Are inter-sample overs the new audiophile bogeyman?

[Sokel]

EDIT. Tested it again. Definitely no headroom created by the DAC V1's volume control. There must be something set up incorrectly in your test.

Can't see the noise floor so to see if lowers as well.
If that so there's a chance that windows or REW are at play, either lowering the DAC or the ADC (unlikely for the latter)

 

[OHtaru]

Where did you get the test signal? What player did you use? How did you lower the device volume?

I create the test signal by myself. I used Foobar2000 and I also tried others, which didn't change the conclusion. Other sampling rates and bit depths didn't make a difference either. The volume is adjusted through the volume slider in the Windows taskbar. When adjusting, the DAC displays the volume as -xx dB, and I think this adjusts the volume of the DAC.
Edit: I'll test using other host devices to see whether there will be any difference.

 

[jkim]

I create the test signal by myself. I used Foobar2000 and I also tried others, which didn't change the conclusion. Other sampling rates and bit depths didn't make a difference either. The volume is adjusted through the volume slider in the Windows taskbar. When adjusting, the DAC displays the volume as -xx dB, and I think this adjusts the volume of the DAC.

Use the Mutltone Loopback Analyzer to generate the signal (44.1/24)---it is different from just a sine tone at 11025 Hz. More importantly, there's another way to demonstrate that DAC V1 does not apply software volume control in DSP. Simply set a peak filter with a positive gain (e.g., 6 dB) in DAC V1 DSP with no preamp cut, feed a full-scale sine tone at the filter's center frequency, and decrease volume on DAC V1 to compensate for the peak. You will see the signal overloaded.

By the way, I know this not just based on my own testing but also information (private discussion with) from Dmitry. I believe he will eventually implement software volume control.

 

[Sokel]

If I can clarify my reason for raising the question :

I keep seeing posts from the "general readership" questioning a particular devices ability to avoid ISOs - and clearly considering this is a problem worth rejecting said device over.

Example:
https://www.audiosciencereview.com/...ng-e50-review-balanced-dac.26219/post-2438208

Worse - it is now being used by marketing wonks, alongside jitter - as a justification for their overpriced hardware:
https://www.audiosciencereview.com/forum/index.php?threads/extreme-snake-oil.24765/post-2352433

I guess that is some sort of justification for including a test - so that we can demonstrate that lower cost devices also are immune (where they are)

Or by the misinformed, as an argument as to why DACs sound different:
https://www.audiosciencereview.com/...tra-streamer-preamp-review.57342/post-2164852

Cost is definitely not a factor.

Cheapo as dirt (about 120 euro back then), 100 years old E-MU 0204:

(I love AKM)

 

[Sokel]

I create the test signal by myself. I used Foobar2000 and I also tried others, which didn't change the conclusion. Other sampling rates and bit depths didn't make a difference either. The volume is adjusted through the volume slider in the Windows taskbar. When adjusting, the DAC displays the volume as -xx dB, and I think this adjusts the volume of the DAC.
Edit: I'll test using other host devices to see whether there will be any difference.

You need to set the phase as well, have a look above, it's 11025Hz at 45° .

 

[OHtaru]

Use the Mutltone Loopback Analyzer to generate the signal (44.1/24)---it is different from just a sine tone at 11025 Hz. More importantly, there's another way to demonstrate that DAC V1 does not apply software volume control in DSP. Simply set a peak filter with a positive gain (e.g., 6 dB) in DAC V1 DSP with no preamp cut, feed a full-scale sine tone at the filter's center frequency, and decrease volume on DAC V1 to compensate for the peak. You will see the signal overloaded.

By the way, I know this not just based on my own testing but also information (private discussion with) from Dmitry. I believe he will eventually implement software volume control.

This is really strange. The files generated by Multitone Analyzer didn't make a difference. I even tested directly with Multitone Analyzer, and it was the same. As long as the volume is less than -3dB, there's no clipping. I'm thinking the only possibility is with the Windows volume control, but I tested it using the DSP overload method you mentioned and it was correct. The volume control does occur after the DSP. What happend?
Edit: It can still be reproduced using an iPhone, there is no clipping at -3dB, and the noise floor is clean because there is no ground loop.

 

[Sokel]

What happend?

