Jitter problem

[Abtr]

Abtr, I thinked you missed John's point... the question is whether the laptop has bad USB hardware. What known good source have you tried that shows the problem follows the DAC vs the laptop??

Hi Dave. You're correct. John suggested that the problem might be a faulty USB 2.0 implementation in my laptop (HP Probook 450 G2). But I find it strange that there is no change at all in the jitter level with - or without WASAPI. That points in the direction of the DAC/driver as the problem cause. We'll see when I receive the Bluekey USB converter and driver..

[Abtr]

I received the Firestone Bluekey converter today and connected it to the coaxial input of the MX-DAC and it doesn't make a difference. The sound is still not very good and the jitter level is unchanged. So apparently the DAC's USB interface is not the problem (although I noticed that the Bluekey device uses the same Windows USB 2.0 driver as does the MX-DAC).

 

I could try another PC/laptop and Windows 10, but I suspect there's no way of knowing beforehand that it doesn't have similar USB 2.0 problems. Maybe a simple desktop PC and a soundcard with coaxial digital output is the way to go..

[Abtr]

An update. This PC USB jitter phenomenon proves to be very persistent. By now I've tried different (Windows) PCs and OSs (including Windows 10), different USB cables, the Intona USB isolator, and a good USB to SPDIF converter (Audio-gd DI U8). Although some improvements can be perceived, all possible setups produce the same excessive amount of jitter as measured and heard in the analogue audible domain. USB streaming from a PC really audibly removes the shine from the music; it sounds dull, flat and lifeless, as compared to the detail, stage and presence you get from a good optical 16/44.1 CD transport.

 

It's clear to me that the DAC's USB interface isn't the problem. Instead, the PC's USB output must be the problem. Somehow it already 'contains' jitter, most probably resulting from some internal sound processing that is still going on even in WASAPI direct mode. Bit perfect data transfer can't fix a faulty source signal. A telltale sign may be that Windows audio mixer can still be used to control the WASAPI output level. I'm now seriously considering to try a MAC or Linux OS..

[Possum Jenkins]

An update. This PC USB jitter phenomenon proves to be very persistent. By now I've tried different (Windows) PCs and OSs (including Windows 10), different USB cables, the Intona USB isolator, and a good USB to SPDIF converter (Audio-gd DI U8). Although some improvements can be perceived, all possible setups produce the same excessive amount of jitter as measured and heard in the analogue audible domain. USB streaming from a PC really audibly removes the shine from the music; it sounds dull, flat and lifeless, as compared to the detail, stage and presence you get from a good optical 16/44.1 CD transport.

 

It's clear to me that the DAC's USB interface isn't the problem. Instead, the PC's USB output must be the problem. Somehow it already 'contains' jitter, most probably resulting from some internal sound processing that is still going on even in WASAPI direct mode. Bit perfect data transfer can't fix a faulty source signal. A telltale sign may be that Windows audio mixer can still be used to control the WASAPI output level. I'm now seriously considering to try a MAC or Linux OS..

 

Why don't you disable the Windows audio service completely? The mx-dac you have has it's own driver and doesn't require any sort of Windows audio services/sound devices, that's how I use it on Windows

[Abtr]

Why don't you disable the Windows audio service completely? The mx-dac you have has it's own driver and doesn't require any sort of Windows audio services/sound devices, that's how I use it on Windows

I've tried that. I disabled Windows audio in the msconfig -> services tab. After that, only Foobar's ASIO: XMOS USB Audio 2.0 ST 3033 driver works, but the PC's USB still produces the same amount of jitter.

[ecwl]

Are you 100% certain that the CD test track was ripped lossless perfectly to a WAV or FLAC file? That kind of frequency response changes are much more common with lossy MP3 rips...

[Abtr]

Are you 100% certain that the CD test track was ripped lossless perfectly to a WAV or FLAC file? That kind of frequency response changes are much more common with lossy MP3 rips...

Yep, I downloaded the (WAV) test tracks from the Dayton Audio website and burned them on a CD. I subsequently played them via an optical CD transport and there's no jitter. But there's a massive amount of jitter when I play the original files via Foobar2000 and USB..

[ecwl]

I just realized the frequency response curve you displayed also can look like de-emphasis was applied to the track. As to what's really going on, I have no answers.

