mean by "putting on your nerd glasses", I'm with you, but I doubt that's what you meant.
James Jamerson cut more #1 hit records than the Beatles, Elvis Presley, The Rolling Stones, and The Beach Boys, COMBINED. We are not talking about a disconnect between musical quality and popularity here, unless you want to question Jamerson's quality (his bass is the absolute definition of "hit bass", he's the original hit-maker, probably the single best-selling musician of all time).
Quote:
Again, is that problem? If ten million fans are expecting artist to make more complicated and technical stuff than anything they have ever heard before, then you probably get better result if you put your "nerd glasses" on and people will love the result. On next day when you are going to record that pretty looking boy band you probably want to take those glasses off.
|
See above. The best and most successful popular bassist of all time still kicks almost everyone's ass in terms of musical complexity and harmonic/melodic fluidity, and he played with one finger and didn't give spoken interviews because he had a speech impediment, and didn't give written interviews because he was barely literate.
Quote:
...I think things like "how technical and complex your material is" are purely artistical chooses.
|
That's fine, but I could not disagree more. I generally do not think that technical difficulty has much bearing whatsoever on artistic quality. I don't generally think that a player-piano roll is made better by speeding up the playback. I tend to think that musical and artistic quality is facilitated by technical skill, but not conscripted by it.
I do think that virtuosity and human ability to play fast and difficult passages has value, but only to the degree that human musicians are better and more expressive than MIDI samples.
There is something to be said for the fastest player or the highest sung note, but I do not think the Guinness book of world records is the best measure of musical or artistic quality.
On a side note, I should really stop replying to this thread. We're way off-topic, and deep into sub-forum-type stuff that has nothing to do with recording techniques. Once upon a time, this thread was people trading useful advice on practical recording techniques. Now we're just arguing over bass-players in a bizarre pseudo-sub-forum. It should have been locked a long time ago. People should be discussing this stuff in other threads.
Last edited by yep; 02-06-2013 at 08:52 PM.
|
|
|
02-07-2013, 01:25 AM
|
#2243
|
Human being with feelings
Join Date: May 2008
Posts: 678
|
I just stopped by to say that Jamerson is truly a godlike musician and as long as you channel little Jamerson into your music, it definitely won't sound like ass
|
|
|
02-07-2013, 04:28 AM
|
#2244
|
Human being with feelings
Join Date: Dec 2010
Location: in your head
Posts: 258
|
I just came to thank you Yep for sharing your knowledge with as ( and me ). It helped me a lot! You made a big difference in my life. I owe you one in this life time or the next
__________________
you don't wanna know
Last edited by MesS!er35; 02-07-2013 at 04:36 AM.
Reason: grammar
|
|
|
02-08-2013, 07:21 AM
|
#2245
|
Human being with feelings
Join Date: Jan 2013
Location: Finland
Posts: 19
|
Quote:
Originally Posted by yep
That's fine, but I could not disagree more.
|
Well, let's just agree that we disagree
You asked criticism and I did criticize those comments that I didn't agree, but like you said this things are off-topic and maybe I shouldn't have comment them here, so sorry about that.
Here are some my other thoughts that are hopefully more useful to others than comparing James Jamerson and Paganini...
When you are mixing it is always easy to put your tracks louder. When I did my first mixes I had also that habit "let's put bass louder... and drums... and guitars... and vocals... and bass... and drums..." until master was clipping. In some point I started to think that if the overall loudness of the song will be the same, it means that if I put one track louder all other tracks are getting quieter. If I lower the volume of one track all others are getting louder... Instead of thinking "what tracks should be louder" I try to think "what tracks could be quieter". If I can't hear bass enough I try to listen why. Are guitars too loud, maybe bass drum is messing up low frequencies...
I like to have master limiter in my project at quite early stage. That way I have the "roof" and I can't go over it. If I want something louder, I must make something quieter first. It forces me to think that what stuff is really necessary and what stuff doesn't have to be so audible. It also tells me that there might be something wrong in my mix if I need to listen it much louder than I'm usually listening music from my computer.
I also once heard one professional saying that "vocals get louder when adding master limiter to song..." I don't know is that true, my ears can't really hear the difference, but I guess it will more probably do something good than harm to have master limiter while mixing. It's little like having highpass at 40Hz in your bass track... you really can't hear difference, but more likely it's going to do good than harm.
Somebody mentioned something about EQ that keeps the overall volume of track at same level even if you boost or cut some frequencies. I once had plugin like that, but unfortunately I have removed it and I can't remember the name. It's probably still out there somewhere, try google it.
|
|
|
02-08-2013, 01:24 PM
|
#2246
|
Human being with feelings
Join Date: Jan 2006
Posts: 1,595
|
Yep, you have a way of talking about audio that, besides being brilliant, is very personal and engaging. You've had a big impact on many readers... half a million thread views... PDFs of your collected writing...
Perhaps you engage with argumentative posters because you care so much about the subject. It occurs to me though, that should you continue posting your thoughts, when future PDFs of your posts are collected, the arguments of others likely won't be included. So here's to the hope that you continue this beautiful thread and don't get pulled into arguments that aren't worth having.
|
|
|
02-09-2013, 07:55 AM
|
#2247
|
Human being with feelings
Join Date: Feb 2013
Posts: 5
|
'Yep', I just registered with this forum because I simply had to post a "thank you" message to you. I am starting to set myself up to get "into music production" and kept feeling that there's just simply no way to cover all the needed basics. I am somewhere on page 6 of this thread now and am sitting here all giddy and giggly to have found this absolute gem of a thread.