You measure with 192kHz I/O, that's what happened.
DAC must be outputting 44.1kHz.

Edit: Nope, I saw the settings and not the bar.
There's still lots of interference though.

 

[OHtaru]

Use the Mutltone Loopback Analyzer to generate the signal (44.1/24)---it is different from just a sine tone at 11025 Hz. More importantly, there's another way to demonstrate that DAC V1 does not apply software volume control in DSP. Simply set a peak filter with a positive gain (e.g., 6 dB) in DAC V1 DSP with no preamp cut, feed a full-scale sine tone at the filter's center frequency, and decrease volume on DAC V1 to compensate for the peak. You will see the signal overloaded.

By the way, I know this not just based on my own testing but also information (private discussion with) from Dmitry. I believe he will eventually implement software volume control.

Did the THD compensation of your DAC use the No load preset? I reproduced -6dB clipping with this preset. This is reasonable because this preset uses Voltage mode, and the digital volume is always 0dB.

 

[jkim]

Did the THD compensation of your DAC use the No load preset? I reproduced -6dB clipping with this preset. This is reasonable because this preset uses Voltage mode, and the digital volume is always 0dB.

My bad. I forgot the fact that DAC V1 uses analog volume control in the top range of volume control in its Voltage mode!

And the PEQ overloading works differently because the DAC chip can't do anything about it even if the volume control is before the reconstruction filter.

Then, the information on the ES9219 must be incorrect. I will cross out relevant parts in my earlier posts.

 

[OHtaru]

Then, the information on the ES9219 must be incorrect. I will cross out relevant parts in my earlier posts.

I remember seeing somewhere that the ES9219 uses 2-Stage volume control, but I can't find where it is now.

 

[jkim]

Just to confirm, here is the FFT of the ISO test signal fed to DAC V1 with its volume set to -3 dB (in the Current mode of THD compensation):

Yup, not overloaded.

But to take advantage of analog attenuation's noise benefit, DAC V1 should use Voltage mode. And I believe Dmitry will update the DAC V1 firmware in such a way that it uses analog attenuation (24 dB range) combined with software volume control in DSP.

 

[OHtaru]

Back to the topic. ISP always occurs simultaneously with strong transients, and I think its audibility is questionable. Putting ISP aside, lossy compression will produce actual samples over 0dB. Except for Foobar2000 and some professional audio software, I rarely see other consumer players that can correctly handle samples over 0dB. Before I figured them out, I was still happily listening to songs and didn't notice anything wrong. Of course, the pursuit of rational perfection can still bring me a sense of satisfaction.

 

[fcracer]

Example songs would be nice .

I think you both are onto something , it can possibly be an actual issue with digital processing involved , with todays more complex products with EQ and DSP of all kind .

Thanks for mentioning an example product did not NAD had some buggy software in the past .

Soo maybe some test procedure should be used by Amir when testing products that do more than simply DA ? there seems to be opportunity to fudge things badly by software .

Lana Del Ray "For Free" song, at 1:10, there is a phrase, "Waiting for the walking greeeeeeeeeen".

 

[Donuts123]

Inter-sample overs are by no means limited to modern loudness-war casualty mastering. I checked several of my CDs in this thread.

I think you could group CDs into categories depending on how they were mastered:

• Normalised to 0 dBFS sample peak. By definition you will see ISOs there.
• Normalised to 0 dBFS true peak. Here the exact true peak may not match exactly, e.g. some tracks might show +0.04 dBFS true peak in foobar2000. But such tiny overs probably won't cause signifcant distortion.
• No overs at all.

 

[Donuts123]

Suppose you're editing/mixing/mastering a song. None of the individual raw instrument and vocal tracks you originally captured have any clipping or ISOs (unless the recording engineer was incompetent). You probably do all editing and mixing in 32-bit float so oodles of headroom there.

The final stage is when you render to 16-bit 44.1 kHz. To make the most of the 16-bit dynamic range, it actually makes sense in a way to normalise the output to 0 dBFS sample peak. There will of course be ISOs in that output file. If the true peak is +2 dBFS say, that's an extra 2 dB of dynamic range (and S/N ratio) vs normalising to 0 dBFS true peak.

