God of SINAD vs. reality we get from most available music files

[Blumlein 88]

No. Proper dither in the lack of noise shaping should be flat-out white. Most emphatically NOT LF.

But the psd of the room is the real key. THAT is more likely to be pink/brown/something not described by a distribution.

I would suspect the Sibelius, if it's on a quiet part, is A/C noise, then. Wondering where the dither got to.

Looks like HVAC noise to me too. I've recorded in some churches and you get noise like that if the HVAC is running. Those long ducts and the air movement really clutter up the low end. Sometimes you'd see significant levels when no one was playing. I ended up cooling or heating the building a little extra ahead of time so we could turn it off to get a recording done. Of course you can also use a high pass filter if you don't have instruments with output down low.

 

[j_j]

But this Sibelius, in contrast to my other examples, is a 24-bit recording. Wouldn't that move any dither way below and what's shown is the actual noise in the room/concert hall (wherever they recorded it)?

It would certainly drop the level of dither, but I doubt that any recording is actually dithered at 24 bits, that puts rather some constraints on the sources thanks to basic physics.

And, still, the spectrum would be flat. Non-flat dither that is not noise-shaped (which would almost certainly push the noise to higher frequences) does not want to have a lot of lower frequencies. This relates to the decorrelation of quantizing noise.

I really think there's quite some HVAC noise there. This is not unusual.

 

[Blumlein 88]

You can always take that little section and digitally amplify it until it is easily heard. That likely would make it clear whether or not it is some ventilation noise.

 

[restorer-john]

You can always take that little section and digitally amplify it until it is easily heard. That likely would make it clear whether or not it is some ventilation noise.

Could be HVAC, could be traffic, could be subway trains. The amount of extreme LF from traffic is totally underestimated. Even over 10km away from a major highway, it can be detected at our house as it's so quiet otherwise.

 

[j_j]

Could be HVAC, could be traffic, could be subway trains. The amount of extreme LF from traffic is totally underestimated. Even over 10km away from a major highway, it can be detected at our house as it's so quiet otherwise.

Yeah, or airplanes. We're under the SEA approach often. You can see LF noise that seems to look for all the world like something relating to a half-wavelength across the wing spread of the planes going overhead.

 

[spacevector]

When I first made REW measurements, I used to wonder where all the low-bass is coming from. It was neat to learn its just ambient noise.

 

[AdamG]

Could be HVAC, could be traffic, could be subway trains. The amount of extreme LF from traffic is totally underestimated. Even over 10km away from a major highway, it can be detected at our house as it's so quiet otherwise.

This, Trains, Airports, Military Bases, tunnels, highway, bridges, industrial parks or buildings. Low frequency energy can travel large distances and retain energy enough to be measurable. Bowling alley, gun range, paper mills, even neighbors subwoofers.

 

[Blumlein 88]

In my experience the quietest night of the year is Christmas. People are at home, people are off work, many factories are closed. You get the quietest measurements then and lowest low frequency garbage.

Oh, I live about a half mile from a major rail line. I can setup a microphone with good low response (omni) and tell you when a big train is coming many miles before it gets here.

 

[GXAlan]

This, Trains, Airports, Military Bases, tunnels, highway, bridges, industrial parks or buildings. Low frequency energy can travel large distances and retain energy enough to be measurable. Bowling alley, gun range, paper mills, even neighbors subwoofers.

Godzilla.

What is the bloop?

It took years to determine the source of this mysterious rumble from the deep, but scientists determined it originated from an Antarctic icequake.

oceanservice.noaa.gov

Hypersonic research.

Secret Is Out on 'Quakes': It's a Spy Plane : Aviation: Analysts believe that Aurora, a craft that can travel 4,000 m.p.h., could be the cause of frayed morning nerves.

It happened again Thursday morning.

www.latimes.com

 

[mdsimon2]

Inputs: 2x spdif optical, spdif electrical, usb
Outputs: rca
In essence a transparent dac with dsp that I can also use as a preamp.

Both for streaming audio, playing blurays and cds and an occasional youtube viewing. The last preferably through hdmi but more practical is viewing via the tv app and sound through spdif I think. Preferably with a streamer included but not a deal breaker if it's not there.

Hard to beat a miniDSP Flex or SHD if you need that many inputs and 4 channels (or less output). Of course an audio interface with software DSP can also do it but it will be more expensive.

Michael

 

[Marc v E]

Hard to beat a miniDSP Flex or SHD if you need that many inputs and 4 channels (or less output). Of course an audio interface with software DSP can also do it but it will be more expensive.

Michael

Thank you! Are there any audible difference between the flex and shd when used a preamp ? Or is the higher sinad of the shd not relevant anymore in this respect ?