Thank you so very much for taking the time to share all this with the rest of us. I feel truly humbled by your generosity.
|
|
|
02-12-2013, 08:18 AM
|
#2248
|
Human being with feelings
Join Date: Feb 2013
Posts: 5
|
Ok, so now I have two questions to ask:
1st)
I (apparently only 'somewhat') understand that you can measure the generated pressure in air (by a sound source) in db. Then absolute silence is 0db and anything that makes a noise will compress the air and generate a positive db value. Starting jet 200db and whatever.
Also there's the apparent inversion of this concept, where absolute silence is expressed as "- infinity" and maximum output of the equipment is described as 0db. However, this is obviously only partially correct, since the scale usually goes from -inifitiy up to +12db or something!? So what's with the surplus of db above 0 on that scale!? I would assume that 0db is 'maximum', i.e. anything above that will not be supported by the equipment in use? Hopefully someone can shed some light on this for me.
2nd)
I (apparently only 'somewhat') understand bit depth and sampling rate. I understand that - if your hardware supports it in terms of space and speed - highest bit depth and sampling rate is the goal when recording an analog signal into your system. However, does this also hold true when recording only MIDI generated audio? The scenario I'm referring to is:
I setup a project in my DAW of choice. I will only use MIDI software-synthesizers. Does the sampling rate and bit rate still matter for this project? Is the signal that the virtual synthesizer controlled by my MIDI signals generates just a regular audio signal?! It can't be, right? Since it's all digital there is no AD/DA conversion, correct? The synthesizer produces a digital signal and that gets recorded 'as is' in the DAW; or does it!? I'm confused...
Thanks a lot in advance to anyone taking the time to respond to help me.
|
|
|
02-12-2013, 10:24 AM
|
#2249
|
Human being with feelings
Join Date: Jul 2010
Posts: 387
|
Hey, groovester.
To answer your first question, there are several different scales used to measure dB (dBFS, dBv, dBSPA, etc.), and they aren't measuring the same thing. We have to pay attention to which scale is being used, because dB by itself is not a measure of anything - it must be referenced to something else.
- dBFS is decibels full scale, which is what your DAW and other digital recorders/sources/FX use to measure RMS volume of a wave. In this scale, 0 dB is the loudest a sound can be without distorting, and quieter sounds are referenced in terms of negative dB down from there.
- In analog equipment, including mic preamps, EQs, etc., you will normally see dBv, which is a measure of RMS (root mean squared) voltage in an electrical circuit. Typically this is referenced to a VU meter such that a signal shows 0 dB on that meter at either +4 dBv (pro gear) or -10 dBv (consumer gear).
- dB SPL (sometimes written as just dB) is a measure of sound pressure level, and is referenced to the human ear - the lowest volume sound a human can hear is 0 dB SPL. This is roughly equivalent to the sound of a mosquito flying 10 feet away from the listener in a silent room. While there is a lower limit on volume in this scale, there is no upper limit, so variations in volume will be shown as positive dB above that lower limit (0dB).
For your second question, when you convert MIDI data to audio in your computer, your computer does the conversion rather than your A/D converters, but the MIDI is still being converted to audio, so the sample rate and bit depth will still matter. I recommend you record at a 24-bit bit depth and whatever sample rate you prefer. 44.1K is the standard for CD audio, 48K is the standard for digital video, and you will likely find that most pro studios record at either of these sample rates rather than the higher ones. If you want to record at a higher sample rate, feel free to do so if your harware supports it, but it is not necessary in order to acheive professional quality results.
|
|
|
02-12-2013, 10:53 AM
|
#2250
|
Human being with feelings
Join Date: Feb 2013
Posts: 5
|
drtedtan,
thank you very much for your answer! That definitely cleared up a lot of the mist for me and I'll go and read up some more on the dB scales you referenced.
If I may prod a little further on the sampling-rate/bit depth part:
If we're pretty much always consuming music that was (if digitally recorded/produced) ultimately sampled at 44.1kHz or 48kHz then why are so many devices/DAWs offer sampling rates of up to 192kHz? Is it just a marketing ploy? Or is there any sense in working at, say, 96kHz and then do the final export to 44.1/48kHz? I'm being genuinely curious!
|
|
|
02-12-2013, 01:01 PM
|
#2251
|
Human being with feelings
Join Date: Jul 2010
Posts: 387
|
Good question, groovester. Before I answer that, let me say something I intended to include in my previous post, but somehow left out.
You remember that +4 dBv is the level of pro audio gear (hardware), right? This is equivalent to -20dB Full Scale. Why do we care? Because hardware (and many software plugins) are designed to work best at this level because it leaves plenty of room for them to do their job without adding unwanted distortion along the way. This is why you see Yep, in this thread, and others recommending that people record at a -18 or -20dBFS level. Your FX plugs (and any hardware you might want to send your tracks out to) will work better at this level.
Now to your question regarding higher sample rates. There are certainly legitimate reasons to work at higher sample rates, just know that we are opening a can of worms here.
