There are, in fairness, good engineering reasons for using higher rated converters at the recording stage. Simply put, it's a lot easier to build a good, non-artefacting gently sloped cutoff filter than a steep one. If your converter can work at a higher rate, it's entirely possible that you can get better recordings at the initial conversion stage.
 
Some points to note are:
- after that initial step, there's no value in the higher rate.
- manufacturers don't tend to publish very clear or reliable information about how their converters are operating
- a 44.1 converter may well be already dealing with this via an initial oversampling step
- there's a definite law of diminishing returns here. Once the filter isn't causing artefacts, it isn't causing artefacts. You rapidly reach a point where increasing the sample rate gains nothing.
 
It is all a lot less like rocket science than it's been presented throughout this thread.

Over the years I've become convinced that the clock plays a pretty big role in the sound quality of a given DAC or ADC process. I've personally heard e.g. the old Digi 002 improve dramatically with a different clock. My friend who studied sound engineering told me they run a rack of Behringer converters at his school that they clock from an Apogee, also greatly improving the sound.
 
 Many soundcards that list similar specs sound very different, and what a given chips maximum sample rate is doesn't really seem reflect on what its sound quality is like. It seems akin to the photography thing where they pretend that using more megapixels gives you a better picture. I'm completely convinced you can get a good sound at 44.1K, 24 bit, as long as you have a decent converter and a good clock. Additionally, I don't think you can entirely "solve" a jittery clock problem by doubling or quadrupling the sample rate.

I managed to fix my settings as advised earlier in the thread. The maximum that my soundcard is specified for recording is 96 / 24, but I could set the bit rate up to 64 bit in the preferences, but havnt done any audio recording yet so don't know if that will work, I'm just focusing on getting the midi backing tracks done for my ideas atm.

John

Goddard

John

Goddard

John
"This is probably a good place to ask this question. I did my first-ever sampling last weekend, and ended up using Audition to do the task. I recorded them 44.1KHz mono (drums), but Audition defaulted to a 32-bit float on saves so I just used that. Is that 32-bit buying anything at all, or just wasting space on me?"
 
 
For recording yes its wasting space. Keep in mind that your converters are incapable of recording anything above 24 bits. So the file that was created is 24 bits plus a lot of padding. This adds nothing useful to the recording at all.  
 
Now for processing it a very different story. However, I am of the opinion that the file on disk doesn't need to be greater than 24 bits even after processing.   I am sure I am alone in this view.  

No, the 32-bit floating-point file data saved by Audition isn't padded, it's just being represented in floating-point format (and using all of those 32 bits for its data) rather than in the fixed-point/integer 24-bit format as was output by the ADC. If it were being saved as 32-bit integer data instead, then yes, it would include padding.
 
Perhaps Noel (if he's still around) would care to confirm whether/when Sonar converts to integer format for disk storage (as WAV files iirc).

Where is the audio that is filling those extra bits coming from? My understanding is that floating point will reduce rounding errors if the file has been processed and with a 32 bit FP audio engine used. Otherwise it will be exactly what the converter created. 

Your understanding is incorrect/incomplete:
 
http://en.wikipedia.org/wiki/Single-precision_floating-point_format
 
 

Sorry I'm not following you. If the recorded data/audio is only 24 bits how can converting it into 32 bits add further data? It will not become audio with any more detail or precision. It is limited to the 24 bits that the A/D converter produced. If it is processed it is possible that the resulting audio will have greater precision due to an internal 32 bit FP audio engine but it will be inaudible.
 
It really should be similar to recording at 16 bits and then converting the 16 bit file to 24 bits. You gain nothing except a patted file with the original 16 and a bunch of zeros filling out the rest of the file. Its like having a container that you fill up. No interpolation occurs simply because the file is bigger.  
 

A simplified analogy often given is to "scientifc notation" of decimal (base 10) numbers, where the decimal point position is represented in an exponent (power of 10).
 
