Let's develop an ASR inter-sample test procedure for DACs!

[splitting_ears]

Great, now we can see (and hear) those harmonics. You can even go 2 octaves lower. It's the low frequencies that are most likely to clip, and when they do, the harmonics are sprayed all over the mids and treble where our hearing is most sensitive.

I think you have misunderstood me. I made this graph to compare the expected behaviour (harmonics related to the fundamental being clipped) with the incredible graph posted by @edechamps (where it looks more like white noise than a nicely ordered harmonic series).

This appears to be a nearly perfect square wave. Not unexpected with hard clipping.

Yes it is. And how do you interpret this, by comparison?

 

[MRC01]

Please. This has nothing to do with production quality. The decoded peaks will always be higher than the encoded peaks, minus one *very* specific constructed case.
This is not up for debate or judgement, it is a logical consequence of the sampling theorem.

I disagree. It has everything to do with production quality. It's a result of the loudness wars, studios mastering albums to be as loud as possible.

Sure, decoded peaks are higher than encoded peaks. This can be seen in the Shannon-Whittaker reconstruction formula, which shows that the amplitude of the analog wave for each sample is an infinite sum involving ALL sample points. But in a properly made recording they don't go over full scale. That infinite sum decays very quickly as adjacent sample points get further away. You're missing the forest for the trees, using a rare exception that is easily avoided, to justify egregious abuse of the recording format to promote every-louder recordings.

 

[popej]

You're missing the forest for the trees, using a rare exception that is easily avoided, to justify egregious abuse of the recording format to promote every-louder recordings.

No, we will be promoting bugy DACs, which in turn improve mastering quality
But we need to test them first.

 

[pkane]

Yes it is. And how do you interpret this, by comparison?

This is an overflow error doing calculations in integer arithmetic, with values exceeding the maximum.

 

[splitting_ears]

This is an overflow error doing calculations in integer arithmetic, with values exceeding the maximum.

Well, I have to admit that you lost me with that one
Would you mind explaining a bit more?

 

[danadam]

Sounds like you're asking for ReplayGain scanning with loudgain.

Or volume normalization offered by streaming services.

 

[kemmler3D]

missing the forest for the trees, using a rare exception that is easily avoided, to justify egregious abuse of the recording format to promote every-louder recordings.

I don't think anyone is trying to justify anything here. I see this thread as looking for a method of evaluating how DACs might react to faulty recordings, nothing more.

Unfortunately, good music has been put into compromised or even broken masters for decades. Looking for equipment that can compensate for that may or may not encourage bad habits at the studio, but that's also not really our problem. I would say this is not philosophically any different than looking for a tonearm that can handle improperly mastered bass on vinyl.

 

[popej]

About overflow: https://en.wikipedia.org/wiki/Integer_overflow

 

[splitting_ears]

Again, and because that bears repeating: most DACs will not "render the peak at >0dB", because most DACs are not designed with ISOs in mind. It's "undefined behavior", to borrow a term from software engineering. When presented with ISOs, some DACs will clip inside the DAC itself; other DACs will break horribly and output a waveform that may be corrupted beyond recognition. The amp doesn't even factor into this equation.

I couldn't have said it better. Aren't you just curious? A whole new territory to measure?

I'm not going to keep posting in this thread indefinitely if nobody cares. And please don't try to paint me as some kind of advocate of loudness wars. That's very much against the way I work on a daily basis as a mastering engineer.

Now, could we please step back from all this debate and discuss together this compromise I have tried to suggest on post #115?

Why not, as a first step, specifically test ISP behaviour on devices with no digital volume control, or no volume control at all?

In other words, CD players, stand-alone DACs, digital interfaces, or DACs with analogue volume controls. For these, the problem is neither intuitive nor easy to solve by the user. I think almost everyone uses their music player at maximum volume so it's a valid concern.

It won't hurt anyone (or any ratings), it won't encourage stupid practices, simply promote good engineering or point the user towards a reasonable solution. @amirm , sorry to tag you on this specifically, but I'd very much like to know your opinion on this.

 

[MRC01]

I don't think anyone is trying to justify anything here. I see this thread as looking for a method of evaluating how DACs might react to faulty recordings, nothing more.

Fair enough! As an engineer I am curious about it, even though I do see it as accommodating intentional abuse of the recording system. Yet one could also see this as testing how DACs handle special cases, much like testing CD players to see how scratched a CD must be before they fail to play it.

However, the following sounds to me like fatalism that amounts to accepting as standard practice intentional abuse of the recording format.

... This has nothing to do with production quality. The decoded peaks will always be higher than the encoded peaks, minus one *very* specific constructed case. This is not up for debate or judgement, it is a logical consequence of the sampling theorem.

