Really dispels alot of bullshit you hear about digital from analogue purists
Last edited by a moderator:
Really good video on digital audio
- Thread starter Trevoire520
- Start date
And here are some differenting opinnions considering the video. Just because someone made a video of it, doesn't mean that they are right.
http://thewombforums.com/showthread.php?t=21362
http://thewombforums.com/showthread.php?t=20609
E: The one thing that really rises my eyebrows in that video is the notion of tape hiss and bit noise floors being identical and analog tapes having 'less bits' than digital.. To my knowledge that's really inacurate and sounds more like a weak attempt to make a jab towards analog gear. There is more to analog signal contra digital than the noise floor.
Usually my gripe with these types of things is the scientific and mathematical approach. Math is all fine and dandy. But we don't live in a mathematical world as I have said before. If every digital hardware manufacturer, plugin coder, algorithm designerm and whatnot had done their homework and used the finest parts with no expenses in mind, we wouldn't have these discussions. But we don't live in that kind of world. We live in reality, where the inplementations of math have severe handicaps when compared to the mathematical universe.
E II: Usually the guys bashing the 'analog purists' are consumer grade digital users who have never even stood in the vicinity of quality analog playback. Not to mention quality digital playback. And usually their only experience with 'analog sound' is their analog modelled plugin collection.
http://thewombforums.com/showthread.php?t=21362
http://thewombforums.com/showthread.php?t=20609
E: The one thing that really rises my eyebrows in that video is the notion of tape hiss and bit noise floors being identical and analog tapes having 'less bits' than digital.. To my knowledge that's really inacurate and sounds more like a weak attempt to make a jab towards analog gear. There is more to analog signal contra digital than the noise floor.
Usually my gripe with these types of things is the scientific and mathematical approach. Math is all fine and dandy. But we don't live in a mathematical world as I have said before. If every digital hardware manufacturer, plugin coder, algorithm designerm and whatnot had done their homework and used the finest parts with no expenses in mind, we wouldn't have these discussions. But we don't live in that kind of world. We live in reality, where the inplementations of math have severe handicaps when compared to the mathematical universe.
E II: Usually the guys bashing the 'analog purists' are consumer grade digital users who have never even stood in the vicinity of quality analog playback. Not to mention quality digital playback. And usually their only experience with 'analog sound' is their analog modelled plugin collection.
It's a simplification, sure, but I think the only thing he was getting at was dynamic range. It probably could have been worded more concisely, but when teaching the basics, you usually have to compromise between technical accuracy and confusing your audience with jargon. I'll have to watch the video again.
My understanding of the SNR of tape is that while it might not be an absolute noise floor, but since you always want to minimize noise, it effectively serves the same purpose. A half-inch two track might have an SNR of 65 dB and have a bandwidth extending to 20 kHz, as soon as you go thinner or try to shove more tracks on the tape, you get more noise and start rolling off some of the top end.
So while it isn't entirely analogous (heh) with digital, a lot of the general concepts apply. Is this an accurate/complete understanding? I got my start with audio programming languages, so my grasp on electronics or analog recording isn't always as strong as it should be.
That's something that's always been true though, isn't it? There are bad tape machines and there are bad converters. At the end of the day, good gear is good gear. They sound a little bit different, but neither of them sound bad, so it ultimately comes down to personal taste (as well as a ample helping of economics).
Genius Gone Insane
http://www.¯\(°_o)/¯.com
egan.
daylightdies.com
I think his point -whether he knows it or not-- is that alot of the science and math here is glossed over. Plenty of mediocre A/D's have filters which cause aliasing within the audible range. Those kind of errors are only going to appear when the content extends beyond the nyquist frequency. Also, it may be audible long before it's visible on a zoomed out waveform on a scope. Go look at that massive interface benchtest thread on gearslutz for real life examples.
I do think the 192 camp is somewhat missing the point (and shooting for overkill) but the truth is that some converters (generally cheaper ones) do sound better (and measurable perform better) running at 88.2 or 96k.
I do think the 192 camp is somewhat missing the point (and shooting for overkill) but the truth is that some converters (generally cheaper ones) do sound better (and measurable perform better) running at 88.2 or 96k.
Sloan
Sounds like shit!
TheWinterSnow
Den Mørke Natt
For starters he never said tape his is the same thing as bit noise floor. However there is some truth to that. His approach was that noise and its overall signal strength directly effect dynamic range, in the audio world that would be tape hiss, vinyl noise etc. In digital the determining factor of dynamic range to noise is the result of the bit depth.
