Discussion in 'Recording/Live Sound' started by jjboogie, Feb 27, 2009.
Not my inference at all and sorry if it came off that way. :BEER
There are many reasons why someone might prefer something. One pretty high end Grammy winning mix guy I know likes playing certain types of drum tracks back off his old PT Mix system at 44.1 precisely because he likes how it sounds. Even though I've grown to hate the sound of a PT Mix system I love this guys mixes and respect his decision on that because it's based on making his tools work for him.
I also agree with you that some hugely successful people, not only engineers but in all genre's of life can pass on info that makes you realize they have no idea what they're talking about in spite of their success. I guess the stars just line up for some people or something. My only concern regarding that is that I don't become one of them. I can assure you that the people I had the pleasure of testing this stuff with weren't of this ilk.
All in all I'm all for whatever makes people happy with the music they write, produce or master. I found working at higher sampling rates on certain types of projects to make a huge improvement in the music I've been producing. Whether that translates down to the person hearing my music on some cheap mp3 player or TV speaker isn't of much consequence to me. This isn't scientific, but I can say with 100% certainty that anything that can be used to improve any aspect of the production process from start to end will at the very least produce an intangable positive effect on the end result. And usually tangible in my experience. Can I prove this - no, but I know with every fiber of my being that it is so.
So I urge people to put down the magazines and journals and just go try something. Record a few projects at higher sample rates and then step back and see if it made any kind of improvement to your end product or enhanced the process of getting there. If it doesn't then you know and if it does then you've found a new way to enhance your creative output. Either way you've not lost a thing in the process.
Or don't.... it's not going to change my perspective on things one way or the other. I'm sure the Beatles would have been just as creative and accomplished the same things musically if they had allowed themselves to be reigned in by all those engineers in the white lab coats telling them about all those theoretical limitations.
Well put. :AOK
I've read the mix article and several others over the years and it seems like everyone can create or cite a test that supports either side.
Personally I know from listening that higher rates and bit depths sound better (closer to the sound of the source). All else being equal why would lower sample rates be better?
In overly simple terms, bit depth represents level choices for each sample. 16 bit audio has 2 to the 16th power (65,536) "amplitude" choices for each sample. 24 bit audio has 2 to the 24th power (16,777,216) choices for each sample. Since every individual sample is a digital representation of an analog source having more amplitude levels to choose from will always allow you to get closer to the source (except in cases where the source actually falls perfectly on one of 16 bits 65,536 break points). But even when they do, there is less interpolation required to get to the next sample when the bit depth is higher so the line graph created by those 2 time slices is a closer approximation of the analog source.
The same logic applies to sample rates in a second of 44.1 audio there are 44,100 tiny slices of sound and at 96k there are 96,000. The digital converters are required to do less interpolation when they have more samples per second. That is math.
Personally, I disagree with the arguments about the quality of the source. The entire world understands G.I=G.O. but when we are discussing recording, we are talking about accurate representations of the "garbage" not the quality of the source.
I also know that all of this is secondary to the quality of the engineer. (which is different then the source) in that a great engineer will get always get better sounding tracks out of 16/44.1 than a crap one will get out of 24/96.
It's all minutiae though. I think most of us got into this to make music, so make it. The more you do, the better you get. The math side of my head wants to figure this stuff out for fun, but the music side really just wants to share my love for music with anyone and everyone who will listen.
Your explanation was neither simple nor relevant. This isn't an issue of amplitude (bit depth) - we know that 24 bits provide enough "slices" of amplitude - it's an issue of the rate at which we slice along the horizontal line, the sample rate. There is a point at which the magnifying glass is focusing on details we CAN'T notice under normal playback.
The average listener can only discern between (about) 3db of volume increase or decrease. That's bit depth. Current digital rates bit depths are adequately "deep." The average ear has almost infinitely more potential to hear a wide spectrum of different frequencies. That's sample rate.
I don't think anyone said "lower rates were better", but since you brought it up......
It has been demonstrated that recording material outside of the human hearing range (above 20k) can cause beating with other tones within the audible range, therefore creating inaccurate representations of what we hear. The non-linearities of analog gear can be guilty of this as well--creating distortion within the audible range while attempting to reproduce the high frequency waveforms. This comes up when the Neve stuff does.
I'm just sayin'......
Which was my standpoint in the argument I was making.
I will not dispute that 96 khz is better overall for sound quality. You would have to be a fool to think otherwise. The whole argument is based on the assumption of content delivery on either CD or DVD mediums.
I have found that when I convert something from 96 khz masters to 48 khz (dvd pcm audio) the final product did not sound like I wanted it to. The 96 khz master apparently had harmonic distortion down in the spectrum that I could hear. Therefore, coloring the audio and making me mix the mastering EQ in a manner that was a 'false positive'.
If I mix from the beginning at 48 khz, I know exactly how everything sounds all the way up to delivery.
If I compare a 96 khz audio sample to a 48 khz sample I can definitely hear that the 96 khz is better. But unless you produce for 24/96, doing this may actually cause your mix to sound different than you intended. Even if you down convert using BarbaraBatch or Voxengo's R8Brain.
Well, it's simple to me! As far as relevance is concerned, the only thing that makes samples different than the original source is the interpolation between time slices and between bits (assuming perfect recording equipment). The less interpolation the better (more like the source). That is why bit depth is relevant. It's a staircase, if the treads are longer or the risers are higher, then the points between the crossings are interpolated to a greater degree and therefore less like the original.
As far as the notion of what average listener can discern goes, I completely disagree and have personal experience to back it up. The average listener may not be able to describe or define the differences they hear, but if we all start mixing in 3db steps that would be just fine according to your argument.
Mixing is not simply about what a non-trained ear can describe. Go to a track you've mixed and raise the guitar 1.5db and ask your wife to compare the 2 and I bet she'll know that they are different. And that's 1/2 of what you are claiming an "untrained ear" could discern.
Finally, I am not a non-trained ear, I am a trained ear. So I choose to make my work the best I can do for myself (and I know there are many better than me). But the idea that I should be mixing for people who won't know the difference between 3db is silly.
Funny thing is I avoided this thread for a while because they always turns into some kind of p---ing match. That's not my intent, I am just sharing my opinion on the subject which is based on my experience. YMMV.
Which is why I quantified my use of the word better by adding, "more like the source". And as I also said and am prepared to back up, every side of this argument has been "demonstrated" ad nauseam. I just base my opinion on my observation. Flawed as it may or may not be, it is still my best description of my observations made over my career in the recording arts.
Everyone can take these tools and apply them as they see fit and if it comes out good, then it is good.
Yes, this topic has gone WAY past simple matter of opinion, but this subject always does. Each time. People take it very personal. It's a bit odd.
Your post was not relevant because the discussion concerns how necessary it is to use high "sample rates." You're making references to bit depth, but the entire world and their mothers are using 24 bit. It's not being debated because it's standardized. The "steps" (in amplitude) you refer to require no interpolation - no middle ground guesswork which has no digital representation (it's a strange choice of words for bit depth) - because 24 bits are all that's required to record amplitude. You said yourself that there are 16 million "steps" when given 24 bits. Even 16 bit (standard CD) completely covers ALL the different steps. It's merely a matter of further suppressing the noise floor at this point, and providing a smooth tail for long reverbs, but 24 bit gives us sufficient room for both.
The reason the word "simple" in your post was funny to me is because you'd inserted a related, yet un-debated aspect into this, um... debate, with an attached explanation which was anything BUT simply stated. Your point is correct - we DO need enough "steps" in our bit depth - but we got 'em in spades. 15 million of them. If we had a recording of the Big Bang (and I swear NASA has this), 24 bits could cover it PERFECTLY! Now I'm not saying 48KHz would... If it were MY job to record this, I'd point a Schoeps towards the center of the universe and you just might see my converters set to 192KHz. Of course, I'd have another set to 96, just in case
There are lots of different opinions out there (and in here) but there is definitely no second take for the big bang. I like the Schoeps for that, but maybe a 57 too, off axis.
You certainly wouldn't have to worry about reflections.
And you'd find plenty of people to argue about mic choice. I imagine I'd even stumble into that train wreck...
a real man would contact mic it.
And that man would be Adam.
(Or Atom, for you science geeks)
So to the OP...
If I had typed up my own (unprofessional) opinion on the subject, this would be it.
Actually someone did ask for a clarification on the differences between sample and bit rate....makes sense to me
Would you mind expanding on this a bit, maybe explain what "cause beating with other tones" means? Does this mean that recording at 96 or 88.2 will actually sound worse when played back at 44.1? Different?
Also, I'm thinking it's only a matter of time before the CD standard rate goes up, or at least it wouldn't surprise me at all. What then, if you've recorded everything at 44.1?
Also, if humans can only hear up to 20khz, why do people rave so much about something like the 28k switch on the Avedis MA5 preamp? Most people claim it adds "air" to the sound which indicates that 28k can be perceived, maybe subconsciously rather than heard?
I have a feeling that that switch does more than add 28k, it probably slops the high end a bit. That and it could suffer from the placebo effect, you hear something because you want to hear it (or because you payed alot of money).