Some people believe that high sample rates sound better for one of a few reasons. Some believe it sounds better because your A/D converter has a very steep low pass filter which cuts all sound above a certain frequency. These people believe that at 44.1K, the filter creates a ringing that is audible in the range of human hearing, but by increasing the sample frequency, the frequency of the ringing is raised up high enough that it is out of the range of our ability to hear it.
Other people believe that higher sample rates allow you to capture frequencies that are too high for us to hear themselves, but that interact with the lower frequencies to create a better, more natural sound.
Some people have an A/D converter that sounds better at a certain sample rate than it does at other sample rates so they use the better sounding one (this is hardware specific - I recommend you test your hardware at its various rates to see if you hear any noticeable differences).
And some people record at higher sample rates because they want to archive their work at the highest quality they can to "future proof" their work in case a higher standard is established in the future.
I'm sure there are plenty of other reasons, too. Just note that recording to 24 bit (instead of 16 bit) yields a good boost in quality for very little extra computer storage/processing resources used. Going up to the higher sample rates will cause a greater strain on your computer for (in most cases) very slight improvements in sound quality, so if you question your computer's capabilities in the slightest, its probably not worth the added computer overhead (but test your A/D converters anyway to see if it makes a difference in your specific case).
So to sum this up, I wouldn't say that high sample rates are just marketing hype, but neither would I say that using them will make a significant difference in most people's recordings.
|
|
|
02-12-2013, 01:18 PM
|
#2252
|
Human being with feelings
Join Date: Jan 2006
Posts: 1,595
|
Quote:
Originally Posted by drtedtan
+4 dBv is the level of pro audio gear
|
Do you mean +4 dBu?
|
|
|
02-12-2013, 02:20 PM
|
#2253
|
Human being with feelings
Join Date: Jul 2010
Posts: 387
|
Quote:
Originally Posted by Timothy Lawler
Do you mean +4 dBu?
|
Yes, because dBv (with the lower case v) and dBu are the same thing.
dBV (with the capital V) is different and yet another measurement involving decibels.
Confused yet, guys?
|
|
|
02-12-2013, 03:29 PM
|
#2254
|
Human being with feelings
Join Date: Dec 2007
Location: Walnut Creek, CA
Posts: 805
|
Quote:
Originally Posted by groovester
drtedtan,
thank you very much for your answer! That definitely cleared up a lot of the mist for me and I'll go and read up some more on the dB scales you referenced.
If I may prod a little further on the sampling-rate/bit depth part:
If we're pretty much always consuming music that was (if digitally recorded/produced) ultimately sampled at 44.1kHz or 48kHz then why are so many devices/DAWs offer sampling rates of up to 192kHz? Is it just a marketing ploy? Or is there any sense in working at, say, 96kHz and then do the final export to 44.1/48kHz? I'm being genuinely curious!
|
It's a marketing ploy and obviously a very effective one. As far as I know there has been no objective verification that humans can hear the difference. But the ploy works because we almost always prefer bigger numbers. The other part of the ploy is the idea that it might improve things and it's basically free now that processing power and storage are so inexpensive, so why not?
Watch the responses to see how effective the ploy has been.
Fran
|
|
|
02-12-2013, 05:29 PM
|
#2255
|
Human being with feelings
Join Date: Jan 2007
Location: Silicon Gulch
Posts: 544
|
.
Fran, don't be silly, of course the higher sampling rates help, it helps me imagine I can hear all those delicate highs up around 48Khz.
Now they need to improve Television too! No reason why TV cannot show infrared and ultraviolet colors, so we can appreciate the HiRes live lightwave experience! Imagine feeling the warmth of the Sun and actually getting a sunburn while watching Hawaii 5-0 !!!
.
__________________
Inundated by a Perfect Storm of Gluten-Free Artisanal Bespoke Quinoa Avocado-Toast Toilet Paper.
Mahope Kakou (Later Dudes)...
|
|
|
02-12-2013, 08:00 PM
|
#2256
|
Human being with feelings
Join Date: Aug 2008
Posts: 251
|
44.1kHz? We should all record at 11kHz. 44.1k is for gullible chumps.
:P
|
|
|
02-12-2013, 08:13 PM
|
#2257
|
Human being with feelings
Join Date: Nov 2012
Location: 'straya
Posts: 9,409
|
There was a sample rate article just recently linked in a post by ChrisHarbin, seems theres some kind of scientific concensus that the 'optimal' sample rate lies somewhere [from memory] between 60 and 70. Nearest standard then [in that case] being 88.2 - 96kHz
Edit: found it
http://trustmeimascientist.com/2013/...-when-it-isnt/
Last edited by morgon; 02-12-2013 at 08:21 PM.
|
|
|
02-13-2013, 03:02 AM
|
#2258
|
Human being with feelings
Join Date: Feb 2013
Posts: 5
|
Thanks for all the replies! I can see that I'm not the only one confused about sampling rates.
The more I think on this sampling rate business the more confusing it's getting, tbh. Bit depth is easy to understand. But trying to get the concept of our ear being "able to pick up frequencies between 20Hz and 20kHz and interpreting them as sound" together with how a signal is being recorded digitally and what the reciprocal effects are between "hearing" and "sampling (and re-playing)" feels a bit mind boggling.
I have to think and read some more on this...
I'll be back later, surely with another question.
