yamamoto2002

Is the source code and training dataset available on GitHub / paperswithcode / HuggingFace to reproduce their result?

Thank you for sharing Interesting sound comparison! It seems listening position contributes a lot to sound difference, a Siltech is more direct sound of program material, while many room reflected sounds are blended on a Vivid Audio.

sorry it is redundant, my post deleted

I think Luxman Balanced XLR output is 600Ω. Still, their balanced input is high-impedance, so, you are correct, it is low output impedance to high input impedance link when 600Ω transformer is not used. This is from their latest CD player D-03x specifications.

On XLR balanced interconnect, there are two different forms: low output impedance to high input impedance link (I don't know English term for this, it is literal translation from Japanese term): this is explained by Paul. “impedance matched” 600Ω output to 600Ω input signal transfer: This method is used to maximize the signal power transferred, it is used between active device and passive device, such as the preamp to your galvanic isolation transformer. If you'd like to Y split and isolate the XLR signal of preamp, it is better to Y split after isolate, than Y split then isolate each branches.

According to TPA3255 (BT20A amplifier module) datasheet, it is very efficient class-D amplifier and the system efficiency is 90% (on 300W output, the system draws 333W and 33W is dissipated as a heat). Idle power consumption is 2W. (For comparison, class-A amp efficiency is 25%) On musical signal input, there is no concern for overheating. If this device is used to amplify continuous sine signal to drive 300W electric motor, it may overheat. When it is happened, output power is automatically reduced to prevent silicon degradation.

Last I checked, "SoX sound exchange" can perform de-emphasis. SoX is available on Apple Macintosh, search "sox macos port" Following is the example to input inFile.flac, perform 50/15 microseconds de-emphasis linear-phase FIR, write result as outFile.flac sox inFile.flac outFile.flac fir 0.00031102739649091732 0.00036885685453166665 0.00057659816229764602 0.00082952248115418167 0.001449970925925961 0.0022387507393955598 0.0039420948394406248 0.0063363475292175873 0.011215231698621361 0.018685854350970088 0.033951734838472802 0.059076192885731141 0.11592376177923121 0.49018811103703697 0.11592376177923121 0.059076192885731141 0.033951734838472802 0.018685854350970088 0.011215231698621361 0.0063363475292175873 0.0039420948394406248 0.0022387507393955598 0.001449970925925961 0.00082952248115418167 0.00057659816229764602 0.00036885685453166665 0.00031102739649091732 How to design FIR filter to de-emphasis CD https://yamamoto2002.blogspot.com/2020/11/designing-de-emphasis-digital-filter.html

I'd like to say, those computer will depreciate rapidly see bargainhardware Recently I'd built proxmox server with dual processor Xeon and 512GB memory CPU costs $85 each (Intel Xeon Processor E5-2697A v4) Memory costs $500 total (64GB ECC DDR4 LRDIMM x8)

My understanding of recording, mixdown and mastering workflow is as follows. Someone should create different master data for respective media to meet target media requirements and sound preference of its typical audience

BTW, this is my implementation of DSF to FLAC conversion (Windows app) https://sourceforge.net/projects/playpcmwin/files/others/DsfToFlacGui100.zip/download https://sourceforge.net/p/playpcmwin/code/HEAD/tree/PlayPcmWin/WWDSPCpp/WWSdmToPcm.h

Mastering process of DSD release and FLAC release of the same music may be different. I read it on local sound engineer magazine, they intentionally makes DSD package more hi-fi (less effects, more dynamics, closer to real performance) on mastering stage for home listening quiet environment, while on CD (PCM) mastering, sound is more punchy and compressed because it is heard on cars, smartphone headphones on train commuting with a lot of ambient noise. Generally it is natural DSD mastered product and PCM mastered product of the same music have the different sound, because those two are different product for different audiences There are several different method to convert PCM to DSD but there is essentially one method to convert DSD to pcm: Sinc function convolution. Converting DSD to PCM tends to produce low sound volume PCM, so it is necessary to match the level before comparing DSD and converted PCM sound. This is my result and YMMV but honestly speaking I cannot hear the difference of DSD and its 44.1kHz downsampled PCM, those two sounds the same after rigorous my listening comparison. when I first thought to find the different sound feature, after comparing by switching the portion several times, the difference disappears, and eventually understand my idea that found the difference is delusion and sound feature is presented on both data with the same manner.

