[Dobrescu George]

  If you consider audible sounds as sonic as in something a human can hear, what would you think that ultrasonic means?

well... the people who made the defectoscopy meter thought that 16 khz is ultrasonic...
 
I personally consider ultrasonic anything above 24-28Khz, but because i think that there are people who can hear until those ranges. 
 
Of course, i might be wrong.

 

[Steve Eddy]

well... the people who made the defectoscopy meter thought that 16 khz is ultrasonic...

I personally consider ultrasonic anything above 24-28Khz, but because i think that there are people who can hear until those ranges. 

Of course, i might be wrong.

Let's for the moment say that there are.

At what levels?

If you're familiar with the equal loudness curves, you'll see that our hearing is least sensitive at the extremes, meaning that frequencies at the low end and the high end need to be louder compared to mid frequencies to be heard. And at the high end, sensitivity drops pretty quickly.

Now consider the spectral makeup of real instruments. Certainly no instrument has a fundamental in that range, only harmonics. And those harmonics drop off pretty quickly with increasing frequency.

So given this, and our decreased sensitivity at high frequencies, and given the loudness of the fundamentals, are there any instuments that produce harmonics at those frequencies that would be sufficicient to be audible even to those who may have some sensitivity up to 24-28 kHz?

se

 

[StanD]

  well... the people who made the defectoscopy meter thought that 16 khz is ultrasonic...
 
I personally consider ultrasonic anything above 24-28Khz, but because i think that there are people who can hear until those ranges. 
 
Of course, i might be wrong.

By definition ultrasonic is above audible give it up already. You are not doing a good job of mincing words.

 

[KeithEmo]

 
OK, I would have guessed the ''perceived headstage'' of audioplayers like the Sansa could be explained by that... BUT if the differences are quite small and -88 is already really good I think it has nothing to do with the ''headstage'' differences between audio players.
 
So how could the Sansa's headstage be small then, ignoring placebo.

 
I would have to disagree entirely there. The "cost" of having too much crosstalk is that you can't have a sound appear in one side and not the other. While someone may record a sound in only one channel on a test disc, this simply doesn't happen in real life, and it wouldn't matter anyway (you've surely never heard a situation with speakers where sound from the left speaker couldn't reach your right ear at all - unless you put your head in a hole in the wall with each ear in a different room - and most people complain that headphones have too much separation). In reality, any system with a separation of 30 dB or 40 dB or more is going to reproduce any actual music you have just fine....  at least as far as channel separation.
 
When you talk about imaging and sound stage, the most critical factor is almost certainly phase response (between different frequencies in each channel, and between the channels at various frequencies), since timing and phase are what our brains mostly use to figure out where something is apparently located.

 

[cjl]

   
I would have to disagree entirely there. The "cost" of having too much crosstalk is that you can't have a sound appear in one side and not the other. While someone may record a sound in only one channel on a test disc, this simply doesn't happen in real life

Sadly, this isn't entirely true. Some recordings (especially classic rock) hard pan sounds to one side, to my great annoyance. They're OK on speakers, but near unlistenable on headphones. Of course, this isn't the kind of thing that would be helped by better channel separation.

 

[maverickronin]

  Sadly, this isn't entirely true. Some recordings (especially classic rock) hard pan sounds to one side, to my great annoyance. They're OK on speakers, but near unlistenable on headphones. Of course, this isn't the kind of thing that would be helped by better channel separation.

 
DSPs to the rescue.
 
Alternatively...

 

[KeithEmo]

  Slight problem, I will need it to support 192khz (I know there is no sound difference), but I wanted it to run some tests!
 
Also, am I correct in thinking that a USB to Coaxial output is strictly digital, and therefore unless the product is completely faulty, there shouldn't be any degradation in the fidelity?

 
Yes - and no.
 
A digital audio signal consists of data and a clock. Unless, as you say, it's completely faulty, a USB-to-S/PDIF converter should be able to convert the data without losing or altering any of the bits. This leaves the clock.... and any variations in timing it might have - like jitter. In the old days, many USB devices, including converters, used USB isochronous mode, which allows the computer to provide the output clock for your USB data. Because of the way USB works, in this mode, the output jitter is often so bad that it exceeds even the ridiculously high numbers that "jitter doubters" agree are audible. Luckily, most modern DACs and converters use asynchronous USB mode, where the clocking is controlled by the DAC or converter instead of the computer, and the jitter should remain within reasonable levels. (Benchmark DACs are immune to jitter anyway, so it shouldn't matter anyway.)
 
However, some DACs are not immune to jitter and, with them, at least some of us are convinced that we can hear the difference between converters that have good jitter specs and those that do not. (Again, since the Benchmark has its own jitter removal circuitry built in, this should definitely not be an issue with it.)
 
