Comment by stcredzero | Hacker News Reader

Comment by stcredzero | Hacker News Reader

Parent Full thread

I know people that work in IT and still pretend that audio transmission over USB can add jitter to music and that a DAC can "open" or "close" "the stage" by low quality processing of the signal.

Audio transmission over USB with cheap converters and poorly chosen levels can result in loss of sound quality. I'm pretty sure you can tell the difference between 8 bit, 12 bit, and 16 bit sound. Choosing levels poorly can reduce your resolution.

Also, there are cheap analog to digital converters which use voltage comparison, which basically linearly "search" for the voltage they should be encoding. So those fundamentally do introduce jitter. (Large changes in signal voltage get encoded a bit later than small ones.) I'm sure there are ways that jitter can be introduced the other way around by cheap DACs. However, if you add more bits and increase the sampling speed, those should largely go away.

Devices like DACs also include a small amplifier, and if those aren't properly designed, there can be some distortion introduced there as well.

That said, there are times when people think they hear a difference, and it's woo. There are other times when the difference is real, however. Audio is a real field with some real science and know how behind it. Please don't grant yourself instant superior expertise just because you've read some stuff online.

(On more than one occasion, I've gotten the "all audio is woo" knee-jerk from programmers who were just reacting to hearing the words "frequency response" not knowing what it was, but labeling it as woo. Ironically, they did remember stuff about Fourier transforms. Go figure.)

> Also, there are cheap analog to digital converters which use voltage comparison, which basically linearly "search" for the voltage they should be encoding.

Integration type ADCs are not necessarily cheap. And you always put a sample-and-hold circuit in front of integration type ADCs, which completely eliminates the sampling time variance.

Also many people are unaware that the Nyquist theorem states, that the _average_ sampling frequency must be at least twice the highest signal frequency. The sampling points may indeed jitter around a little bit, as long as the jitter is distributed evenly. As a matter of fact, adding a little bit of sample jitter can be used as a poor man's dithering, to shift the LSB noise spectrum above the reconstruction filter cutoff frequency.

Jitter is mostly a problem with sample clock distribution. Too much jitter and the sample clock no longer arrives with the right delay with respect to the data, which may create large transients, e.g. whenever the numeric value rolls over some power of 2 (a lot of bits change); this problem could be avoided using Gray codes, though.

Oh, and then there's Sigma-Delta Modulation.

Regarding your usage of 'cheap', I would read this as the difference between a 50c or $5 component. I can understand spending more to get a reputable brand compared to buying the cheapest no name item from AliExpress.

However audiophiles are a whole different level. People spend tens of thousands of dollars on items that have questionable difference in quality. I guess people do the same with cars though, a Toyota Corolla will get you to work just as well as a Rolls Royce Phantom.

Wine might be a better comparison (at least a Rolls and a Toyota look different). Many could tell the difference pretty well between a $5 and $20 bottle (hint: the $20 one is the one not loaded with sugar), but not a whole lot of people can tell the difference between a $20 bottle and a $200 bottle. The law of diminishing returns kicks in pretty quick with both wine and audio. I think with audio once you’re in over a few hundred bucks for your set that wall comes up pretty quickly.

You got it. There are degrees here. You can be an "audiophile" and not break the bank.

> difference pretty well between a $5 and $20 bottle (hint: the $20 one is the one not loaded with sugar)

Wine loaded with sugar??? AFAIK, in a dry wine all the sugar is converted to alcohol- even in the case it's been added before the fermentation, which is still pretty rare, at least in major wine producing countries. That is definitely not the difference between a $5 and $20 bottle of wine (in Italy you can still get excellent wine for $5 a bottle if you buy it from the producer, btw).

whether it's rare depends on the wine. most champagne has added sugar before fermentation.

search for "brut nature" which for cava, champagne and the like means it's only the sugar in the grapes themselves.

If it's added before fermentation, it's still not in the final product, as it's been converted to alcohol. But you're right, champagne, even the best brands, has some sugar added after the fermentation, about 12g/l. It's part of the method. It's still a rather tiny amount if you consider that an orange juice without added sugar has about 90g/l (and you do drink it by the litre, as opposed to champagne...)

This is incorrect, there is always "Residual Sugar" [0] in wine, even when it's dry. Of course, GP is also not correct as there are plenty of wines with high RS that are very expensive while I can get a 7g/L RS wine for around 4€ here.

[0]: https://en.wikipedia.org/wiki/Sweetness_of_wine#Residual_sug...