The trouble is that a (probably large) proportion of consumer playback hardware and signal chain doesn't handle ISOs properly. So while your final 44.1/16 file could play back perfectly (and probably did when you played it in your DAW software), it probably won't be in the hands of a typical end user. Especially if it gets lossily encoded to MP3 or whatever where even more headroom is needed.

 

[markstein]

Suppose you're editing/mixing/mastering a song. None of the individual raw instrument and vocal tracks you originally captured have any clipping or ISOs (unless the recording engineer was incompetent). You probably do your all editing and mixing in 32-bit float so oodles of headroom there.

The final stage is when you render to 16-bit 44.1 kHz. To make the most of the 16-bit dynamic range, it actually makes sense in a way to normalise the output to 0 dBFS sample peak. There will of course be ISOs in that output file. If the true peak is +2 dBFS say, that's an extra 2 dB of dynamic range (and S/N ratio) vs normalising to 0 dBFS true peak.

The trouble is that a (probably large) proportion of consumer playback hardware and signal chain doesn't handle ISOs properly. So while your final 44.1/16 file could be played back perfectly (and probably did when you played it in your DAW software), it probably won't be in the hands of a typical end user. Especially if it gets lossily encoded to MP3 or whatever where even more headroom is needed.

It simply makes no sense to stupidly normalise the sample peak to 0 dBFS. There is so much dynamic range available, so please use it.

 

[antcollinet]

It simply makes no sense to stupidly normalise the sample peak to 0 dBFS. There is so much dynamic range available, so please use it.

Exactly - there is no music that needs the full dynamic range of even 16 bit PCM. Lowering the level by 3dB from sample peak would make no audible difference to the sound quality.

 

[skankhobag]

I know what they are - I know how they happen.

I've never knowingly heard distortion coming from intersample overs. When I first got my MiniDSP flex, I set the input gain to -3dB to avoid any risk. I recently set it back to 0dB and:

1 - Have not noticed any difference
2 - Have not ever heard anything I could attribute to intersample overs.

Is it not the case that only badly mastered music (basically clipping in any case) will cause them. And even then - if it only happens for an extremely brief period in the music it is going to be almost impossible to detect audibly in any case.

Yet there seems to be all sorts of FUD talked about it in conjunction with DACs. We even have a whole thread discussing how to test DACs for it. It feels to me like it is the "new jitter" Something that people can hang their hat of audble differences on - without it being an actual problem in reality.

What am I missing?

Has anyone else heard the effect of inter-sample overs - if so, which track, at what time? And what does it sound like?

A couple points:

• Inter-sample overs are nothing new. I remember reading about them 20+ years ago.

• If you can’t hear a difference between 0dB and -4dB, that doesn’t mean that your equipment won’t produce inter-sample overs. The material being played may not create them. They are mostly caused by mastering engineers over compressing the audio signal and running the peak output at 0db rather than -1 or -2dB (which is just plain stupid).

• Any good DAC today accounts for inter-sample overs and, since the subject is now passè and very wonky, no manufacturers really talk about it anymore in their sales lit.

 

[JSmith]

Any good DAC today accounts for inter-sample overs

You sure? Most DACs do not actively correct for inter-sample overs from my understanding.


JSmith

 

[jkim]

I believe, going forward, ANYONE wanting to discuss issues related to ISO/ISP should first have a look at Archimago's recent study on the topic. As usual, it is an amazing, comprehensive analysis!

And I agree to his conclusions:

My recommendation of +3dB discussed a few years back for DAC intersample overhead I think remains fair. This will easily cover the vast majority of music out there (including 96% of my Electronica tracks). However, if you're very much into electronica/techno/EDM, maybe a DAC with +4.25dB headroom would given you a bit more assurance - or simply attenuate your DAC digital volume by a few dBs, or even better, use ReplayGain with something like -18 LUFS target.

And more importantly,

I think the ultimate conclusion to all of this is simply this: instead of insisting that our DACs have high intersample overload headroom, it would have been much better if the music were created respecting 0dBFS as the limit. I would argue that music with high intersample peaks - let's say more than +1dB which is higher than 99% of my classical music collection is likely not great sounding "hi-fi" material anyways.

I would still like to have 3 dB headroom for peace of mind. IMO the best, device-side solution to ISO is software volume control (in DSP before signal sent to a DAC chip), like in RME ADI-2 series or Qudelix 5K.