 
 
 

 

[uchihaitachi]

   
If you are set on getting the DAC1, get the version with the USB.  This device would certainly be the end game for almost everyone, so you may as well future-proof it by getting one with the most options available.  It will also make it easier to sell once you come to your senses and use a $50 DAC later. 

 

I know it's inaudible but I was offered the DAC1 at such a good price. Can't resist the temptation!

 

[KeithEmo]

Could you possibly explain what you mean by it's ground loop and optical can solve that?

I presume the other USB to Coaxial output that is selling for crazy amounts is purely snakeoil?

 
The electrical ground on computer power supplies is notoriously noisy. Since USB DACs and USB-powered converters run on the computer's electrical system, sometimes this ground noise can leak into the analog circuitry (you hear funny little noises, and sometimes you can actually "hear" what's happening on the computer screen). The way to eliminate or prevent this is to electrically isolate the ground on the computer from your audio system. Since Toslink connections are optical, and so don't have a ground connection at all, as long as your audio is passing through a Toslink connection, it will be isolated, and so immune to this problem. The Coax and other inputs on many devices are also electrically isolated, but not all, and many USB connections are not.
 
While a lot of the fancy USB audio gadgets are indeed snake oil, and that's especially true for fancy USB audio cables, some of the higher-end converters do have jitter specs that measure far better than the lower cost ones. So, in terms of measured electrical performance, you are sometimes getting something for your money. As usual, some folks find there to be an audible difference, while others insist that there is not. (Since some DACs, like the Benchmark models, have circuitry that renders them immune to jitter, while the sensitivity of other DACs to jitter may depend on specifics of their design, it's not unreasonable to expect some DACs to be affected a lot more by jitter than others.)

 

[cjl]

   
Yes - and no.
 
A digital audio signal consists of data and a clock. Unless, as you say, it's completely faulty, a USB-to-S/PDIF converter should be able to convert the data without losing or altering any of the bits. This leaves the clock.... and any variations in timing it might have - like jitter. In the old days, many USB devices, including converters, used USB isochronous mode, which allows the computer to provide the output clock for your USB data. Because of the way USB works, in this mode, the output jitter is often so bad that it exceeds even the ridiculously high numbers that "jitter doubters" agree are audible
 
 
 
 

I'd be curious to see a source for that - I was under the impression that even isochronous dacs can have jitter levels way below anything that would be remotely audible.

 

[KeithEmo]

Unfortunately it doesnt

If you were in my position, use a benchmark dac1 with a converter or get the benchmark dac1 usb for a slight premium 150USD more, what would you go for?

 
Considering that even a cheap converter that goes up to 24/192 is going to cost between $50 and $200 or so, and it will be an extra box and an extra cable to find homes for, I'd spend the extra $150 to get it all in one box. (Besides, as someone else already said, the resale value on a DAC without a USB input isn't going to be nearly as good if you do decide to sell it later.... and, besides all that, not all converters work all that well, and I would expect the USB section build into the Benchmark to be pretty good.)
 
(Also, as someone else suggested, you might also want to look at a Schiit Audio Bifrost.....    )

 

[KeithEmo]

  On the other hand, coaxial (rather than optical) S/PDIF is not isolated, and is therefore not protected from ground loops. The expensive USB to coaxial S/PDIF converters claim to reduce jitter, and may or may not implement isolation as well. In other words, they are indeed likely to be snake oil, as jitter is not normally audible with a DAC of any reasonable quality. Optical S/PDIF may have higher - but still inaudible - jitter than coaxial, that is why the converters use the latter, as the people buying them tend to worry more about jitter than ground loops.

 
Actually, you've got to read the specs - many Coax inputs are in fact isolated, usually by coupling transformers (although certainly not all of them).

 

[dprimary]

Unfortunately it doesnt

If you were in my position, use a benchmark dac1 with a converter or get the benchmark dac1 usb for a slight premium 150USD more, what would you go for?

I would just pay more for the USB version. One less thing to go wrong.

 

[Steve Eddy]

The electrical ground on computer power supplies is notoriously noisy.

It's not so much the ground as much as it is the interchassis currents caused by the fault return path (that third pin on the IEC).

se

 

[dprimary]

  You might want to look into a Schiit Bifrost Uber USB DAC. It has all 3 inputs, sounds great and costs less.

The Bifrost is a completely different class of equipment from the Benchmark. Schiit is great sound low price, designed for home use. Benchmark is balanced, rack mountable. I buy gear from Schiit pretty often, but I don't view it as a professional equipment.