It also depends on the wine... The difference between a typical $10 bottle of pinot noir and a better $30 bottle is like night and day... But most varieties aren't as delicate and hard to make, so the differences get way more subtle.

Yeah, and sometimes one car is better than another car, even if they have the same engine ($5 wolfson DAC), because more love is given to everything around the engine.

If you just want to put together a cheap car with a good engine, that's a different thing from wanting to build a comfortable or speedy car with a good engine. Someone who isn't in the market for that extra level will just be a sceptic and say "this Audi is just a volkswagen with a prettier body", but if you sit in one you'll feel the difference. (not sure if Audi's have VW engines)

"More love" in this case doesn't have to mean more than maybe a retail price of $50 or so, to get a DAC with a frequency response so flat and a noise level so low that any difference to more expensive items is purely academic.

Digital to analog conversion is a solved problem, with inexpensive commodity parts. Even if you buy an expensive DAC, the studio gear used to produce the music you listen to has gone through dozens of A->D->A conversions, all done using standard commodity parts. No one builds fancy audiophile-approved DACs and ADCs into mixer or studio gear.

In my home setup, I'm using a DBX DriveRack PX as a DSP EQ and crossover between my main monitors and subs. It's almost a decade old, but its signal/noise ratio is on par with the newest gear you can buy. Any competently built pro gear from the last ~10-15 years with 24-bit conversion can be plugged right into any analog signal chain and be 100% audibly transparent.

People worry way too much about the electronics, and way too little about their speakers and rooms.

You think they're overpriced, but I bet you no one is getting (very) rich off these $400 DACs. Even more, I think many of them profit so little every once in a while a company pivots or goes out of business.

An Audi also has just a couple thousand of extra love worth of physical cost added to it, but the retail price is easily upwards of $20.000 above the equivalent VW car. That's because if you sell fewer units, the r&d and marketing costs per unit are much higher, as well as the required profit per car sold.

If you get a good deal on your hardware, you might have that good DAC in a nice box with a pretty manual for $50 physical costs, then put it on a prime shelf in a nice Hi-Fi store for $130 (30% of $400), maybe $50 to target and prime your customer, and then use $150 to pay your audio engineer, your CEO, maybe outsource the rest of the company infrastructure, and then 8% interest to your lenders. That's 1.7M profit at 10.000 units sold (assuming that you do all of this within a year).

I have no idea if 10.000 units is close to reasonable for an indie hardware startup, but calculations become a little more complex if we're going to have to take a multi year ramp up in account..

edit: Sorry I got caught up a little in the startup aspect of it. Your system sounds great, but your DAC (and DSP+crossover system) also cost $400, so you're not making a point?

No one is getting rich off DACs, except for the makers of outrageously overpriced 4- or even 5-figure retail price units. After all, DACs are commodity hardware, boutique devices serve primarily serve as audio jewelry, with no technical benefit.

The DAC I use cost less than $30, yet performs well beyond the limits of human hearing. The DriveRack did have an MSRP around $300 when it was new, but usually $200 on sale, but it does do a whole lot more than a DAC, and the base price level for pro gear is somewhat higher than for consumer gear.

As I said, people focus way too much on electronics, when even basic devices are more than good enough. I don't even have to use the DAC anymore, after I switched to DisplayPort for my monitor, which has a perfectly good analog stereo output. The only reason I keep the Toslink DAC hooked up is pure laziness.

It's almost as if there's a middle ground to be had somewhere.

There are definitely things which can perceptually detract from audio which people just don't think about.

I'm about to buy a couple power conditioners for my electronics because the wiring in my new apartment is garbage and every little spike in voltage comes through as a loud pop in my speakers and headphones.

But more subtly, my audio is constantly being degraded because I have a few cheap dimmers which cause a constant interference pattern in my electrical circuit that weren't a problem in my last apartments, and I can even hear fluctuations caused by my refrigerator if I listen closely.

And all of these things which can subtly degrade audio are 100x more of an issue when you're recording audio. Run any given song through a spectrogram analyzer and you'll usually see a horizontal line or two which could be anything from a little cable noise to a CRT screen.

Wow, that is pretty bad wiring. I can hear an audible hum from my subwoofer when my wife runs the vacuum or blow dryer but that’s about it. And those are about the highest pulling appliances you could have in a household and the sub is pretty big with pretty powerful amp.

Brushed AC motors degrade power quality, although they're normally equipped with filters.

Right, similarly to my dimmers that constant gating of power draw as the brushes make and break contact just wreck your AC circuit.