Perhaps it wasn't intended this way. I don't mean anything personal. My point is that if the decoded peaks go above full scale, the recording is too loud. And it is not mysterious or complex. It is easy to detect when this will happen, and easy to correct for it. So there is no excuse for any recording to do this. It is intentional abuse of the recording format.

 

[amirm]

In other words, CD players, stand-alone DACs, digital interfaces, or DACs with analogue volume controls. For these, the problem is neither intuitive nor easy to solve by the user. I think almost everyone uses their music player at maximum volume so it's a valid concern.

If it isn't intuitive or known by the user, how would they seek out such DACs? It is not like you are proposing a solution that retroactively solves their problem.

In contrast, I am proposing a solution like that. By having the player manage this, all DACs automatically get a solution. The feature is built-into Roon already:

Please note that I am not dismissive of the problem. It is just that the bar is extremely high in adding new tests to reviews.

 

[MRC01]

Even an old DAC chip like the WM8741 has 2 settings related to this: one for overall gain, another for the digital filters.

 

[popej]

By having the player manage this, all DACs automatically get a solution. The feature is built-into Roon already

So you sacrifice some dynamics, but how do you know if it is really needed for a particular DAC?

 

[amirm]

So you sacrifice some dynamics, but how do you know if it is really needed for a particular DAC?

Presumably because you have an audible problem. Roon has an indicator for this which I use in addition to noise when clipping occurs due to EQ. I actually tune that value less than theoretical value as it is content dependent. But you can set it to a high number and not worry in that regard.

If you don't have an audible problem, then there is nothing to worry about. Alternatively, if you are hearing something, you can use that feature to increase headroom and see if it is still there. If it is, then that wasn't your problem. If it is, then you have solved it.

 

[popej]

If you don't have an audible problem, then there is nothing to worry about.

The older I am, the less I perceive audible problems Still I would like to have technically good equipment.

I don't expect to hear hard clipping of ISP and I don't think crippled DAC are common. But the problem of a test for ISP is interesting itself. I'm surprised, that people come here only to discredit the idea. My thinking is that a correct test can at least show whether the problem is worth worrying about.

 

[mdsimon2]

If it isn't intuitive or known by the user, how would they seek out such DACs? It is not like you are proposing a solution that retroactively solves their problem.

In contrast, I am proposing a solution like that. By having the player manage this, all DACs automatically get a solution. The feature is built-into Roon already:
View attachment 322807

Please note that I am not dismissive of the problem. It is just that the bar is extremely high in adding new tests to reviews.

And if you use the SPDIF input on your DAC you are just out of luck?

Michael

 

[AnalogSteph]

Even an old DAC chip like the WM8741 has 2 settings related to this: one for overall gain, another for the digital filters.

Proof that the folks at Wolfson Micro knew what they were doing (first discussion of ISOs dates from 2003ish, so being aware of the issue pre-2007 was quite feasible), but I don't think it's a feature that really caught on long-term, and now Wolfson is basically dead and all the relevant DACs are EOL. Of course nowadays your average playing device is a computer anyway, so giving up a bit of headroom should be fairly trivial.

 

[amirm]

And if you use the SPDIF input on your DAC you are just out of luck?

Michael

Why? Roon player will output correct values over SPDIF just as well. Or do you mean from a non-computer source? If so, once again, buy a DAC with volume control. That is the correct solution regardless of this issue anyway.

 

[mdsimon2]

Why? Roon player will output correct values over SPDIF just as well. Or do you mean from a non-computer source? If so, once again, buy a DAC with volume control. That is the correct solution regardless of this issue anyway.

Non-computer source. And if your DAC has an ASRC upstream of the volume control (cough, cough miniDSP) then what?

Michael

 

[levimax]

I have been following this thread which I find interesting but I am not sure if one of the central premises that "turning down the DAC volume prevents digital clipping" is correct. I have a SMSL DO100 DAC with a volume control which I use with Foobar2000 and it's convolver with FIR corrections for room and speaker. To test for digital clipping I run a 0 dB sweep and listen and it is quite obvious when clipping occurs. I find I need to use -8 dB attenuation with the convolver to prevent all audible clipping. After reading this thread I thought rather than use the attenuation in the convolver I would just set the convolver attenuation to 0 dB and turn down the volume on the DAC until the 0 dB sweep did not audibly clip. This did NOT work. No matter what the volume level on the DAC was set too there was audible clipping when running the sweep (I have a analog pre-amp so I can listen for the clipping at reasonable volume levels). I am not sure how the volume on the SMSL DO100 works but it does not appear to work in the digital domain and I have to believe many other DAC's are the same and digital clipping could be a bigger issue than most people realize.