While they are not the same kind of noise...actually with dithering it is all white noise, but at the end of the day you have noise that limits the dynamic range, that was all what was trying to be said.
In the first video he talks about this and I thought it was common knowledge but, the quality of the low pass filter in the converter directly effects aliasing. Everything that is above the Nyquist Frequency is folded back into the audible range, the sharper the cutoff, the lower the sampling frequency can be. Not only that but with a sharp filter, even when you have aliasing, it affects a smaller portion of the signal and with lesser magnitude. Higher sampling rates allow for more shallow rolloff slopes but have more aliasing. If you increase the sample frequency, the new Nyquist Frequency allows for the aliasing caused by the lower rolloff slope to be moved above the audible range.
What that means is that if you have a very expensive A/D D/A setup seen in most mid to high end studios, you can track at 44.1KHz/16-bit and there would be no difference with higher sampling rates/bit rates. However for more consumer level products that don't have the best filtering, yeah they may have some noise going on at lower sampling rates. Its worth noting though that since DSP power has become more robust that high end sharp and accurate filters can be made in software and save a lot of money, this means that newer consumer grade digital audio products don't need higher sampling frequencies as they already have as little aliasing (if none) as much higher end equipment would.
From my experiments recently working on my new DI I came across something more concerning. I have a cheap $200 USD interface from which originally came out in 2003 I believe, a sine wave coming out looks clean, no stairstep, no aliasing. What is really concerning more than sample rates is tracking levels including the signal strength you send out to your monitors and outboard gear. The amount of distortion near and at 0dBFS in consumer grade products will make you cry. The fact is my interface as outdated as it is still produces perfectly clean signals with no digitization just like the video, its your tracking levels that really make a massive difference.
I will write/post of blog with pictures and my findings to augment the discussion. I was going to write a blog about it anyway.
While they are not the same kind of noise...actually with dithering it is all white noise, but at the end of the day you have noise that limits the dynamic range, that was all what was trying to be said.
In the first video he talks about this and I thought it was common knowledge but, the quality of the low pass filter in the converter directly effects aliasing. Everything that is above the Nyquist Frequency is folded back into the audible range, the sharper the cutoff, the lower the sampling frequency can be. Not only that but with a sharp filter, even when you have aliasing, it affects a smaller portion of the signal and with lesser magnitude. Higher sampling rates allow for more shallow rolloff slopes but have more aliasing. If you increase the sample frequency, the new Nyquist Frequency allows for the aliasing caused by the lower rolloff slope to be moved above the audible range.
What that means is that if you have a very expensive A/D D/A setup seen in most mid to high end studios, you can track at 44.1KHz/16-bit and there would be no difference with higher sampling rates/bit rates. However for more consumer level products that don't have the best filtering, yeah they may have some noise going on at lower sampling rates. Its worth noting though that since DSP power has become more robust that high end sharp and accurate filters can be made in software and save a lot of money, this means that newer consumer grade digital audio products don't need higher sampling frequencies as they already have as little aliasing (if none) as much higher end equipment would.
From my experiments recently working on my new DI I came across something more concerning. I have a cheap $200 USD interface from which originally came out in 2003 I believe, a sine wave coming out looks clean, no stairstep, no aliasing. What is really concerning more than sample rates is tracking levels including the signal strength you send out to your monitors and outboard gear. The amount of distortion near and at 0dBFS in consumer grade products will make you cry. The fact is my interface as outdated as it is still produces perfectly clean signals with no digitization just like the video, its your tracking levels that really make a massive difference.
I will write/post of blog with pictures and my findings to augment the discussion. I was going to write a blog about it anyway.
JeffTD
Senhor Testiculo
He would probably agree and say that he enjoys math because it doesn't have to operate within the bounds of real-world applications.
Most people seem to forget that interfaces and AD/DA's are also analog devices and the circuitry, crappy psu's and so on do make just as much difference as the digital filters and algorithms..
Keeping nominal levels in and out is good gain stageing, but also important because if your gear isn't top notch from top to bottom (like most of our isn't), high levels in, out or both will fuck up your sound cause they make the cheap ass psu's and their cheap ass friends cry like a little baby.
drew_drummer
Dancefap
Maybe stick to arguing points instead of ad hominem?? Doesn't do your argument much service when you resort to character assassination on a broad-brush basis like this.