 

[jhaider]

I also understand why for many this would pose as much of a problem as many of us use an avr for home cinema, which almost always has eq in it. And a seperate hifi solution.

Do you mean in different rooms or in the same room? If the latter...I don't get why anyone would do that?

In my situation however, I want to combine video and audio which leaves me with 4 options:
1) buy a minidsp to get eq for audio, movies, youtube etc. 500 to 1500 euros
2) buy Roon and accept only eq for streaming audio. 600 + 500 euros for a nuc
3) fix eq in every source possible with software. No solution for youtube and movies as I use 2 channel instead if an avr.
4) buy an avr with less than transparent spec. Probably the best for 2 channel available is the nad m33 for 6000 euros.

I don't see that as remotely a dilemma, but it does highlight the fundamental problem with the focus on SINAD. Either you run everything through a 2-channel system and use a modern AVR/intergrated/preamp or a miniDSP or similar between preamp, or if you are multichannel use a modern AVR. In real world use, either one is likely just fine for either use case. Apparently per Amir's measurements AirPlay limits SINAD to 90dB or thereabouts. But so what? It's fine. I have no qualms AirPlaying into an WiiM and then going digital out to RME ADI-2 for headphone/IEM use. It's fine.

As for "less than transparent spec," based on what? If the choice is between processing to improve the fit between room and system (and maybe anechoic correction of the loudspeakers) that might have lower SINAD than something else, or some super-SINAD box that results in shit frequency response in the modal region - how is that possibly a dilemma?

The main problem I think is that many hifi companies are small and not capable of implementing eq transparently. Let alone audiophiles willing to pay for it...

"Transparency" isn't the issue, IMO. The product design issues are more prosaic - programming the math, designing the UI. The greater issue in my view is simply will - it's easier to chase marginal improvements on already inaudible measured values then it is to make more flexible gear, and there's an cheerleader ecosystem in place to support companies that pursue that approach.

 

[solderdude]

A more recent very LF noise source is windmills.
In our pretty crowded country they are often too close to homes

 

[Thomas_A]

LPs have their share of LF noise and one good example of fan noise is the Cowboy Junkies Trinity Sessions. Even above the LP noise I guess.

 

[bluefuzz]

On DSP, I use powerful and transparent version in my software player. I have no use for it in the DAC. To wit, I don't use the one in my RME DAC.

To be sure, it is a highly valuable feature if you can't do the same in software. But if it is important to you, why don't you pick a player that can do it?

I think a lot of people, if not most, have need of the DSP/room correction for multiple sources, e.g. tv, blu-ray, a streamer or two, plus maybe an analogue source. That's not practical to do in software ...

 

[PierreV]

full report worth a read.

Environmental noise in Europe — 2020

The report presents an updated assessment of the population exposed to high levels of environmental noise and the associated health impacts in Europe, based on the new World Health Organization (WHO) recommendations. The report also documents actions being taken to manage and reduce noise...

www.eea.europa.eu

 

[tmtomh]

I don't see that as remotely a dilemma, but it does highlight the fundamental problem with the focus on SINAD. Either you run everything through a 2-channel system and use a modern AVR/intergrated/preamp or a miniDSP or similar between preamp, or if you are multichannel use a modern AVR. In real world use, either one is likely just fine for either use case. Apparently per Amir's measurements AirPlay limits SINAD to 90dB or thereabouts. But so what? It's fine. I have no qualms AirPlaying into an WiiM and then going digital out to RME ADI-2 for headphone/IEM use. It's fine.

As for "less than transparent spec," based on what? If the choice is between processing to improve the fit between room and system (and maybe anechoic correction of the loudspeakers) that might have lower SINAD than something else, or some super-SINAD box that results in shit frequency response in the modal region - how is that possibly a dilemma?

"Transparency" isn't the issue, IMO. The product design issues are more prosaic - programming the math, designing the UI. The greater issue in my view is simply will - it's easier to chase marginal improvements on already inaudible measured values then it is to make more flexible gear, and there's an cheerleader ecosystem in place to support companies that pursue that approach.

I get your points here (and I clicked Like on your post). At the same time, though, I don't quite get the level of heat around banging on against SINAD. SINAD is a property of each individual piece of gear, and is useful for ensuring that no component in your chain is potentially adding unnecessary levels of noise or distortion (compared to similarly featured gear that could be had for the same or lower price). The need for DSP in the modal region is a system question - as the recent few pages of this thread show, there is not just one answer to that question; it depends on one's system(s), source components, preferences, and so on.

