Blind Comparison LIGHTNING ROUND: Beethoven's 'Waldstein'!

[G. String]

Lesson learned for next time. I used Audacity to join the movements in each of the files, after carefully ripping 320 kbps MP3 files from CD sources (except Biss and Bavouzet, which were MP3 downloads at the same bit rate). But then Audacity automatically compressed each file without asking me, down to something like 128, which is low enough that you can tell the difference. If I'd discovered this at the start, it could have been resolved, but since it was after creating all five, honestly irritation/exasperation set in.

So my next game likely won't require any file-editing, or if it does, it will be after a careful comb through all the options and settings. Another issue is, suppose I had provided FLACs of the recordings I owned on CD; my download copy of Bavouzet could have been FLAC, but ClassicsOnline doesn't send me lossless files so audiophiles might have perceived Jonathan Biss at a disadvantage.

My humble advice from experience and technical knowledge is, to convert classical music to mp3, use dbpoweramp with
MP3(lame),
Target: Quality(VBR),
Variable bit quality: -V 0 (estimated bit rate 240kbps) "extreme" --> this means the slider sits on the far right,
Encoding: Does not make perceptible sound difference so does not matter.
settings.

Very easy to use and basic functions are free.

[amw]

I know that people are not supposed to be able to pick out high bit rate MP3 (ie, 320) when compared to CD, yet I've been able to do it, at least when comparing a CD to a CD-to 320 MP3-back to CD conversions. 

I suspect that has to do with sample rate rather than bit rate. MP3s are usually 44.1kHz, which basically means the range of frequencies is compressed to fit inside 22.05kHz. (The sample rate of an uncompressed recording is something like 1.4MHz; frequencies in music can typically be up to 120kHz.) ~21kHz is supposed to be the maximum humans are capable of hearing, but I doubt that to be true. Even if not directly audible, the higher frequencies act as harmonics modifying the sounds we can hear. This is particularly noticeable comparing compressed cymbals, snare drums, violins, etc, to uncompressed ones (at least for me). Of course everyone's hearing is different, and changes over time, as well.

I can't tell the difference between 192kbps (or VBR) and 320kbps MP3 though.

For Audacity, when exporting an MP3 there should be an option on the export dialog. I think it literally says "Option". Then you click on it and can change the bitrate and whatever (and I think it remembers that for next time).

[0spinboson]

I suspect that has to do with sample rate rather than bit rate. MP3s are usually 44.1kHz, which basically means the range of frequencies is compressed to fit inside 22.05kHz. (The sample rate of an uncompressed recording is something like 1.4MHz; frequencies in music can typically be up to 120kHz.) ~21kHz is supposed to be the maximum humans are capable of hearing, but I doubt that to be true. Even if not directly audible, the higher frequencies act as harmonics modifying the sounds we can hear.

http://www.audiocheck.net/audiotests_frequencycheckhigh.php < Feel free to test it. (Mind that not all sound cards will be able to reproduce high frequencies reliably; same goes for speakers/head phones ofc.)

[Pat B]

I suspect that has to do with sample rate rather than bit rate. MP3s are usually 44.1kHz, which basically means the range of frequencies is compressed to fit inside 22.05kHz. (The sample rate of an uncompressed recording is something like 1.4MHz; frequencies in music can typically be up to 120kHz.) ~21kHz is supposed to be the maximum humans are capable of hearing, but I doubt that to be true. Even if not directly audible, the higher frequencies act as harmonics modifying the sounds we can hear.

No.

The sample rate of CDs is 44.1kHz. Typically mp3 compression does not change the sample rate.

The bit rate of uncompressed CD-quality audio is about 1.4 Mbps.

20kHz is the generally accepted audible maximum for humans. Some people have a lower limit, some have a higher limit. Nobody can hear 120kHz, and even if you could, your playback equipment probably can't reproduce it. Harmonics are no different. If you can't hear a 30kHz fundamental then you can't hear a 30kHz harmonic. The highest-pitched orchestral instruments go up to about 4kHz, so even at the very top of their range a 44kHz sample rate signal can represent many harmonics.

mp3 depends a lot on the encoder. The early (1990s) encoders were really bad (e.g. underwater cymbals). Newer ones are better. It wouldn't surprise me if some of the older ones are still in use.

Mods feel free to move this discussion to a more appropriate thread.

[DavidW]

Brian just said that it was doubly compressed so I think we can drop the technical arguments.  The samples did sound worse than radio.  Yuck!

[Ken B]

http://www.audiocheck.net/audiotests_frequencycheckhigh.php < Feel free to test it. (Mind that not all sound cards will be able to reproduce high frequencies reliably; same goes for speakers/head phones ofc.)

Cool. I did the audio checks too. Hearing still good, but cannot hear 19k and 18k intermittent, maybe need a quieter room. Or its the headphones 

[amw]

No.

Hmm. I suspected I was talking nonsense but knew there was something going on with high frequencies that I remembered.

