1 Topic by switch6343 (edited by switch6343 2022-07-05 03:21:42)

  • From: Central Switzerland
  • Registered: 2022-02-23
  • Posts: 29

Topic: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

I carried out an experiment today. I changed the sampling rate in DSP of JRiver Media Center v29 from 44.1 to (2x 44.1) 88.2 kHz.

I noticed inmediately that the sound on my Sennheiser HD650 Headphone had much improved, voices more in the front, more present, voices of backing voices much more differentiated. It gave me the impression that I had been present during the recording in the studio.

Also, the echo in the recording (live recordings in a church of a monastry) was much more present. This discovery has been a real ear opener for me.

I did another test be increasing the sampling rate from 44.1 to (4 x 44.1) 176.4 kHz.  I just listened very briefly, very unsatisfactory, a bad experience. No comparison to the very good resolution at 88.2

I listened to these music sources: Roy Orbison "Black & White", CD1 of Ane Brun "It all starts with one", Bensound "The Jazz Piano" (nur Piano and Standing Bass), various albums of the Monks Choir of the Monastry Chevetogne (Belgium)

I am sharing my experience and would be very pleased to know of any other owner of an RME ADI-2 Pro (FS Black Edition) with same of similar upsampling experience.

Sorry, for initially posting in German language.

2 Reply by ramses (edited by ramses 2022-07-05 06:38:25)

  • ramses
  • DAW Mastermind
  • Online
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 13,182

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

Sorry, but upsampling doesn't make the sound better, it stays at the quality of the last A/D conversion, nothing is being enriched or improved.

What kind of D/A filter are you using? It would be conceivable that the treble roll-off of a D/A filter (e.g. NOS or other) was shifted into the inaudible range above ~20 kHz. This roll-off at single-speed (44.1/48 kHz) could also be compensated by PEQ settings, which have already been posted here in the forum, see e.g.: https://forum.rme-audio.de/viewtopic.ph … 26#p130026

Other possibilities:

1. different placement of the headphones on the ear. I know that the Sennheiser HD800 series has extremely large ear cups that are also not tight. There you never have a defined position with the the ear (which I find quite terrible). Perhaps it is similar or the same with your model.

2. Psychoacoustic effect ("You wished for it").

Kein Grund zur Entschuldigung. Es gibt auch einige (wenige) Postings auf Deutsch hier im Forum, aber Sie erreichen natürlich mehr Anwender, wenn Sie auf English posten.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub14
  • KaiS
  • DAW Mastermind
  • Offline
  • From: Germany
  • Registered: 2020-02-05
  • Posts: 3,523

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

switch6343 wrote:

...changed the sampling rate in DSP ... from 44.1 to (2x 44.1) 88.2 kHz.

I noticed inmediately that the sound on my Sennheiser HD650 Headphone had much improved, voices more in the front, more present, voices of backing voices much more differentiated.

... increasing the sampling rate from 44.1 to (4 x 44.1) 176.4 kHz.  ... very unsatisfactory, a bad experience. No comparison to the very good resolution at 88.2..

Interesting, never heard such significant difference here.
Listening to Tidal “Master” streaming, sample rate changes all the time, 88.2 happens quite often.

Even other situations where I use sample rate conversion never changed the sound in the way you described.

This leads to the idea that something is kind of “wrong” with JRiver’s sample rate conversion algorithm at 88.2 kHz.

This “wrong” could create extra harmonics, similar to something like a hot driven tube amplifier or Aural Exciter, a technique to brighten up recordings in the studio, beyond what an EQ can do.

KaiS's Website

4 Reply by KaiS (edited by KaiS 2022-07-05 08:27:59)

  • KaiS
  • DAW Mastermind
  • Offline
  • From: Germany
  • Registered: 2020-02-05
  • Posts: 3,523

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

ramses wrote:

1. different placement of the headphones on the ear. I know that the Sennheiser HD800 series has extremely large ear cups that are also not tight. There you never have a defined position with the the ear (which I find quite terrible). Perhaps it is similar or the same with your model.

HD 650 is very much better in this regard.

The huge positional dependency of Sennheiser HD 800’s sound is the result of internal reflections.
If you have one I can give you information on how to largely improve on that.

KaiS's Website

5 Reply by ramses (edited by ramses 2022-07-05 09:00:28)

  • ramses
  • DAW Mastermind
  • Online
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 13,182

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

KaiS wrote:
ramses wrote:

1. different placement of the headphones on the ear. I know that the Sennheiser HD800 series has extremely large ear cups that are also not tight. There you never have a defined position with the the ear (which I find quite terrible). Perhaps it is similar or the same with your model.

HD 650 is very much better in this regard.

The huge positional dependency of Sennheiser HD 800’s sound is the result of internal reflections.
If you have one I can give you information on how to largely improve on that.