So once again - and sorry to repeat myself but I think it bears repeating - @pma 's bizarre crusade against SINAD in favor of high-frequency square waves is framing this discussion in a way that makes no sense. It would be awesome to have a thread dedicated to the challenges and options for integrating DSP in multichannel systems, in systems with analogue source components, and so on.

And, crucially, such a thread would be full of information and free of the heat that's in this thread, because it would not be framed as a silly war against SINAD.

 

[GXAlan]

I think SINAD will be “done” in the next 5-10 years. We still see bad products, but the AVR-X8500 broke the 100 barrier 5 years ago, and we are seeing great desktop DACs beyond the 120 dB range. A few product cycles from now, I can imagine products generally all doing well. SINAD will only be unimportant because everyone has stepped up.

Looking at our audibility threshold

Strict
Dynamic range, linearity, SINAD: 120 dB
THD, IMD, noise, crosstalk, jitter: -120 dBFS / 0.0001%
Frequency response, channel balance: ±0.1 dB
Output impedance: 0.16 ohms

Is there more crosstalk or IMD testing to be done?
1) we see that multitone noise patterns can vary with frequency. Some are flat, others spread wider as you go up in frequency.

Are there equal SINAD products with “vastly different” 32-tone results? Or does SINAD closely correlate with the 32 tone results?

2) The current 32 tone is evenly spaced 0 dB tones. Is there a difference if you have a mix of 0 and -10 dB tones?

3) what if you did a crosstalk test with 32 tones instead of a sweep? If you played 32 tones that were 1/6th octave shifted between the left and right, might be really interesting to see if there are differences.

Is there any benchmark optimization going on?
In the PC industry, once a benchmark is popular, time is spent optimizing around the benchmark even if it’s a single use case.

What if you did your dashboard at 500 Hz or 2kHz?

If you shifted the tones so they are not evenly spaced and “stress” the region of IMD humps, does it change?

Does output impedance/input impedance vary with frequency? At least when I look at Stereophile, for amplifiers, there is still a big non linearity when they look at frequency response in a simulated speaker load.

But that speaker load is reflective of 1995 speakers

Real-Life Measurements

As mentioned by two readers in this month's "Letters," amplifiers are used to drive loudspeakers but are almost exclusively measured into resistive loads. The reasons for this are twofold: 1) real loudspeakers both produce neighbor-annoying sound levels and tend to break when driven with typical...

www.stereophile.com

Would different speaker loads reflecting popular speakers today be any different?

—-
The next frontier is applying science to mixing and mastering

Modern mixes are compressing dynamic range so that it sounds good on phones and cars, but is that science? Or the same truth as cables and power conditioners? For all we know, in a formal, scientifically designed study, it may prove that you don’t need to compress dynamic range but instead have the car audio system or phone do a speaker specific EQ and speaker specific dynamic compression based upon performance characteristics.

 

[Blumlein 88]

snippage......
The next frontier is applying science to mixing and mastering

Modern mixes are compressing dynamic range so that it sounds good on phones and cars, but is that science? Or the same truth as cables and power conditioners? For all we know, in a formal, scientifically designed study, it may prove that you don’t need to compress dynamic range but instead have the car audio system or phone do a speaker specific EQ and speaker specific dynamic compression based upon performance characteristics.

I don't think mixing and mastering is going to be a science. At one time doing recordings on reel tape without noise reduction you had to gain ride. Meaning you knew when music would get loud and when it would get quiet. Your turned down loud parts to prevent overloading. You turned up quiet parts so they would be loud enough to be heard and not lost in noise. Essentially manual compression. Hardware compression has attack and release times so it can do that automatically. Various attack and release settings however have a sound of their own. So it became an artistic effect of sorts. So whomever recorded music doesn't want you or some auto system messing with that.

I can imagine music either coded or where it is run thru to decide gain riding for the conditions of your use, but I just don't think that is going to happen. Listening in a quiet living room, listening in a car, listening on a business flight or in noisy conditions could all be done differently. I just don't think you'll ever get music production to give you the uncompressed version because any compression will change the overall sound.

OTOH, I've seen studio people absolutely murder music with super high compression even while being told, "we don't want it compressed". I do not know how the fashion of compressed to the nth degree became so entrenched in those people's business you literally almost cannot pay them not to do it. Some will argue with you that you do want it super-compressed you just don't know it. No,....REALLY,,,,,,I DO NOT!

 

[pma]

Looking at our audibility threshold

Strict
Dynamic range, linearity, SINAD: 120 dB
THD, IMD, noise, crosstalk, jitter: -120 dBFS / 0.0001%

You never will be able to tell it from those numbers being -80dBFS, with a real-life music recordings. Wasting time, wasting efforts. Rather make the thing reliable. The more simple the better. Keep it cool, no local overheated islands.