[Abtr]

The frequency response curve for the noise test signal looks as if at some point in the audio chain the USB signal is processed and reconstructed based on the timing of the computer clock which can be very jittery. It doesn't appear to be software that is doing this. Possibly some cost effective shortcut in the integrated audio hardware uses the analogue output of the computer DAC to resample and reconstruct the digital signal, including the analogue result of jitter from DA conversion. At least that's how it sounds. And there's no way any reclocking device can remove this jitter because it has become 'encoded' in the digital signal.

 

I didn't yet try a modern desktop PC (with separate sound card), so the problem may be limited to laptops/notebooks with very compact circuit designs. I wonder whether an external USB to SPDIF sound card would bypass any DA-AD conversion or other possible audio processing going on in the PC's integrated hardware (PC USB output -> External sound card USB to SPDIF -> DAC). A USB sound card might receive the same degraded signal to begin with, but I'm not sure..

[sandyk]

At least that's how it sounds. And there's no way any reclocking device can remove this jitter because it has become 'encoded' in the digital signal.

I am finding similar too, when saving to a USB memory stick plugged directly into a rear motherboard USB port. However, if I connect the Regen in line with the USB memory stick when saving to it, and play back using the Regen in line too, that the version with the USB Regen in line sounds markedly better than the same without the Regen in line originally. The direct version does give the appearance of Jitter having become "encoded" in the digital signal as you are suggesting.

[Abtr]

I am finding similar too, when saving to a USB memory stick plugged directly into a rear motherboard USB port. However, if I connect the Regen in line with the USB memory stick when saving to it, and play back using the Regen in line too, that the version with the USB Regen in line sounds markedly better than the same without the Regen in line originally. The direct version does give the appearance of Jitter having become "encoded" in the digital signal as you are suggesting.

The only way I see jitter becoming 'encoded' in the digital/USB signal is that (for some reason) all digital audio first goes through the computer DAC before it's sent to analogue output and/or to AD conversion for digital/USB output. Copying or reading a file to or from a USB stick only involves the OS file system and must be absolutely bit perfect, otherwise (e.g.) copied program files would be corrupted and consequently fail to run. So any perceived difference between audio files being copied and read to and from a USB stick with or without the Regen in line must be imaginary. Maybe you can do some measurements..

[mansr]

The frequency response curve for the noise test signal looks as if at some point in the audio chain the USB signal is processed and reconstructed based on the timing of the computer clock which can be very jittery. It doesn't appear to be software that is doing this. Possibly some cost effective shortcut in the integrated audio hardware uses the analogue output of the computer DAC to resample and reconstruct the digital signal, including the analogue result of jitter from DA conversion. At least that's how it sounds.

 

That's not how any PC hardware works. It certainly wouldn't be cost effective.

[Norton]

Yep, I downloaded the (WAV) test tracks from the Dayton Audio website and burned them on a CD. I subsequently played them via an optical CD transport and there's no jitter. But there's a massive amount of jitter when I play the original files via Foobar2000 and USB..

 

I was going to ask what happens if you use the notebook as the physical cd transport, still connected to DAC via USB? Is that what you mean here? If so that is interesting.

[mansr]

Yep, I downloaded the (WAV) test tracks from the Dayton Audio website and burned them on a CD. I subsequently played them via an optical CD transport and there's no jitter. But there's a massive amount of jitter when I play the original files via Foobar2000 and USB..

Are those files available for anyone to download?

[sandyk]

The only way I see jitter becoming 'encoded' in the digital/USB signal is that (for some reason) all digital audio first goes through the computer DAC before it's sent to analogue output and/or to AD conversion for digital/USB output. Copying or reading a file to or from a USB stick only involves the OS file system and must be absolutely bit perfect, otherwise (e.g.) copied program files would be corrupted and consequently fail to run. So any perceived difference between audio files being copied and read to and from a USB stick with or without the Regen in line must be imaginary. Maybe you can do some measurements..

 

No, it's not imaginary!

I demonstrated this to a friend here today, who is also a C.A. member, via both my DIY DAC and his Bricasti M1 DAC. The files were stored on a Corsair Voyager USB memory stick which was plugged into a front USB port of the Oppo 103, which was used as a Transport, using an Uptone USB Regen in line powered by a regulated +9V supply derived from a 12V Li Ion battery.

We also listened to a recently burned CD-R from the Oppo 103 with the same worse sounding files burned to it, as well as the same .wav files saved directly to an internal SSD powered via a 12V to 5V regulator followed by a John Linsley Hood designed shunt regulator.

Even though the file numbers were in different sequences , but in pairs on the same CD, my friend correctly identified the different versions via both DACs.

The pairs of .wav files on the CD-R have the same .md5 checksums after being ripped back to HDD again.

 

If you wish to discuss this further with me, please do it via a PM as I have zero interest in arguing with the usual suspects about this.

[Abtr]

That's not how any PC hardware works. It certainly wouldn't be cost effective.

I would think so too. But as I said, that’s definitely how it sounds (and measures) and reclocking and/or USB isolation makes no difference.

 

Are those files available for anyone to download?

You can download them here: http://www.daytonaudio.com/Omni_Mic_Test_Tracks.zip They're also available on CD. Track 1 is the (pseudo) noise track and track 2 is the sine sweep. To use them you need the Omnimic V2 software and a calibrated microphone. With the noise track, basically FFT is performed on the played back noise.

 

I was going to ask what happens if you use the notebook as the physical cd transport, still connected to DAC via USB? Is that what you mean here? If so that is interesting.

I get the exact same result if I play the test tracks from the PC's optical drive (degraded sound).

[davide256]

An update. This PC USB jitter phenomenon proves to be very persistent. By now I've tried different (Windows) PCs and OSs (including Windows 10), different USB cables, the Intona USB isolator, and a good USB to SPDIF converter (Audio-gd DI U8). Although some improvements can be perceived, all possible setups produce the same excessive amount of jitter as measured and heard in the analogue audible domain. USB streaming from a PC really audibly removes the shine from the music; it sounds dull, flat and lifeless, as compared to the detail, stage and presence you get from a good optical 16/44.1 CD transport.

 

It's clear to me that the DAC's USB interface isn't the problem. Instead, the PC's USB output must be the problem. Somehow it already 'contains' jitter, most probably resulting from some internal sound processing that is still going on even in WASAPI direct mode. Bit perfect data transfer can't fix a faulty source signal. A telltale sign may be that Windows audio mixer can still be used to control the WASAPI output level. I'm now seriously considering to try a MAC or Linux OS..

 

This behavior parallels exactly why I gave up on trying to use a windows PC as source feeding a DAC and went to purpose built streamers feeding a DAC... better results with a device whose hardware was chosen with audio as the priority. Mac seems to be the only device where personal computer users are consistently happy with off the shelf digital audio to DAC

[ecwl]

I think abtr has two problems. One is probably jitter. Which is probably why music sounds better off the Toslink from a CD player. But the other problem is that at least with the test file, it is played differently between the PC and the CD player optical output because there is no way jitter can cause such a dramatic shift in frequency response. Otherwise, everyone else would measure hitter the way abtr did it. The question then is, when abtr plays music from his PC, is every file's frequency response messed up or just the test track. No idea. Good luck figuring things out.

[Abtr]

I think abtr has two problems. One is probably jitter. Which is probably why music sounds better off the Toslink from a CD player. But the other problem is that at least with the test file, it is played differently between the PC and the CD player optical output because there is no way jitter can cause such a dramatic shift in frequency response. Otherwise, everyone else would measure hitter the way abtr did it. The question then is, when abtr plays music from his PC, is every file's frequency response messed up or just the test track. No idea. Good luck figuring things out.

I played a CD with the test tracks from the computer's optical drive over WASAPI and USB to the DAC. That should implement bit perfect data transfer. With good reclocking, that should give a jitter free analogue signal. Currently I have the Intona Isolator and the Audio-gd USB to SPDIF converter to do the reclocking. But neither device makes any difference in the amount of degradation (jitter) of the noise test track played from the PC relative to the same noise test track played from a good CD transport. The latter showing no frequency degradation at all (relative to a pure sine sweep which is much less susceptible to jitter).

 

A white noise signal contains every frequency in approximately the same amount (and e.g. pink noise contains every frequency in an amount proportional to 1/f). If a noise signal is played back, then all frequencies should be present in the analogue result in the original amount. A so called Fast Fourier Transform (FFT) algorithm can decompose the analogue signal into all the frequencies it contains and their relative magnitude. This shows that playing back a noise signal from a PC leads to severe degradation of the magnitude of mid and high frequencies. You can easily hear the difference in pitch in an A/B comparison between a PC and a decent CD transport playing back the same noise signal.

 

Jitter = timing errors and in music this leads to phase shifts and cancellation of especially higher frequency components, resulting in a lack of detail and stage or presence. I've used this method to assess the jitter levels of different redbook CD players (which were generally much lower then the PC's jitter level) and it always provided a confirmation of the level of ease/unease/irritation I experienced with a player. Although noise exaggerates the effect of jitter in the analogue domain, it can definitely be heard in regular music.

 

So where does this PC jitter come from? Since it can’t be removed from a bit perfect (WASAPI) stream by reclocking, it must somehow have been encoded in the digital signal by the PC. This could be some DA-AD conversion stage in the PC but it can also be a bad/cheap hardware implementation of the system sound level control, which is apparently always functioning in the sound chain and accessible (e.g., via Windows mixer) even in WASAPI direct modus..

[Abtr]

Another update. I tried a brand new Apple Mac Mini. I played the WAV test tracks through the native iTunes player as well as through a Decibel player, using the USB 2.0 output and the build in optical (Toslink) output of the Mac. Both media players are supposed to provide bit perfect digital audio, but I got the same disappointing results: massive jitter and degraded sound.

 

I'm aware of reviews stating that a good audio setup on a PC or Mac will put even the best redbook CD players to shame and that, e.g., Archimago's measurements indicate that just about any laptop/notebook computer can produce bit perfect digital audio (see: Archimago's Musings: MEASUREMENTS: Part I: Bit-Perfect Audiophile Music Players (Windows). and Archimago's Musings: MEASUREMENTS: Bit-Perfect Audiophile Music Players (Mac OS X).). Personally I must conclude however that, based on my own measurements and listening, this is far from the truth and I will inform Archimago of my own experience with computer audio versus playing from regular optical CD transports.

 

I must say that I'm more than surprised about the total lack of reports on the internet recognizing this computer audio problem. I tried anything I could (except dedicated sound cards and/or dedicated digital sound streamers) to get a decent sound from a computer, but it seems impossible. I tried different bit-depths and sample-frequencies, but there remains a whole audio world missing as compared to the lowly 16/44.1 CD. The question is, what's going on here?

 

Does anyone in the computer audio community ever do some comparative A/B listening with the same music played back from a computer versus a CD transport? If so, then what transport, DAC, amp, et cetera are you using? For example, my Musical Fidelity MX-DAC can sound absolutely gorgeous using optical CD input, but it sounds like crap using computer input (USB or optical) and the same music. I can personally only tolerate electronic music form a computer for longer periods. All else quickly becomes unbearable. Anyone with reasonable hearing should be able to discern the difference..

[One and a half]

Another update. I tried a brand new Apple Mac Mini. I played the WAV test tracks through the native iTunes player as well as through a Decibel player, using the USB 2.0 output and the build in optical (Toslink) output of the Mac. Both media players are supposed to provide bit perfect digital audio, but I got the same disappointing results: massive jitter and degraded sound.

 

If the Mac Mini was purchased yesterday, the USB ports are type 3.0, therein may lay the problem. Your chain from the computer recommended be : Mac Mini USB 3.0 -> USB2.0 hub [unpowered] (can also be a regen or RUR) -> Intona -> DAC.

The OSX version should be Yosemite or earlier (but not Lion). El Crapitan has issues with USB.

 

The MF MX-DAC could use a good, well grounded linear power supply. I wonder how the 0V on the PSU is determined. If the AC source on the PSU has two prongs, there could be elevated AC hum on the ground connections creating havoc. It's sometimes easy to determine elevated AC voltages, in a darkened room, make the USB connection to the DAC and see if there are small sparks at the shell. Other than that a scope or VERY good digital meter can measure.

[Jud]

I haven't read the entire thread, but with the level of problem you're experiencing, I would suspect a hardware fault in the DAC, since this continues to happen with several different sources. Try a different DAC (ideally another of the same model, but if not, just any other DAC you can).

[Abtr]

If the Mac Mini was purchased yesterday, the USB ports are type 3.0, therein may lay the problem. Your chain from the computer recommended be : Mac Mini USB 3.0 -> USB2.0 hub [unpowered] (can also be a regen or RUR) -> Intona -> DAC.

The OSX version should be Yosemite or earlier (but not Lion). El Crapitan has issues with USB.

USB 3.0 is fully compatible with USB 2.0. The only difference is the transfer rate (USB 3.0 is about ten times faster than 2.0). A USB 2.0 device will happily operate at USB 2.0 speed when connected to a USB 3.0 port, and USB 2.0 driver software functions perfectly well with 3.0 ports. And I have the El Capitan OS X. I've never heard of any USB issues. Anyway, I used the Toslink output of the Mac so any USB issue is irrelevant.

 

The MF MX-DAC could use a good, well grounded linear power supply. I wonder how the 0V on the PSU is determined. If the AC source on the PSU has two prongs, there could be elevated AC hum on the ground connections creating havoc. It's sometimes easy to determine elevated AC voltages, in a darkened room, make the USB connection to the DAC and see if there are small sparks at the shell. Other than that a scope or VERY good digital meter can measure.

The MX-DAC sounds brilliant with a CD transport connected to its Toslink input (it has two actually). But the Mac mini connected to the same Toslink input, sounds not so brilliant to say the least. So the DAC or its power supply (SBooster linear grounded PSU), or for that matter any component after it, cannot be the problem here.

 

I haven't read the entire thread, but with the level of problem you're experiencing, I would suspect a hardware fault in the DAC, since this continues to happen with several different sources. Try a different DAC (ideally another of the same model, but if not, just any other DAC you can).

I've tried a different DAC with exactly the same (measurable) result. And as I explained above, the DAC cannot be the problem. The problem happens only and consistently with audio fed to the DAC by a computer, be it Windows or Mac, connected through USB or Toslink. The problem consistently does not happen with audio from a CD transport connected to the same DAC through coax or Toslink. This leads to the logical conclusion that the problem is located somewhere in the computer software or hardware. Somehow what should be a bit perfect stream is changed into something less perfect..

[sandyk]

USB 3.0 is fully compatible with USB 2.0. The only difference is the transfer rate (USB 3.0 is about ten times faster than 2.0). A USB 2.0 device will happily operate at USB 2.0 speed when connected to a USB 3.0 port, and USB 2.0 driver software functions perfectly well with 3.0 ports. And I have the El Capitan OS X. I've never heard of any USB issues. Anyway, I used the Toslink output of the Mac so any USB issue is irrelevant.

The MX-DAC sounds brilliant with a CD transport connected to its Toslink input (it has two actually). But the Mac mini connected to the same Toslink input, sounds not so brilliant to say the least. So the DAC or its power supply (SBooster linear grounded PSU), or for that matter any component after it, cannot be the problem here.

 

I've tried a different DAC with exactly the same (measurable) result. And as I explained above, the DAC cannot be the problem. The problem happens only and consistently with audio fed to the DAC by a computer, be it Windows or Mac, connected through USB or Toslink. The problem consistently does not happen with audio from a CD transport connected to the same DAC through coax or Toslink. This leads to the conclusion that the problem is located somewhere in the computer software or hardware. Somehow what should be a bit perfect stream is changed into something less perfect..

 

It would be interesting to see if somebody else with a Mac Mini could try and replicate your results.

I doubt that there would be anywhere near such a problem with a Mac Mini that uses a good Linear PSU such as the John Swenson designed JS2, especially if Kelvin Sensing was enabled.

[Abtr]

It would be interesting to see if somebody else with a Mac Mini could try and replicate your results.

I doubt that there would be anywhere near such a problem with a Mac Mini that uses a good Linear PSU such as the John Swenson designed JS2, especially if Kelvin Sensing was enabled.

Yes, that would be very interesting. And that's exactly what I'm concerned about. If a better power supply to the Mac would improve the quality of its digital audio output, then that would conclusively prove that (even in so called bit perfect modus) some sort of DA/AD conversion is going on, such as (zero) sound level adjustment..