Thanks Kai, very kind from you, but not needed. I only had one here for a test a couple of years ago.

BTW .. on the pictures of the HD650, which I found in Internet, it was difficult to see whether they also have such large ear cups like HD800.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub14
  • From: Central Switzerland
  • Registered: 2022-02-23
  • Posts: 29

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

Thank you Ramses for your feedback. Please note, that I have my RME ADI-2 PRO fs Black Edition run in default settings, I haven't any settings changed, plain vanilla. After my purchase I experimented with some settings, but I noticed that any change in settings did not improve my listening experience.

An important benchmark for me is when I listen: do I experience fatigue. I experienced fatigue during listening at 44.1 kHz sample rate. It was not as bad, that I stopped listening after a few minutes, I was able to listen for an hour or so. But any experience of fatigue has completely disappeared. The sound has become like satin.

In the mean time I started to compare. again I disabled the 88.2 settings, and listened to classical music at 44.1. Fatigue was the result. I listened to concertos of Hummel, Albinoni, Bizet. Changed the setting to 88.2 and the music becomes relaxed, I don't get nervous while listening. May be I am oversensitiv. I have very well trained hearing and I am used to listen to music at low volume level (-20 dB, - 30dB), depending on the recording. And I have been in live concertos for almost all of my life, starting in Rotterdam, subsequently at the KKL in Luzern, Concerthall of the LAC in Lugano, Theatre sociale (Opernhaus) in Como, many many pop/rock/jazz concerts in Zürich.

My experience is to not experience with filters at all, I have had my bad experiences with Room Settings and calibration with special calibration software and microphones. Until I found out that setting my (high end) amp to bypass resulted in taking away a thick curtain and fatigue immediately disappeared. I was able to listen only a few minutes, and the fatige was unbearable. But that is another story. I also used once a VST filter in JRiver Media Center. I removed it, and the sound was back as I wanted. That specific VST filter was meant to flatten the frequence curve of my Headphone. I don't listen to music with presumed logic (a straight frequency curve, based on measurements, is). Plain vanilla, default settings. But the very postive change in listening experience of the sampling rate to 88.2 surprised me big, I can assure you.

So, to conclude: the experience of fatigue or the lack of it is for me the ultimate benchmark. Instinctively I think that phase differences in the frequency range from 20 Hz - 20 kHz (and may be even above) is the cultprit of causing fatigue. But I am unable to prove, I don't know if that can be measured.

My listening experience is not based on science, its my subjective personal experience. I just can't explain WHY my listening experience is better at a sampling rate of 88.2 kHz. I am not intending to try to get it explained. I am just curious if other persons have had the same experience as I have made.

My twin brother, who also is the proud owner of an RME ADI-2 PRO, has made the same experience now as I have. He now also changed 44.1 to 88.2 in JRMC.

I don't think it's imagination on my part, but for sure, I am not able to explain.

7 Reply by ramses (edited by ramses 2022-07-05 14:30:57)

  • ramses
  • DAW Mastermind
  • Online
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 13,182

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

You could try the D/A filter "Slow", see also KaiS's evaluations:

https://forum.rme-audio.de/viewtopic.ph … 41#p186341
https://forum.rme-audio.de/viewtopic.ph … 58#p178258
https://forum.rme-audio.de/viewtopic.ph … 68#p177168

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub14

8 Reply by switch6343 (edited by switch6343 2022-07-06 00:53:16)

  • From: Central Switzerland
  • Registered: 2022-02-23
  • Posts: 29

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

excellent ramses, I read the postings of your links. I'll try that out with the "slow" filter, as mentioned in the posting of KaiS. I will revert back to you, but that will be sometime during the next weekend.

And KaiS, a great many thanks for your feedback as well. Much appreciate.

I had extensive tests with a Beethovern concerto by Josef Suk, two tracks only. I switched about 5 times between 44.1 and 88.2. Of course it was not a blind test, but @88.2 is a clear winner.

I have not doubt that the listening difference is NOT due to the processing of the RME DAC. I guess JRMC doesn't play well with RME DAC. May be the processing in JRMC of 44.1 is not well calibrated, but I cannot speculate what the root cause is or could be. I have a short and very good (Oehlbach) USB cable (3 x isolated) between my HP NOtebook and my RME DAC, so this can be excluded as a possible cause.

In any case, I consider this as a nice journey to try to find out what the cause could be.

On another note. I discovered a strange error message popping up in JRiver Media Center. I wanted to listen to Josef Suk's Beethoven concerto, but unfortunaely I hadn't yet turned on my RME DAC.

This was the error message:
Quote
Something went wrong with playback.
Details:
Playbank could not be started on the output "ASIO" using the format "88.2 kHz 2ch".
The ASIO device "ASIO MADIface USB" does not support the sample rate of 48000 Hz.
Unquote

Should I report this message to technical support of RME? What seems strange to me is the last phrase "The ASIO device "ASIO MADIface USB" does not support the sample rate of 48000 Hz.