Existing dCS owners will appreciate if exterior design is not changed very much from predecessor models, like Accuphase amplifier or Leica M series or Toyota Century, it is better the family does not notice the replacement

there are many foreign language songs heard as inappropriate Japanese

search modern coachbuilds to find more expensive cars. IMO most of them are fail but there are several beautiful bespoke cars

I vaguely remembered some person ordered one-off, money-is-no-object DAC to famous DAC designer about 10 years ago and it costs about 200K

I'm fan of Varèse, have several CDs of Ionisation and Intégrales, intense but somehow feel of tradition of classical music from the pieces I feel some sonata form from this piece Ionisation

I think the SQ difference comes from slight level difference of volume level. It is well known the listening comparison fails with mere 0.2dB ～ 0.3dB level difference, louder sound is perceived as more expressive, more detailed, wider sound stage. Human ears do distinguish the 0.2dB level difference, but it is sometimes heard as quality difference, not heard as level difference.

I first suspected the 6dB level difference (SACD gain) but re-read your post and you are confirmed it is level matched, I don't have any clue sorry

There was a discussion of music complexity on this forum in January 2024. As Rick said that in the video at 11:40, Many chord progression possibilities are explored on 1970s. Now all the songs use only 3 or 4 codes. IMO pop music is more and more focused to be listened by kids or very distracted adults, and it is getting worse for concentrated listening of adults. IMO making music touching, be moved, thrilled is the most important, and there is no problem to apply technology advancements, Moog, Roland 808, FM synthesizers like Yamaha DX7, PCM synthesizers like Yamaha Motif, DSP, or AI, this becomes the sound signature/merkmal of the era. The most complex chord changed song appeared on Billboard 100, with 52 different codes. "After the Lovin'" Engelbert Humperdinck 44 different codes used. "One" Metallica Those use 33 chords. "Last Dance" Donna Summer, "With You I'm Born Again" Billy Preston

I agree about kernel modifiability of Linux, I read lengthy linux/Documentation/Configure.help and had build small sized kernel (400KB) that includes functionalities only I use, when Linux was version 2.0.28. On Apple computers, I actually don't know about recent MacOS or iOS... Clearly they don't welcome hobby amateur programmer. I don't understand why I should pay money to Apple to write free app and deliver it to friends' iPhone. (On Android, it is free)

Interesting. so, competitor OS prohibit for 3rd party antivirus to run in kernel mode or patch the kernel, but, Microsoft cannot prohibit them as others did, because of EU region antitrust concerns... It is somewhat difficult to perform recurrence prevention by fixing the root cause.

I block the connection to OneDrive with firewall rules but the blocking list should be updated when Microsoft decided to use different domain name. onedrive.com, onedrive.live.com, 1drv.com, storage.live.com, skydrive.live.com, settings.live.net, livefilestore.com, ... The same nonsense happens on blocking Facebook. facebook.com, fb.com, fb.net, facebook.net, instagram.com, instagram.net, threads.net, ...

My understanding is, Microsoft Windows has a fundamental design flaw that enables 3rd party app to crash instantly and prevent next boot. Usually operating system do not allow such action to any 3rd party app. I'm a seasoned Microsoft fan and do not surprised at all about it, you should feel lucky when Windows works and do not crash. if it crash often, increases the joy when it runs well.

OK I recorded home AC power anomaly events using PicoScope 4444 and PicoConnect 442 CAT III probe. This happens when a laser printer starts. Transient noise. it happens when Desktop computer switched on. Also happens on distant lightning strike Instant voltage Sag. I suspect it happens when some high power electric appliance switched on. This is introduction video of pico4444 found on youtube

On February or March 2024, It seems OneDrive silently altered the sync settings and now it automatically copies "My Documents" "My Pictures" and "Desktop" onto OneDrive when the computer is logged on Microsoft account to Windows 11 computer. If you are using Windows 11 Home operating system or logging in Windows 11 Pro computer with Microsoft account, it is necessary to check sync settings of OneDrive. For security reasons, it is better to not store items onto "My Documents" "My Pictures" "My Videos" "My Music" "Desktop" because it is typical target folder of “ransomware”