In a 32-bit floating-point binary (i.e., single-precision) representation, the 1-bit "sign bit" plus 23-bit "signficand" (sometimes called the "mantissa" as in logarithms) portions together carry the basic 24-bit integer binary data (actually 25 bits' worth thanks to the "implied" leading "1" bit) while the 8-bit "exponent" portion carries the data indicating the position of the binary decimal point. The use of a floating point representation enables, by virtue of the exponent, a vastly wider range of possible values than can be represented by a 24-bit integer binary number, which is why a 32-bit foating-point audio/mix engine doesn't suffer risk of clipping (whenever the value would exceed the significand's maxium possible value, the exponent value can be increased).
 
If you stilll can't comprehend from the preceding and from that wiki page how all the data of a 24-bit integer PCM sample (plus 1 extra bit's worth!) is represented by all the bits of a 32-bit floating point binary number representation then I'd suggest searching for another online explanation of floating point binary math which is hopefully easier to grasp.

John T

Goddard

John T
Sampling and playback are not the same, and don't exhibit the same problems. That the mac can playback 192khz audio has nothing to do, for good or ill, with what he discusses in the article.

Uh, a Mac laptop's codec chip can record (sample) @24/192 too:
 
http://www.realtek.com.tw/products/productsView.aspx?Langid=1&PFid=28&Level=5&Conn=4&ProdID=138
 

I'm sure it can, but it's not relevant to the point the guy is making. He's simply saying that a bog standard consumer grade converter running at 44.1/16 is actually very good these days. If you want to get stuck on the fact that he's used the throwaway description "laptop mac converter", then go for your life. It's got nothing to do with the matter at hand.

Ah, so your earlier criticism of the guy for some historical inaccuracy about videophones was somehow relevant to some point the guy was making, but my pointing out that the guy's own "Mac laptop converter" example undercut all his hooey about 192kHz sampling being problematic isn't relevant and has nothing to do with the matter at hand?

Goddard

John T

Goddard

John T
Sampling and playback are not the same, and don't exhibit the same problems. That the mac can playback 192khz audio has nothing to do, for good or ill, with what he discusses in the article.

Uh, a Mac laptop's codec chip can record (sample) @24/192 too:
 
http://www.realtek.com.tw/products/productsView.aspx?Langid=1&PFid=28&Level=5&Conn=4&ProdID=138
 

I'm sure it can, but it's not relevant to the point the guy is making. He's simply saying that a bog standard consumer grade converter running at 44.1/16 is actually very good these days. If you want to get stuck on the fact that he's used the throwaway description "laptop mac converter", then go for your life. It's got nothing to do with the matter at hand.

Ah, so your earlier criticism of the guy for some historical inaccuracy about videophones was somehow relevant to some point the guy was making, but my pointing out that the guy's own "Mac laptop converter" example undercut all his hooey about 192kHz sampling being problematic isn't relevant and has nothing to do with the matter at hand?

It wasnt a criticism, it was a mainly humorous observation following up on bitflipper's post. I hope I'm being clear about this. I think the guy makes very good arguments. He seems to have gaps in his knowledge. But you know, so do we all. They matter when said gaps are fatal to the arguments we're making. In this case, hey are clearly not.

John T

Goddard
 
You've misunderstood. A converter is actually sampling its input at a far higher frequency (in the MHz "low radio frequency" band) than "ultrasonic".
 

This is flat out wrong. You're conflating the device's means of operation with its function. A converter is sampling its input at its functioning sample rate. There's all kinds of jiggery pokery going on with how it does that, some of which yer man touches on in his article.

No, it's not wrong, at least not as concerns Delta-Sigma converters employing oversampling, which for many years have been the prevalent type of ADCs and DACs used for audio (although thinking back, Lavry's and perhaps also some others' early "mastering" converters may have used conversion bits having a different archtecture (multi-pass flash?) way back when).
 
Your "conflating the device's means of operation with its function" statement is nonsensical and naiive.  Perhaps you should learn a bit about the "jiggery pokery" how of things before posting any further arguments. This might be informative:
 
http://skywired.net/blog/2011/07/how-delta-sigma-works-anti-aliasing-advantage/
 
And this might lend a bit more perspective than that facetious "Science of..." blog:
 
http://www.benchmarkmedia.com/discuss/feedback/newsletter/2010/08/1/unique-evils-digital-audio-and-how-defeat-them
 

John T

Goddard

John T
Well, I'm not actually arguing, or at least not at the moment. I'm just finding Goddard strangely elliptical and a bit unconvincing. He certainly doesn't seem interested in explaining or discussing anything, but he does seem awfully keen for us all to think he's really clever. I'm trying to give him the opportunity to impress us.

I really don't care whether you or anyone else is impressed, only that misinformation isn't perpetuated unchallenged, or taken in by the naiive and gullible who don't know any better, when those who should know better praise it.
 
You and others criticized that blogger for historical inaccuracy earlier, and I've merely done the same wrt obvious technical fallacies in that article.

In principle, this is a fair point but it's undermined by the fact that you've been profoundly incorrect in most of your claims. For example, your thing about MHz sampling. That was just nonsense.

 
MHz sampling (oversampling) is a reality in DSM audio ADCs and DACs, even if you are unable to comprehend that it occurs. You're confusing the actual sampling rate of a DSM converter (the frequency at which its input is sampled) with the converter's output sample rate.
 

John T
If you've a counterpoint to that, or anything else in this conversation, I'm honestly eager to hear it. If you're not willing or able to to counterpoint anything, though, well... I dunno what to do with that.

 
I've already responded as much as I care to. Done playing your time wasting games. Don't really care if you're persuaded. I'll leave it to others and posterity to assess any technical merit in what I've posted.

Very little of what you wrote has anything to do with bit depth, I would say.
However, music is an emotional thing too, so if you are feeling better using 32 bit fp, knock yourself out.

Sander...Just trying to understand the implied additional headroom/processing implications... If "I would say..." is based on a real understanding of what God was sayin' (couldn't resist!) then thats good to know although "very little" tells me "very little"... I for one don't comprehend the full breadth and credabilities of the arguments yet but my curiosity was piqued by Goddard's comment...
 
It doesn't make me feel better or worse... information/data does that for me... I admit to being stupid or ignorant about a zillion things... but try to understand everything...just haven't delved and practice suspending judgement/positions till I understand...and never get stuck there, as new info always unfolds...
 
One thing I think I learned this week is to always record stereo to retain track bin imaging opportunities...all my mono tracks are stuck there... thus, smart to record stereo just in case...BitFlipper's eloquent solilaque on mono/stereo really helped me decide what works best for me.... Seems the same for 32fp, lacking reasons not to as I can always reduce it if the issues are clarified...but never truly realize any advantages by going there after the fact. I was hoping for some real data on the above...it helps wade through the human tendencies of denialism/false beliefs on either side of many issues which stem from ignorance or lack of compelling/comprehensive data/understanding.
 
 

thank, John... I'll take a look... I'm a dummy in this "realm" at this point and want to understand...
My whole objective is to ideally keep multiple signal chains from overloading and if it is a specious argument, it's worth knowing...gotta lot of reading to do on both sides of the fense within this thread...

OK, Gurus...
I stepped away to pour over a number of cited references... and of all of them, this one seemed to ring true or at least was couched in a fashion that I can relate to... what I came away with is Sonar probably does this internally anyway, at least for effects (does it do it for busses and tracks?) and therefore the real world distinction between say, what I'm running (96/24, primarily for Latency with my Forte) and 96/32 is non-existant...except for much bigger files/disk&ram demands... 
 
http://www.sonicscoop.com...pth-is-probably-wrong/
excerpt:
"32 Bits and Beyond
Almost all native DAWs use what’s called “32-bit Floating Point” for audio processing. Some of them might even use 64 bits in certain places. But this has absolutely no effect on either the raw sound “quality” of the audio, or the dynamic range that you’re able to play back in the end.
What these super-high bit depths do, is allow for additional processing without the risk of clipping plugins and busses, and without adding super-low levels of noise that no one will ever hear.  This extra wiggle room lets you do insane amounts of processing and some truly ridiculous things with your levels and gain-staging without really thinking twice about it. (If that happens to be your kind of thing.)
To get the benefit of 32-bit processing, you don’t need to do anything.  Chances are that your DAW already does it, and that almost all of your plugins do too. (The same goes for “oversampling,” a similar technique in which an insanely high sample rate is used at the processing stage).
Some DAWs also allow the option of creating 32-bit float audio files. Once again, these give your files no added sound quality or dynamic range. All this does is take your 24-bit audio and rewrite it in a 32-bit language.
In theory, the benefit is that plugins and other processors don’t have to convert your audio back and forth between 24-bit and 32-bit, thereby eliminating any extremely low-level noise from extra dither or quantization errors that no one will ever hear.
To date, it’s not clear whether using 32-bit float audio files are of any real practical benefit when it comes to noise or processing power. The big tradeoff is that they do make all of your projects at least 50% larger. But if you have the space and bandwidth to spare, it probably can’t hurt things any.
Even if there were a slight noise advantage at the microscopic level, it would likely be smaller than the noise contribution of even one piece of super-quiet analog gear.
Still, if you have the disk space and do truly crazy amounts of processing, why not go for it? Maybe you can do some tests of your own. On the other hand, if you mix on an analog desk you stand to gain no advantage from these types of files. Not even a theoretical one..."
and...
"All Signal, No Noise
To give a proper explanation of the mechanics of just how the relationship between bit depth and noise floor works (and why the term “resolution” is both technically correct and so endlessly misleading for so many people) would be beyond the scope of this article. It requires equations, charts, and quite possibly, more intelligence than I can muster.
The short explanation is that when we sample a continuous real-world waveform with a non-infinite number of digital bits, we have to fudge that waveform slightly in one direction or another to have it land at the nearest possible bit-value. This waveform shifting is called a “quantization error,” and it happens every time we capture a signal. It may sound counter-intuitive, but this doesn’t actually distort the waveform. The difference is merely rendered as noise.
From there, we can “dither” the noise, reshaping it in a way that is even less noticeable. That gives us even more dynamic range. At 16 bits and above, this practically unnecessary. The noise floor is so low that you’d have to go far out of your way to try and hear it. Still, it’s wise to dither when working at 16 bits, just to be safe. There are no real major tradeoffs, and only a potential benefit to be had. And so, applying dither to a commercial 16-bit release remains the accepted wisdom.
Now You Know
If you’re anything like me, you didn’t know all of this stuff, even well into your professional career in audio. And that’s okay.
This is a relatively new and somewhat complex field, and there are a lot of people who can profit on misinforming you about basic digital audio concepts.
What I can tell you is that the 22-year olds coming out of my college courses in audio do know this stuff. And if you don’t, you’re at a disadvantage. So spread the word.
Thankfully, in a field as stimulating, competitive and ever-evolving as audio or music.
Keep on keeping up, and just as importantly, keep on making great records on whatever tools work for you – Science be damned."
 

I think I'll call it a night... and thanks to all for an interesting mental exercise... I like Justin's closing thought...
 "lifelong learning is half the point of getting involved..."
 

John
I think Goddard is confused about 1 bit audio recording and the rest of the audio recorders. 
 

God only knows if that's it. What's evident, though, is that he's got absolutely no idea what he's talking about. It's sort of fascinating.

OK before I let myself get caught up in putting some one down let it end here and now. Often threads like this can and often do result in conflict. That is not the way we should let things happen any more. Goddard was only posting what he thought was true. There is no sin in that. 
 
I like it when some one posts a correction when they believe it is needed. We need to let members feel this place will not jump on them just because of a disagreement. 
 
We need to keep this place free of intimidation or make people feel uncomfortable no matter what is posted.   

Nod as good as a wink...
 
I think I heard something more from my vocals using 48k a long time ago but it was very subjective and subtle.
 
I tried higher rates but could tell no more difference and lost half of my UAD resources.
 
48k runs smoother in my DAW because my korg synth is 48k native, so is my TC interface and my VL2.
 
Lot less stumbling and stalling in 48k for me.