Bad wiring and a sub-par setup will make that bleed right through your speakers.

Not sure why you got downvoted, I've seen crappy electricity in my own home office after I got a UPS for my computer system. It would have its fans kick in regularly to modulate the voltage that came in at a crazy high ~255v (Australia).

I got the APC Line-R 1200VA Line Conditioner With AVR installed in front of the UPS and couldn't be happier now - it steps down the voltage and gives the UPS peace of mind.

Try the APC, not the most expensive thing but it's worked well for me.

Installing a UPS behind the conditioner sounds like a good idea. I have been meaning to get one of those... Thanks for the rec on the APC, I bookmarked it for comparison later. What UPS are you using?

I am using the CyberPower Value GP LCD 1500VA/900W Line Interactive UPS, but it looks like it's been updated to be named "SOHO" now:

https://www.scorptec.com.au/product/UPS/UPS/44933-VALUE1500E...

I also have one of their smaller BRICs for my HTPC / home NAS JBOD server: https://www.scorptec.com.au/product/UPS/UPS/40279-BR850ELCD

The voltage in that part of the house appears to be better, but the circuit is easily tripped when running toaster + micro or similar, so having these UPSes has taken a lot of frustration away.

Actually just checked my invoice and I got them both in 2012, which is a little while ago now, and they're still good with their original batteries.

Good DACs and ADCs are useful for recording studios who have a lot of analog outboard gear (where the signal leaves and enters their digital setup multiple times). But for normal listening where there are only one or a few conversions, the point of diminishing returns is pretty low.

Most of the perceived gains in quality of higher end equipment (DACs, Amps etc.) come from the higher voltage of the outputs (where you connect your speakers and headphones), this translates into an increased loudness/volume level.

So if you want to do blind tests between different kind of equipment, you have to measure that voltage and make sure it's equal across different pieces of equipment (you can adjust in realtime by volume knobs), otherwise the louder one will always win and perceived as higher quality.

I have a decent hi-fi audio system at home, and there is a clearly audible difference between playing from my high-end CD player, and playing FLAC from the audio out of my Macbook.

The difference is not one of loudness but of 'energy'. A snare hit played from the high-end cd player sounds like someone actually hitting a snare; you can feel there's actual energy in the hit, while the macbook out sounds flat, even at higher loudness.

I'm not sure what is the exact cause but I've always guessed that it comes from the qualities of the built-in amps in the DACs.

> The difference is not one of loudness but of 'energy'

Aren't those the same? Put more joules into the speaker wire and more joules of sound pressure waves will come out of the speaker cone, which just means the sound will be louder (until you reach the mechanical limits of the speaker, of course, but no DAC in the world could possibly change those).

>Aren't those the same?

Not really. If you take some of the transients out of recorded drums they will sound lifeless, without "energy", but not necessarily quieter.

That sounds like amp presets to drop bass, which would make perfect sense on an Amp expected to be used on laptop speakers.

I think the general problem is mid-range speakers and reasonable settings is all it really takes to have quality sound, but people want to be above average and therefore look for ways to discriminate between better sound and worse sound differences they can't "yet" hear. From there you get into placebo and random correlations.

It can be the case that the output of your Macbook isn't really flat, or that you have some kind of software processing enabled. You could test a $20 class compliant Behringer UCA202 USB interface against it, which is flat, and try to disable processing or use a software which just passes the signal to the interface.

It can also be your room. If you're not exactly located in the same listening position the listening experience can change a lot, even with the same setup.

Your room is basically an equalizer and should be regarded as a part of your audio system. That's how some Hifi sellers can trick you into more expensive setups, they just change the speaker positions or your location to optimized defaults. It's crazy how they can trick people.

A harder hitting snare can be caused by changes to the upper mids and lower highs, by anything that is located between you and the speakers, and then the later reflections of the room which also add up.

Consider audio benchmarking.

For Widows there is Rightmark, maybe you could run the Macbook's output against a decent Windows DAW.

audio.rightmark.org

Another possibility is that your CD player might not be flat. Marantz, Yamaha and other manufacturers add their slight sound signatures to their higher and models.

I never benchmarked my Yamaha CD-S300 to other devices, but its flyers and web site never says it's completely flat on the analog outputs. The reviews of the player notes that it has the distinct Yamaha signature too.

Also "Monitor" and "Flat" devices as their name implies sounds pretty flat and dull.

All-in-all you should consider taking response graphs of your devices rightmark.

Reply on news.ycombinator.com