Please note that I have the latest ASIO MADIface USB driver installed, issued by RME in December 2021.

In fact I wanted to play the Beethoven album @88.2 instead of @44.1 So, the 48000 Hz seems to be odd to me as additional information in the message.

pls advise.

9 Reply by MC

  • MC
  • MC
  • Administrator
  • Offline
  • From: Germany/Thailand
  • Registered: 2007-04-06
  • Posts: 25,582

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

That error mesaage has no further meaning. You tried to start playback with a missing driver, so whatever happened is invalid.

Regards
Matthias Carstens
RME

MC's Website

10 Reply by ramses (edited by ramses 2022-07-06 08:08:51)

  • ramses
  • DAW Mastermind
  • Online
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 13,182

Re: Upsampling in DSP JRiver Media Center from 44.1 to 88.2 GREAT SOUND

switch6343 wrote:

[...] I have not doubt that the listening difference is NOT due to the processing of the RME DAC. I guess JRMC doesn't play well with RME DAC. May be the processing in JRMC of 44.1 is not well calibrated, but I cannot speculate what the root cause is or could be. I have a short and very good (Oehlbach) USB cable (3 x isolated) between my HP NOtebook and my RME DAC, so this can be excluded as a possible cause. [...]

What you mean by calibration? Where should calibration be needed and cablibration of what ?

Oehlbach USB cable? Regarding cables. I am also not buying the cheapest to get also good plugs but Lindy Premium cables are fully sufficient. USB2 cables can be up to 5m long according to standards and even at that length it wouldn't produce any other sound. This is digital transfer of (audio) data. The cable has no influence on the sound.

I'm not clear right now if you understand how the components work end to end when audio is played. Could it be that you are still very much in the analog world? The audio player opens a driver, ideally the RME ASIO driver. Then the only objective is to send the stream of digital data unaltered through the ASIO driver to the device. The ASIO driver is able to communicate with the Windows kernel to be able to send data through USB to the ADI-2 Pro/DAC on the shortest way without having to go through other Windows audio subsystems. The same lossless transfer of (audio-) data you can get e.g. by using Windows drivers like WASAPI, best be used in exclusive mode.

And with this fantastic unit you can check the lossless transfer of audio data yourself in an "end-to-end" fashion from audio player to the DSP of the ADI-2 DAC/Pro. This is the so-called "Bittest". These are small test files (.wav format) which you can download for free (for each sample rate there is one single file) and playback on your player. These files are short, brought to the point. There is no lengthy transfer needed. You can also use the repeat function of your audio player if you want.

The purpose is to check "end to end" whether audio is played back completely lossless. The positive test result is shown on the display of ADI-2 DAC/Pro if all was ok, lossless transfer of (audio) data. This includes the complete digital path from the audio player to the DSP of the ADI-2 DAC/Pro. After that the D/A conversion takes place.

This way you can validate lossless audio transfer "end to end" and more than reliable transfer of digital data is not required. No expensive cable can make this transfer any better. And even the cheapest 5m USB cables, according to standards, needs to be able to transfer data (including audio data) without any error. Otherwise IT would be a mess if you would not be able to make reliable backups of Operating System and user data!

If the Bittest fails different reasons are possible
a) volume not set to 100% on the audio player (the application)
b) not the correct driver chosen, Windows Mixer and audio features in the OS could also change audio data
c) USB transport errors

So if the bit test fails you can (latest) install the RME ASIO driver and keep it open while using the ADI-2 DAC/Pro.
As long as this driver settings Window is open CRC checks are being performed by the RME ASIO driver.
If you have a transport error over USB, you know that the issue is between computer and ADI-2 DAC/Pro.
Maybe faulty cable.

The advantages of ADI-2 DAC and Pro in this regard I have also presented here in this blog article, by the way, can also be read in the manual. https://www.tonstudio-forum.de/blog/ent … ses-en-de/

So there is nothing to "calibrate", no idea what you think, this is a normal transfer of digital data from a player over digital connections, which all work lossless. Any jitter is prevented by SteadyClock FS technology and the final D/A conversion happens on the ADI-2 DAC/Pro with the built-in D/A converter using its own Femto Second Clock. Behind the DAC there are analog amplifier stages. Here you then have additional quality by the four reference levels design in the analog domain together with the feature autoreflevel. Autoreflevel in combination with the four reference levels ensure high SNR and dynamics over a wieder range of volume settings on the ADI-2 DAC/Pro itself.

Using the Slow D/A filter changes nuances of the D/A conversion in the D/A converter chip. I think here you can fine tune a little and I am curious whether you can hear the change and whether you like it. Please report back.

I hope this little overview was informative / interesting for you. Have a nice day and see you back on forum wink

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub14