Harsh piano recording of Clair de Lune

[RDoc]

I'm a fan of Debussy, but have found several recordings of his piano music hard to listen to for very long.
I'm hearing a "harshness" in some of the notes on some recordings, specifically around 0:36 - 1:00 and on in Clair de Lune from the Thibaudet "Arabesque Debussy Piano Works" album, but not in the Seong-Jin Cho "Debussy" album. The difference is in the sound of the top notes in the repeating phrases with the Thibaudet recording notes sounding distorted and harsh, whereas the Seong-Jin recording sounds musical, sharp but not harsh. If it means anything, I'd compare it to a photograph that was over sharpened.
I hear it both through KH 170ii speakers (Denon 3700H used as preamp with Wiim as source), and also through WH 1000X M3 earphones connected with LDAC from a PC. I'm using Tidal HiFi with "high" quality recordings. It's present when played at moderately high but not deafening levels but not at lower levels.
This seems to be an issue only with a few solo piano pieces, I never hear anything like it on any other music.
Any thoughts?

 

[mhardy6647]

1) Do you have any degree of tinnitus?
2) Whether you do or not -- you're probably dealing with "bad" (relative to your expectation/taste/enjoyment/tolerance) recording(s).

 

[majingotan]

Is it this album? Zero issues with the piano recording from Qobuz through my Susvara (no DSP, no EQ). Not a recording issue for me at least. The above recording to me have the mics a bit further from the piano and therefore have room reflections that are clearly captured in the recording.

Seong-Jin Cho's recording is definitely a lot cleaner, mics are placed very close to the string and are clearly evident from the recording. Also, there are less room reflections and more fundamental note emphasis on this recording

I'm leaning towards your physical condition of either fatigue or hyperacusis. Both recordings are PERFECTLY mastered for me at least.

 

[RobL]

The first one triggers my tinnitus…not sure why, the high notes almost sound discordant.
I can tolerate this version much better than either of those:

 

[RDoc]

I'm going with the tinnitus explanation. It's not obvious to me that it's a particular trigger, but I do have it and that makes sense. Unfortunately it really does affect my enjoyment of the Thibaudet album, and I like him a lot.

 

[mhardy6647]

I'm going with the tinnitus explanation. It's not obvious to me that it's a particular trigger, but I do have it and that makes sense. Unfortunately it really does affect my enjoyment of the Thibaudet album, and I like him a lot.

Sorry to click "like" but as a fairly recent member of Club Tinnitus I, at least, empathize. I find certain sounds on certain recordings 'ring' (for lack of a better word) when they didn't before. Aggravating as all get out.
You have my sympathy and empathy.

 

[majingotan]

I'm going with the tinnitus explanation. It's not obvious to me that it's a particular trigger, but I do have it and that makes sense. Unfortunately it really does affect my enjoyment of the Thibaudet album, and I like him a lot.

I hope you get better from your tinnitus. Thanks for the album recommendation! Now that it's getting late night here, I'm thoroughly engaged with the musical performance while writing this

Wow goosebumps with the Arabesque Debussy Piano Works by Thibaudet . Beautiful recording, perfect piano rendition indeed!

 

[JIW]

Using Qobuz on macOS, on the Thibaudet which is quite loud there is something that sounds like clipping. Like a fuzzy rustling kind of sound on the attacks. Even on the first few notes. For reference, volume on the second one is set 5 dB higher but it is no louder. The second also has some similar noises but not to the same degree.

Try the version by Alice Sara Ott from Nightfall. That one sounds fairly clean (24/96 on Qobuz).

Listening to the other two made me think my DAC or my HD600s may be broken.

 

[majingotan]

Using Qobuz on macOS, on the Thibaudet which is quite loud there is something that sounds like clipping. Like a fuzzy rustling kind of sound on the attacks.

Are we listening to the same Thibaudet album? Mine has ZERO clipping, ZERO fuzzy rustling with Bit Perfect lossless delivery from Laptop to DAP USB out to DAC (No EQ, NO DSP no digital signal manipulation)

 

[JIW]

On the Thibaudet it is still there at low levels. It is just hard to hear.

 

[JIW]

Are we listening to the same Thibaudet album? Mine has ZERO clipping, ZERO fuzzy rustling with Bit Perfect lossless delivery from Laptop to DAP USB out to DAC (bitperfect lossless)

View attachment 338669

Yes in 16/44.1.

Listen at 0:40 or thereabout. Definitely fuzzy. Not aggressive though. It is easier to hear at a higher volume. But even at a volume corresponding to an actual performance it is audible.

Same on YouTube.

I use Qobuz on macOS 13.2.1 on a MacBook Air M1 2020. Output is set in Qobuz to SMSL DL200 over USB and system output is set to another device. Volume in Qobuz is set to full. Core Audio is bit perfect for a single program. Connection is bit perfect, therefore. Headphones are Sennheiser HD600 with the new, revised pads. Reconstruction filter is number 3, the brickwall, linear phase, full attenuation at Nyquist. Filter 4, fast linear with slightly greater bandwidth but hardly any attenuation at Nyquist, gives the same though.

Edit: I also tried with AKG K371s and a Sabaj Da2 DAC/Headphone amplifier as well as the MacBook's analog output. Same result.

 

[majingotan]

I have two setups that I can try and see if the same scenario happens in multiple cases

First setup is the following:
I tried multiple ways to feed my DAC to see if there's gonna be any difference. My DAC is Schiit Yggdrasil+ More Is Better (118 dB SINAD) is on default out of the box configuration (oversampling enabled). Headphones are Hifiman Susvara driven by ampsandsound Forge tube amp. As expected, there's no difference between the following:

1) Built-in YouTube app from my DAP to Schiit Yggdrasil+ More Is Better via USB out. No fuzziness or any issues on the chorus or fast paced sections. No clipping.

2) Hooked up the USB cable from my laptop directly to DAC and use Roon. Again, No fuzziness or any issues on the chorus or fast paced sections. No clipping.

3) YouTube from laptop to Yggdrasil+ MIB using ASIO4ALL. Again, No fuzziness or any issues on the chorus or fast paced sections. No clipping.

Second setup is my portable setup, Campfire Andromeda IEM driven directly with the SP2000 DAP (no tubes in the chain, no EQ or any DSP):
The SP2000 DAC filter is default AKM4499EQ (same DAC chip as Topping D90) minimum phase apodizing filter (I think this is either Mode 1 or Mode 5 on the Topping D90)

No fuzziness or any issues on the chorus or fast paced sections. No clipping whatsoever as well. Based on my personal experience with this exercise, I speculate that I don't get any discomfort with this album that some others can perceive as tinnitus inducing or clipping. Heck I even got goosebumps earlier on how amazing this album sounded to me. One can download the track and load it up to Audacity/Adobe Audition and see if there's a loudness wars issue with the track.

 

[JIW]

I tried it again.

I have two setups that I can try and see if the same scenario happens in multiple cases

First setup is the following:
I tried multiple ways to feed my DAC to see if there's gonna be any difference. My DAC is Schiit Yggdrasil+ More Is Better (118 dB SINAD) is on default out of the box configuration (oversampling enabled). Headphones are Hifiman Susvara driven by ampsandsound Forge tube amp. As expected, there's no difference between the following:

1) Built-in YouTube app from my DAP to Schiit Yggdrasil+ More Is Better via USB out. No fuzziness or any issues on the chorus or fast paced sections. No clipping.

View attachment 338682

2) Hooked up the USB cable from my laptop directly to DAC and use Roon. Again, No fuzziness or any issues on the chorus or fast paced sections. No clipping.

View attachment 338690

3) YouTube from laptop to Yggdrasil+ MIB using ASIO4ALL. Again, No fuzziness or any issues on the chorus or fast paced sections. No clipping.

View attachment 338710

Second setup is my portable setup, Campfire Andromeda IEM driven directly with the SP2000 DAP (no tubes in the chain, no EQ or any DSP):
The SP2000 DAC filter is default AKM4499EQ (same DAC chip as Topping D90) minimum phase apodizing filter (I think this is either Mode 1 or Mode 5 on the Topping D90)
View attachment 338713
No fuzziness or any issues on the chorus or fast paced sections. No clipping whatsoever as well. Based on my personal experience with this exercise, I speculate that I don't get any discomfort with this album that some others can perceive as tinnitus inducing or clipping. Heck I even got goosebumps earlier on how amazing this album sounded to me. One can download the track and load it up to Audacity/Adobe Audition and see if there's a loudness wars issue with the track.

Looking at measurements both the Susvara and the Andromenda have a recessed upper midrange compared to Harman target and Crinacle target for the Susvara and Crinacle target for the Andromeda. This may make it hard for you to hear what I hear on the HD600 which are known to be slightly forward in the upper midrange and measure showing no recession there.

 

[Taiga]

I just played both the Thibaudet and Cho recordings briefly and notice what I would describe as increased reverb or sustain in the notes of the Thibaudet recording. I‘m probably not using the best descriptor, but it does have a sharper less rounded feel to the notes. The Cho is more pleasant to listen to. My wife agrees.

 

[JIW]

The Thibaudet is from 2000 or before and was originally released on Debussy - Images, Études: Complete Works For Solo Piano Vol. 2 on Decca.
On Qobuz it is here: https://open.qobuz.com/album/0002894602472

It sounds much the same with all the same issues as the one on Arabesque although the one on Arabesque sounds a fair bit louder to me. According to YouTube's stats for nerds, the content loudness of the original is -12.6 dB whereas the content loudness on the Arabesque is only -7.5 dB, i.e. it is 5 dB louder.

Also in the Menuet of Suite Bergamasque it is even worse in the beginning around 10 to 15 seconds. On the Ott recording it is clean.

Further, it all happens predominantly in the left channel. In mono it is more subdued but still present and centered. Switching my headphones around such that the left channel is on my right ear, the issues move right.

 

[majingotan]

Thanks for responses!

I probably don’t perceive those extra reflections as harsh or annoying at all hence why I commented that I don’t hear anything wrong with the Thibaudet recording other than the Korean pianist recording as far cleaner sounding with less emphasized reverberations.

Ultimate question though is if any of these Thibaudet recordings induce listening fatigue or tinnitus to you guys?

 

[JIW]

I also used YouTube to check the loudness of the Cho. It is -13.0 dB and thus 5.5 dB less than the Thibaudet Arabesque but only 0.5 dB less than the original Thibaudet. The difference in level alone, the Arabesque Thibaudet being almost twice as loud as the Cho, may explain why it sounds harsh if played on the same volume setting. OP should check if the original Thibaudet causes the same issues.

 

[JIW]

Thanks for responses!

I probably don’t perceive those extra reflections as harsh or annoying at all hence why I commented that I don’t hear anything wrong with the Thibaudet recording other than the Korean pianist recording as far cleaner sounding with less emphasized reverberations.

Ultimate question though is if any of these Thibaudet recordings induce listening fatigue or tinnitus to you guys?

I can listen to it just fine. It does not cause me any discomfort, it just annoys me and distracts me from the music.

 

[antcollinet]

Sorry to click "like" but as a fairly recent member of Club Tinnitus I, at least, empathize. I find certain sounds on certain recordings 'ring' (for lack of a better word) when they didn't before. Aggravating as all get out.
You have my sympathy and empathy.

Me too. I blame Motorhead : 1982 Leicester De Montford Hall.

I call it "resonate" rather than "ring" but probably the same thing.

 

[SMc]

The fuzzy sound is likely the piano's dampers exaggerated by close miking.