Last edited by groovester; 02-13-2013 at 03:07 AM.
Reason: Minor grammatical correction.
|
|
|
02-13-2013, 04:16 AM
|
#2259
|
Human being with feelings
Join Date: May 2008
Posts: 678
|
I personally use 44.1 kHz sample rate and find no need to go higher. I know that the optimal would be 50-60 kHz and that is what audio professionals wanted back in the early 80s when Sony and Philips decided that 44.1 kHz is "enough". That is perhaps one reason to use higher rates (88.2k or 96k). I personally have no need for that last fraction of an inch.
If you are using plugins in your DAW, some of them work better with higher sample rates, especially the older ones. EQs, synths, dynamics and amp sims could really benefit from it. But of course if you use good plugins with internal oversampling and such, there is no need for that, they already work "perfect" in 44.1k rate.
Also it is good to know that some older plugins may sound completely wrong in sample rates they are not intended to be used.
Another plus for higher rates would be that latency gets smaller. for example if you use 128 samples latency in 44.1k (44100 samples per second), it is about 2.9ms, but in 96k it would be 1.3ms, roughly. But with small latencies today this isn't so important anymore.
The cons of using higher sample rates are quite big: the bigger the rate, more processing power that audio needs. If you double your rate, you practically double your CPU use. It is also more difficult for converter to do quality conversion in higher rates and that is just physics.
192k is total scam to sell even more gear. But as you may have noticed, that hasn't really happened, gladly. Marketing has turned to preamps and such and it isn't about that sample rate anymore.
EDIT: one thing I'd like to add about sample rate and digital audio: people often don't quite understand that this jagged cubistic and stepped view of waveform that we see on our computer screens isn't the audio we hear from computer after it passes through DAC (digital-to-analog conversion). If you put analog oscilloscope to the outputs of your soundcard or interface, you just see smooth smooth curves. That is what DAC does. There is no steps or "pixels" in real world (well, perhaps on quantum level but that's whole another ballgame). That is why 44.1kHz digital audio can contain up to 22.05kHz waves, because two samples are enough to contain it. And when it is converted to analog it is identical with the original wave. Not a cubistic one but smooth with peaks and valleys. In practice, 44.1kHz DACs don't go this high because of anti-alias filtering in them, but still, the sampling theory is solid fact and higher rates don't make any difference. They only help you record frequencies that no one can hear. Also do keep in mind that most microphones can't record those frequencies anyway, and also most amps and speakers don't reproduce then either.
Last edited by gavriloP; 02-13-2013 at 04:44 AM.
|
|
|
02-13-2013, 06:50 AM
|
#2260
|
Human being with feelings
Join Date: Feb 2013
Posts: 5
|
gavriloP, great stuff, thanks!
|
|
|
02-13-2013, 07:27 AM
|
#2261
|
Human being with feelings
Join Date: Jul 2010
Location: Online
Posts: 4,896
|
Why do your recordings sound like ass?
They don't!
I read this thread and gained the knowledge I needed to completely EQ the ass outta my tunes.
__________________
it aint worth a bop,if it dont got that pop
|
|
|
02-16-2013, 12:02 PM
|
#2262
|
Human being with feelings
Join Date: Jan 2006
Posts: 2,173
|
Quote:
Originally Posted by Fran Guidry
Watch the responses to see how effective the ploy has been.
|
Not a ploy here, pure Acoustic music will always benefit from higher sample rates (88 or 96k) because the high end is represented better.
For all other music 44.1 is totally fine....
__________________
Yep's First 3 Years in PDF's
HP Z600 w/3GHz 12 Core, 48GB Memory, nVidia Quadro 5800, 240GB SSD OS drive, 3 480GB SSD Sample/Storage drives, 18TB External Storage, Dual 27" Monitors
|
|
|
02-16-2013, 01:07 PM
|
#2263
|
Human being with feelings
Join Date: Dec 2007
Location: Walnut Creek, CA
Posts: 805
|
Quote:
Originally Posted by Smurf
Not a ploy here, pure Acoustic music will always benefit from higher sample rates (88 or 96k) because the high end is represented better.
For all other music 44.1 is totally fine....
|
So you'll supply clips of the same source recorded at the same time that demonstrate this difference?
Fran
|
|
|
02-16-2013, 04:00 PM
|
#2264
|
Human being with feelings
Join Date: Aug 2006
Posts: 2,019
|
Hopefully someone will spin off a new thread dedicated to rigorous discussion of sample-rate (and/or bit-depth), because the debate is ongoing, hotly-contested, and can get into difficult mixtures of highly-technical arguments versus expert-but-subjective ones.
A very short version, relevant to this thread, is that your recordings do not sound like ass because of the sample-rate you are using. It is possible that they might sound better if you were using a different one, but we are mostly talking about stuff on the margins, here.
Some high-level points:
- CD audio is basically perfectly fine for well-made, real-world records. That is, well-made 16-bit/44.1kHz recordings played back on high-quality systems, contain everything that matters, with all the fidelity that real-world listeners are ever going to hear, and then some, for purposes of music-recording. (scientific measurements, audio-forensics, and extreme 99.99th percentile audiophiles and/or synthetic tests excluded). So if your recording sounds like ass, it's usually something else.