Some research led me to this. They're inaudible, but affect our perception of sound.

But then more research led me to this which says, actually, pianos don't make any sound above 15 kHz anyway, so the problem with MP3s probably has more to do with sampling algorithms and so on.

Now I'm confused. Too much science for my tiny arts-major brain. Oh well.

[Pat B]

Hmm. I suspected I was talking nonsense but knew there was something going on with high frequencies that I remembered.

Some research led me to this. They're inaudible, but affect our perception of sound.

But then more research led me to this which says, actually, pianos don't make any sound above 15 kHz anyway, so the problem with MP3s probably has more to do with sampling algorithms and so on.

Now I'm confused. Too much science for my tiny arts-major brain. Oh well.

Regardless of the technical underpinnings, you seem to have a pretty good grasp on what works for you. Beyond that, don't worry about it. Enjoy the music.

[NJ Joe]

Well, I picked the two bottom finishers as my two top choices.  Interesting about Biss:  while immersing myself in the Gulda set last year, I was hungry to read something about the sonatas.  I wound up reading Biss's short book Beethoven's Shadow, and enjoying it thoroughly.  Although he wasn't on my short list of future Beethoven sonata purchases (Fischer, Gilels, Richter, and Serkin were) I will now reconsider, although Fischer still remains at the top.

Thank you so much Brian for organizing this, it was a lot of fun.  As I am almost always under limited time constraints, I really enjoyed the "short clip" format.

[kishnevi]

Those interested in Serkin should take note of his Diabellis, which are apparently not in the box Brian posted.

Bavouzet's second volume is approaching the peak of Mt. ListenTo, so I can't comment directly.  Whether or not you liked this clip,  I would suggest checking into his Debussy cycle.

[amw]

Those interested in Serkin should take note of his Diabellis, which are apparently not in the box Brian posted.

Hey, those are my reference recording of that piece. And I forgot about the Serkin LvB piano concertos I grew up with on cassette. No surprise I liked his Waldstein I guess.

[Todd]

Those interested in Serkin should take note of his Diabellis, which are apparently not in the box Brian posted.

The '57 Marlboro Diabellis are included, but the remastering used is not as good as some others.  The crickets are gone, and the sound is a bit duller.  I had to keep an earlier Sony France remastering to hear the performance in all its glory.

[kishnevi]

The '57 Marlboro Diabellis are included, but the remastering used is not as good as some others.  The crickets are gone, and the sound is a bit duller.  I had to keep an earlier Sony France remastering to hear the performance in all its glory.

My copy is this.  I don't remember any crickets.
No information is given about remastering,  but the copyright date was 2002. 



That was my first recording of the Diabellis.  Followed soon thereafter by the Brilliant box containing Brendel's recordings of all the variations and other pieces that was originally done for Vox;  I remember that the idea that I would now have two recordings of the same piece was a negative.   

[Todd]

My copy is this.  I don't remember any crickets.
No information is given about remastering,  but the copyright date was 2002. 

Tracks 32 and 33 have them.  Well, either crickets or something mechanical that sounds a lot like crickets.  That's the remastering to have, and still probably the best Diabelli recording I've heard.

[kishnevi]

Tracks 32 and 33 have them.  Well, either crickets or something mechanical that sounds a lot like crickets.  That's the remastering to have, and still probably the best Diabelli recording I've heard.

And so I got it as absolute beginner's luck?!

I don't have time tonight for the whole thing, but I'll get it and listen to those two tracks, once the current CD is over (also Beethoven:  Abbado/BPO doing Opp. 92 and 93)

[Todd]

And so I got it as absolute beginner's luck?!

That can happen sometimes.  If memory serves, the CBS Odyssey issue of the same recording was my second or third version, and this reissue came a few years later, and nothing has quite matched, let alone bettered it.  Hearing Anton Kuerti play it live was probably at the same level, but that was but a fleeting experience.

[Bogey]

I cannot believe I chose Serkin....one of my all time favorites!  Kind of proud that I put him in first out of the group.  And it looks like I better take a look at getting #5's cd!  Thanks, Brian.

[Ken B]

And so I got it as absolute beginner's luck?!

I prefer the more euphonious "intuition".

[kishnevi]

I prefer the more euphonious "intuition".

The fact that it was one of the only two versions at my local Borders (not sure how many hundreds of CDs I ended up buying there in the next few years, as it slowly dwindled and then went out of business) was the main reason I got it.  The other version was the Brendel box I got a few weeks later.

Now playing tracks 32 and 33 of that CD.  Not hearing crickets, but I do hear what appears to be some Gould like vocalizing randomly appearing in a ghostlike way.

ETA: there is some further noise on track 34,  including what sounds like someone stepping on something crunchy.

[Todd]

Now playing tracks 32 and 33 of that CD.  Not hearing crickets, but I do hear what appears to be some Gould like vocalizing randomly appearing in a ghostlike way.

Serkin vocalizes a little throughout.  The crickets are noticeable in the quieter passages.