Not trying to prove or argue anything with with that comment.. Just mentioning my personal observation / rant.
E: Wasn't pointing that towards Trevor.
E: How many engineers you know / have heard of that after years and years of extensive experience with quality analog mediums and gear, favour digital based on sound quality rather than the other good attributes of digital mediums? Or budget restrictions? How many 'scientists' and experts have you witnessed (who by their own merits know better than veteran engineers) that want to dispell the so-called-illusion that analog is superior..? How many, that are in both categories?
But all in all, this is a moot discussion because in the end, all that matters are the sonics.. They go hand in hand with audio quality, but they are not the same thing. And that, is not a scientific argue.. It comes down to taste.
The reason these things keep coming up is because everyone wants it to be true. And I do mean everybody. Everybody wants and hopes that consumer grade digital could compete with quality analog gear. Sadly in many cases that just simply isn't true.. Atleast for now. But I hope and believe that we are slowly getting there.
According to some experts, we got there over 20 years ago. According to just about every professional who has ever worked with audio, we aren't there yet.
TheWinterSnow
Den Mørke Natt
huh wut? The above is so not true it isn't even funny.
Ask yourself how many modern studios record on Analog Mediums? There is a reason they don't anymore. Analog consoles and outboard gear is on thing, but when it comes down to it, digital has been winning as the medium for recording audio, dude to its noise floor and ease of use.
As to which is better is completely subjective to the individual.
It is also not a moot discussion, the thing is many newbies are still asking these questions due to myths that still float around and will tend to blame cheap gear for their shit sound. The video was dispelling that if you mixes horribly suck because you have a cheap interface, it isn't the interface, its the user. Hell you still see experienced engineers even hear asking about dithering and bit rate, it by any means is never moot to ask questions, or to dispel some myths due to people asking questions.
The fact is, consumer grade digital audio equipment is about 95% there when it comes to quality. The biggest portion of issues pertaining to converters is the analog portions of the circuits inflicting unwanted color, the biggest one being phase distortion.
Also keep in mind when it comes to noise and aliasing not all DAWs are created equal. Pro Tools being one of the worst, still explains why expensive converters can still sound better at higher sampling rates.
There are so many other aspects to noise, and sound quality other than what was mentioned in the video, way more than I could ever go through in a single post. The video was only there to distinguish the myths of sampling frequencies, bit rate and if output signals look like stair steps. Those three things aren't the only thing that determine good audio quality converters.
egan.
daylightdies.com
Come on dude. That's a revisionist history. Classical recordists are concerned with maximum noise floor and DR but the reason rock studios left analog is cost of entry, cost of maintenance, cost of medium, learning curve and editing power.
Where are you proposing this new aliasing occurring? In the mix bus? I'd really like a citation for this statement.
TheWinterSnow
Den Mørke Natt
And a much greater noise floor, SNR and dynamic range wouldn't also play into effect the switch either. The only big plus of analog was the harmonic coloration, even if that was worth the fuss regarding cost and ease of use, would you willingly have higher noise levels for such coloration?
At the same time when you are using an analog console and outboard gear, how much of a "sonic improvement" would you get from the sound of a tape medium? Then again, would every engineer/musician prefer the sound of tape in a mix? Is that coloration absolutely needed for the best mix possible. I would imagine that for most rock/metal that coloration would be detrimental to the clarity of the mix.
Just my 0.02
Engine accuracy/precision. Not all engines have accurate oversampling from processing errors. Some years ago around here there was a page posted that was on gearslutz showing aliasing for each day, I don't have it unfortunately.
egan.
daylightdies.com
Actually, the other big plus for rock is that tape by it's nature is the fastest and smoothest limiter there is. Whether you mind the noise floor depends on your goals but if you're slamming to tape it's not an issue.
Whether it's for you is a matter of personal opinion but I'm sure other guys that were around in the 90's and 00's will tell you that the push move towards totally digital (as opposed to hybrid) was an economic choice. I still regularly run into folks who won't book the commercial space I work in b/c we don't have a 2". That sound can't be faked (yet). YMMV.
I'm by no means an electronics engineer, just spouting what I've gathered from here and there on the subject. Please do enlighten me. You are saying that the the low cost parts in the circuitry and the crappy psu won't affect the sonics of any given gear? And that the pushing the given gear on the edge won't tax more from the psu? And the quality of the psu (or other parts) doesn't matter how the aforementioned gear handles the signals fed in and out of the circuitry?
Similar threads