- 24 bit recording is always recommended (when available) for tracking, mixing, etc, purely because it allows you to record with high-headroom and extremely high resolution without having to worry about even theoretical loss of fidelity. That is, 999 times out of a thousand, nobody will ever year the difference between music tracked at 16-bit vs 24-bit, but recording at 24-bit allows you to leave plenty of headroom and just not worry about it. You'd have to record at like -50dB peak before you lost any resolution relative to CD quality, so 24-bit allows a huge margin of safety, at the relatively cheap price of 50% increase in disk-space and bandwidth. It is recommended not because it necessarily sounds better (it's almost always indistinguishable), but because there is absolutely zero sonic downside, minuscule cost/performance downside, and the significant practical benefit of never having to worry about headroom/clipping (just set your record levels low and go).
Sample rate is a slightly more complicated question. Setting aside some scientifically dubious (but vehemently-argued) "audiophile" arguments, 44.1kHz is theoretically capable of accurately capturing and reproducing everything that human beings can hear. But there are still some technically-sound reasons to consider higher sample rates. Unfortunately (and unlike higher bit-depth), those reasons do have significant trade-offs in real practical terms, and potentially in sonic/theoretical terms:
- Higher sample rates have a significant performance impact on the CPU, and so are more expensive than higher bit-depth. Doubling the sample-rate generally takes about twice the processing power, which significantly effects things like plugin and track-count. In native "in-the-box" DAW recording, this is typically a much more significant bottleneck than bit-depth, for a bunch of reasons. So there is potentially a non-trivial real-world "cost" to using higher sample-rates.
- Unlike bit-depth, there can actually be a theoretical downside to using higher-than-necessary sample rates, in terms of audio quality. An old rule of thumb was to stick to exact multiples of the target sample rate while tracking(e.g., 88.2 or 44.1 for CD releases, 96k or 48k for DVD releases), since sample-rate conversion ("SRC") is not perfect (SRC has gotten much better in recent years).
- Similarly, higher sample rates can, in some cases, produce greater jitter on some converters (jitter is a kind of timing error that can produce ugly, brittle, "digital" sound), especially if outboard clocks are used.
So unlike bit-depth, higher is not always and automatically "better-or-at-least-as-good", when it comes to sample rate. (OTOH, 24 bit is categorically never worse than 16-bit). The specific arguments get extremely technical around the edges, and their relevance to real-world record-makers is doubtful, but they do exist. That said, there are some real (sometimes marginal, but still real) reasons to consider recording at higher sample-rates:
- Before your audio is converted to digital, it goes through an analog filter that "cuts out" everything above the sample-rate, to avoid digital noise/distortion from supersonic harmonics and artifacts (basically an extremely steep bandwidth-limiting EQ). The quality of this filter is part and parcel of the overall ADC quality, and some are better than others. Higher sample rates can push that filtering up into extreme supersonic frequency ranges, and theoretically can improve sound-quality at the high-end of the audible spectrum.
- Plugins and digital processing can produce "inter-sample" errors or distortions that can effect the audible spectrum. To over-simplify: digital processing that is trying to draw a "curve", sometimes has to "guess" at what happened between extreme high-frequency samples, and incorrect "guesses" can sometimes lead to audible distortions after processing. Higher sample rates tend to result in fewer and less-objectionable distortions.
Note that everything related to sample-rate differences tends to specifically and exclusively affect the extreme limits of frequency and dynamics. For most people, the differences will tend to be theoretical (if that), not audible.
What makes sample-rate different from bit-depth is that higher bit-depth (24-bit):
1. Comes at a minuscule cost in terms of price, performance, and CPU load (especially when already working in a high bit-depth, floating-point audio engine like REAPER's), and;
2. Has absolutely zero sonic downside, even in potential/theoretical terms, and;
3. Has a stupidly obvious practical benefit (ability to record at lower levels, with high headroom and no loss of fidelity).
Sample-rate is a more complicated question, with more specific budget- and setup-specific technical considerations, where "higher" is not only more expensive, but is also sometimes less-good (i.e., the theoretically "ideal" sample rate is a situation-dependent question).
Both of these are like 100,000 steps down the ladder of importance from things like making sure your guitar's bridge and truss-rod are properly-adjusted, and creating a good headphone mix. In practical terms, using modern gear and otherwise-sound recording practices, we are into the realm of talking about whether the singer should trim their eyebrows and wear a less-absorbent shirt to improve the high-frequency early reflections. It's not necessarily bogus, but it's a little bit la-la land for spare-bedroom recordings to worry about this stuff.
TL;DR: record at 24bit always, if you can. If your computer can easily handle it, go ahead and record at 88.1 for CD releases, and 96k for DVD or high-res audio releases. If you're not sure, stop reading threads like this and focus on making better sounds.
|
|
|
02-16-2013, 05:01 PM
|
#2265
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Sample rates..?
Thanks Yep, I take what you have to say very seriously.
I have to say I'm with Fran on this and I didn't see anything compelling in your post to change my mind.
Has there been any professional tests using several qualified engineers/persons, in a very good listening environment, under good controlled conditions, to quantify this in any way?
Maybe the benefits or non-benefits can be explained technically in such a way that these tests would be irrelevant?
|
|
|
02-16-2013, 06:15 PM
|
#2266
|
Human being with feelings
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
|
I switched to 96kHz after doing a simple little test with my soundcard.
Set the sample rate to 44kHz, plug my guitar into the intrument input and enabled monitoring in the soundcard app. Have a little play.
Then change the sample rate to 96kHz and do the same. The diiference is stark, i couldnt believe it. Its like a blanket gets lifted of the speakers.
Like Yep suggested, im led to believe its simply because the filters in my soundcard are working better at 96kHz not because there is 'more high end'.
Its worth trying and deciding for yourself, IMO. Its potentially a free way to improve the sound quality of your recordings.
__________________
Fortune favours the prepared...
|
|
|
02-16-2013, 06:29 PM
|
#2267
|
Human being with feelings
Join Date: Aug 2006
Posts: 2,019
|
Quote:
Originally Posted by Tod
Thanks Yep, I take what you have to say very seriously.
I have to say I'm with Fran on this and I didn't see anything compelling in your post to change my mind.
Has there been any professional tests using several qualified engineers/persons, in a very good listening environment, under good controlled conditions, to quantify this in any way?
Maybe the benefits or non-benefits can be explained technically in such a way that these tests would be irrelevant?
|
A chronic problem is that people are answering questions based on tests that did not ask the question under debate, and then demanding impossible proofs.
- If the question is whether any scientifically-controlled real-world sample of human beings has even heard a statistically-significant difference between a high-quality musical signal recorded at 44.1 kHz versus 88.2 all else being equal, my guess is no. My personal opinion is that nobody can hear the difference between an otherwise-equivalent 44.1kHz versus 88.2kHz recording.
- But if the question is whether theoretically audible digital errors or distortions could be introduced by applying digital processing to a 44.1k signal, that would be mitigated or rendered inaudible by recording at, say, 88.2kHz, then the answer is emphatically yes: that is not only possible, but easily-measured. How obvious the distortions are is a separate question, but if you paid me to do so, I could pretty easily create a signal-path where measurable inter-modulation distortions occurred on a 44.1kHz track, but not on an 88.2k recording of the same signal.
Would those differences "matter" or even appear to a real listener in a realistic recording and playback scenario? Probably not (almost certainly not, unless I was purposely trying to wreck the recording).
People are confusing "theoretically possible" with "important and meaningful". This whole conversation is off in la-la land.
"Audio engineer" used to mean an expert in the transduction of air-pressure into electro-mechanical signal. Then it meant people who knew how to operate transduction equipment to achieve the best sound. Now it means a bunch of crazy people arguing stuff they don't understand on internet forums.
It's like the blind leading the stupid, when it comes to things like inter-modulation distortion. People who can't even properly tune a guitar or side-mic a snare drum want to argue over esoteric and half-understood technical woowoo that has nothing to do with anything.
The argument you are asking me to referee/pick a side on, is a stupid one. It's like arguing over whether the ocean was bluish-green or greenish-blue. One side wants validation that 44.1k sample-rate is "fine" (or whatever), and the other wants to prove that 96k is "better", or something like that. It is an idiotic argument, and neither side is "right".
The best analogy I can think of right now is drinking water. If you distill all your water to the point of zero impurities, it will probably not taste as good, nor be as nutritious as mineral-water. It will definitely be better-tasting and better for you than drinking seawater or raw-sewage, but that's not what we are comparing it to.
We're off in la-la land, arguing over hypotheticals and theoreticals, and justifying them with stipulations and conditions that are meant to win arguments, not to produce good sound.
Take the debate offline, start a new thread. Reasonable sample-rates and bit-depth do not make modern recordings "sound like ass". You can't prove that your recordings sound just as good as someone else's through clever argument. The proof is in the pudding.
I already said that things like sample-rate advantages are mostly theoretical-to-nonexistent. Now you want to me to prove that there is room for error or doubt in that conclusion. I won't, and can't, and you can't prove the contrary. You can't prove a negative.
Take this somewhere else. There is a place for deep and technical discussion of audio-transduction theory, but "why do your recordings sound like ass" is not that place.
If anyone wants to make a serious case that recording at the wrong sample-rate makes one's "recordings sound like ass", this is a good place to do it. But if you just want to argue that this or that doesn't matter, please take it offline.
|
|
|
02-16-2013, 06:52 PM
|
#2268
|
Human being with feelings
Join Date: Jan 2007
Location: Silicon Gulch
Posts: 544
|
.
Quote:
- Before your audio is converted to digital, it goes through an analog filter that "cuts out" everything above the sample-rate, to avoid digital noise/distortion from supersonic harmonics and artifacts (basically an extremely steep bandwidth-limiting EQ). The quality of this filter is part and parcel of the overall ADC quality, and some are better than others. Higher sample rates can push that filtering up into extreme supersonic frequency ranges, and theoretically can improve sound-quality at the high-end of the audible spectrum.
|
For about 15 years now, nearly all audio Analog to Digital Converters have been using Sigma/Delta conversion. This technology replaces the base-rate multi-bit converter with a lower bit depth converter running at a much higher sampling rate. This achieves several important advantages all at the same time:
1) The higher sampling rate moves the analog anti-aliasing converter to a MUCH higher frequency where there are essentially no filter phase effects in the passband. The old analysis done by Lavry (which recommends 60KHz or so sample rate, based on in-band anti-alias filter effects) has been obsolete for some years now.
2) The included digital "decimation" filter combined with the high sample rate but low bit depth (and easily monotonic) converter guarantees monotonicity. This is an important issue in high bit depth converters. In the purely analog domain it is very hard or nearly impossible to achieve 24bit resolution due to even very small errors. I do not know of a single currently available 24 bit audio converter that does not use Sigma-Delta conversion. FWIR - The well-respected "Crystal" audio converters are all Sigma-Delta. My somewhat old interface using Crystal converters samples at a rate of about 380Khz and "decimates" down to 44.1. Newer ones probably sample a bit higher.
3) The sampling noise energy is also moved upward well beyond the base-rate and so much more of it is removed by the now out-of-band anti-alias filter giving a much lower noise floor than an equivalent base-rate converter could achieve.
It is quite silly that folks still bring up the old Lavry arguments that no longer apply to modern converters. We might as well argue about vacuum tube ADC design .
.
__________________
Inundated by a Perfect Storm of Gluten-Free Artisanal Bespoke Quinoa Avocado-Toast Toilet Paper.
Mahope Kakou (Later Dudes)...
Last edited by Kihoalu; 02-16-2013 at 06:59 PM.
|
|
|
02-16-2013, 07:21 PM
|
#2269
|
Human being with feelings
Join Date: Aug 2006
Posts: 2,019
|
Quote:
Originally Posted by Kihoalu
...It is quite silly that folks still bring up the old Lavry arguments that no longer apply to modern converters...
|
People bring up a lot of silly things. Some very famous and expensive studios are using pretty old technology, and some very famous expert names are still making statements based on their experiences from 15 years ago. Moreover, a lot of gear is still in use that was made decades ago.
This is where these hyper-technical discussions tend to run off the rails. Are we talking about engineering technicals, or about user-experience and sound-quality? Because they are two completely different conversations that tend to turn into crazy-talk woo-woo when they merge. Smart people have a way of turning stupid, when it comes to debates about audio quality.
|
|
|
02-16-2013, 08:38 PM
|
#2270
|
Human being with feelings
Join Date: Jun 2010
Posts: 415
|
Sometimes my recordings of acoustic guitar sound too much like a ukulele: not much low end, not much body.
I remember that I thought it sounded good when I played it back after the recording, so I think it probably sounded close to the actual sound. But when I listen now, I feel like I've missed something that was really there.
Assuming I actually am losing the original sound, and it's not just the guitar or performance, what can I do to improve this?
My main concern is that I think it sounds great at first. But then listening to the raw recording later, it sounds weak.
|
|
|
02-17-2013, 10:59 AM
|
#2271
|
Human being with feelings
Join Date: May 2008
Posts: 678
|
In "defence" of Lavry (yeah, like he really need that from me ) He himself has always mentioned those sigma delta oversampling "thingies" and I thought his reason for 60kHz was simply the theoretical gain of those frequencies that some folks might hear. He has been writing about this stuff like few years ago (I've read that stuff in gearslutz and REP forums, among other places). So his saying aren't outdated at all, and I do think he is one of the guys we should listen.
I just thought that this had to be said. His stuff is relevant. Misquoting however, never ends. Vague posts like this probably won't help either
Still, if same card sounds very different on different sample rates, it isn't because of physics but rather how they made that soundcard in the first place. If they want high rates sound better they can make lower rates sound worse, so to speak. That's pretty common philosophy nowadays, when it comes to technology. You know, product lines churning out same stuff and then lower-end models get crippled by driver software etc. With stuff like cameras and such this is so common... ok enough of this aimless ranting, carry on
|
|
|
02-17-2013, 11:19 AM
|
#2272
|
Human being with feelings
Join Date: Jan 2011
Location: Southern California
Posts: 642
|
Quote:
Originally Posted by jacobestes
Sometimes my recordings of acoustic guitar sound too much like a ukulele: not much low end, not much body.
I remember that I thought it sounded good when I played it back after the recording, so I think it probably sounded close to the actual sound. But when I listen now, I feel like I've missed something that was really there.
Assuming I actually am losing the original sound, and it's not just the guitar or performance, what can I do to improve this?
My main concern is that I think it sounds great at first. But then listening to the raw recording later, it sounds weak.
|
Record it better!
If you like how your guitar sounds from where you hear it, put the mic there (next to your ear). You don't hear the guitar as you play it from two inches in front of the sound hole. you hear it from a couple of feet above and a few inches behind.
Play with mic positioning and where you are in the room until you get the sound as close to the finished sound you want to hear in the finished recording. take the time to get the recording right, Don't aim to "Fix in the mix" you'll end up with sounds that require a ton of processing and so sound over processed
It's like cooking, you can add seasoning to the meal after it's cooked, but you can't unburn a steak with seasoning.
|
|
|
02-18-2013, 01:34 PM
|
#2273
|
Human being with feelings
Join Date: Dec 2007
Location: Walnut Creek, CA
Posts: 805
|
Quote:
Originally Posted by jacobestes
Sometimes my recordings of acoustic guitar sound too much like a ukulele: not much low end, not much body.
I remember that I thought it sounded good when I played it back after the recording, so I think it probably sounded close to the actual sound. But when I listen now, I feel like I've missed something that was really there.
Assuming I actually am losing the original sound, and it's not just the guitar or performance, what can I do to improve this?
My main concern is that I think it sounds great at first. But then listening to the raw recording later, it sounds weak.
|
You're suffering under the impression that our hearing is objective/mechanistic. It just isn't so. Just as the taste of wine changes by the taste of the food that precedes it, what we hear is impacted by _what we heard just before_. Another huge factor is the listening level. Emotional state is important too.
Expectation drives our experience as well - I play my guitar and it sounds fantastic, explosive, deep. I put it down and go make a cup of coffee. Return to the same seat, pick up same guitar, play same riff - now it sounds like cardboard. The only change was what I expected to hear - my expectation moved my judgement baseline.
Having played at home recording for ten years or so, I'm more than ever convinced that the most difficult part of recording, and the most critical, is the listening.
Fran
|
|
|
02-18-2013, 02:01 PM
|
#2274
|
Human being with feelings
Join Date: Jul 2010
Location: Online
Posts: 4,896
|
Quote:
Originally Posted by Billoon
I switched to 96kHz after doing a simple little test with my soundcard.
Set the sample rate to 44kHz, plug my guitar into the intrument input and enabled monitoring in the soundcard app. Have a little play.
Then change the sample rate to 96kHz and do the same. The diiference is stark, i couldnt believe it. Its like a blanket gets lifted of the speakers.
Like Yep suggested, im led to believe its simply because the filters in my soundcard are working better at 96kHz not because there is 'more high end'.
Its worth trying and deciding for yourself, IMO. Its potentially a free way to improve the sound quality of your recordings.
|
I just repeated your experiment here but with my voice not guitar.
No difference here that I could discern.Thank God.
__________________
it aint worth a bop,if it dont got that pop
|
|
|
02-18-2013, 03:53 PM
|
#2275
|
Human being with feelings
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
|
Quote:
Originally Posted by Cosmic
I just repeated your experiment here but with my voice not guitar.
No difference here that I could discern.Thank God.
|
Yeah, im sure it wouldnt happen with better quality soundcards.
__________________
Fortune favours the prepared...
|
|
|
02-18-2013, 05:47 PM
|
#2276
|
Human being with feelings
Join Date: Mar 2007
Posts: 21,551
|
Haha. I can't beleive that Yep actually had to defend one of "The Funk Brothers"? Lol
Now that was fun to read.
|
|
|
02-18-2013, 06:16 PM
|
#2277
|
Human being with feelings
Join Date: Mar 2007
Posts: 21,551
|
As to sample rate, i think a recent article floating around has some merit in that regard, and points to how some of that can result in unnecessary debates.
His general premise was that with a maybe not great overall converter design you actually may hear a difference, with your particular sound card, going to a higher rate. Where the personal logic falls apart is people taking that as a literal truth for all converters, extrapolating that experience into a meaning it maybe doesn't have.
It's the "what's true in this case must be true in all cases" fallacy. "96k recording sounds better here (comparing one thing only to itself) so 96k sounds better everywhere."
You may do the same comparison with a Mytek or Lavry and actually hear no immediate difference between 44.1 and 88.2 or 48k and 96k. The only way to know for sure is to try it on one of them.
And as relates to sound in general, if you are of the opinion that clocking helps, those high end converters also typically have much better clocks I think... assuming it's the master clock ... and assuming you buy into the general idea that good clocking makes a difference, another area of debate.
Last edited by Lawrence; 02-18-2013 at 06:24 PM.
|
|
|
02-19-2013, 12:41 PM
|
#2278
|
Human being with feelings
Join Date: Jan 2013
Location: Finland
Posts: 19
|
If you wonder is the difference between two sample rates or what ever audible, you can try program like this.
https://www.youtube.com/watch?v=jt7GyFW4hOI
|
|
|
02-19-2013, 09:50 PM
|
#2279
|
Human being with feelings
Join Date: Jan 2006
Posts: 2,173
|
Quote:
Originally Posted by Fran Guidry
So you'll supply clips of the same source recorded at the same time that demonstrate this difference
|
No, you can do this yourself, just like a poster in this thread did...
There is a difference, even after reading all the great responses here...try it out....and I have also been recording at home for over 10 years...could blow a lot of smoke & mirrors to back up my "level of experience", but that means nothing in a internet world where credits seem to rule...
Meanwhile, I will get back to recording my acoustics at 88 & electrics at 44.1 while the world keeps turning....
__________________
Yep's First 3 Years in PDF's
HP Z600 w/3GHz 12 Core, 48GB Memory, nVidia Quadro 5800, 240GB SSD OS drive, 3 480GB SSD Sample/Storage drives, 18TB External Storage, Dual 27" Monitors
|
|
|
02-26-2013, 08:56 AM
|
#2280
|
Human being with feelings
Join Date: May 2012
Posts: 13
|
Yep, since you are around, I'm charging again
Since I owe you mostly everything I've learned, I believe it's at least fair enough that I share the result of your teachments:
https://soundcloud.com/daysofjuly
I hope the comunity enjoys. After all, this IS home recording: $120 worth Mackie interface, older-than-me beaten-up Squier strat, a fine enough bass and drum pluggin. All of this endorsed by WDYRSLA.
PS: oh, i forgot: $80 or so AKG perception 100 mic.
|
|
|
Thread Tools |
|
Display Modes |
Linear Mode
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -7. The time now is 09:43 PM.
|