Hahah! Yes indeed! I remember those fantastic 1bit dacs. 
Great sound! I had a cd player with one of those.
I think it is important that we see this issue as non-academic, because we have not yet really pushed the limits of 24bit sound. To keep pushing for development, and get actual improvement in sound quality, we fanatics should stand together.
Great sound! I had a cd player with one of those.
I think it is important that we see this issue as non-academic, because we have not yet really pushed the limits of 24bit sound. To keep pushing for development, and get actual improvement in sound quality, we fanatics should stand together.
I am unaware of microphones that are flat to MHz frequencies and could record such transients. I'd be grateful if you could point me to one, and also to any recordings made using it. Clearly whatever digital format is used, it can't be any better than the limitations of the physical sensors that are used to capture the sound.
Can you suggest a specific high resolution grand piano recording that you have in mind, so that I can perform your experiment for myself? I guess I would need to downsample the 16/44.1 version from the high resolution version myself, to rule out any possibility that the publisher had applied different processing to the 16/44.1 and high rez versions.
eClassical - Rachmaninov - Piano Concertos Nos 2 & 3
If you buy the Studio quality recording from eclassical.com you get both the 24/96 and the 16/44,1 files for easy comparison.
The difference is easily discernible on a pair of Dayton PS95-8 wide band drivers driven by a simple source follower amp based on IXYS IXTP32P05T mosfets. This is directly feed from my MiniDSP2X4HD. There is no voltage gain in my amp.
Regards,
Johannes
eClassical - Tchaikovsky & Babajanian - Piano Trios
BIS Records - Tchaikovsky & Babajanian - Piano Trios
Another great 24/96 record with great sound quality.
Available in both 24/96 and 16/44,1 formats in the same download for easy comparison.
BIS Records - Tchaikovsky & Babajanian - Piano Trios
Another great 24/96 record with great sound quality.
Available in both 24/96 and 16/44,1 formats in the same download for easy comparison.
One non technical difference to consider is that hires files are many times marketed for "audiophiles" and can be mastered direrently than the 44/16 versiones, most evidently in the dinamic compresion they apply.
Thus many times you will get better quality with hires but not for the technichal specs of the format...of course there are exceptions also and even upsampled files being sold as "hires"!!, at the end for me each album I buy has a research done before so I know which is the best version, being mastering more important than format...if the same mastering, well, a lot of info on this thread for discussion
Thus many times you will get better quality with hires but not for the technichal specs of the format...of course there are exceptions also and even upsampled files being sold as "hires"!!, at the end for me each album I buy has a research done before so I know which is the best version, being mastering more important than format...if the same mastering, well, a lot of info on this thread for discussion
It's like digital cameras. An 8Mpix camera with a good lens is better than a 16Mpix camera with a poor lens.
Early CDs sounded bad, not because the sample rate and resolution were modest, but because the recordings were made using sharp analog filters that had massive phase shift. Then came oversampling and digital filters.
So a regular CD can sound just fine if the recording process is well managed and an impressive DAC can not fix a bad recording.
OBTW, be sure you don't confuse 192Kbps with 192KHz. Not the same thing at all.
Another point is that the human ear has limitations that certain "distortion" can aid. Do not assume that greater fidelity will sound better. I have heard familiar music become clearer when processed by a modest mp3 compression because the compression removed ~noise. Some communications equipment deliberately adds distortion to make speech clearer.
The Aural Exciter And The Audio Enhancer - A Guide
Early CDs sounded bad, not because the sample rate and resolution were modest, but because the recordings were made using sharp analog filters that had massive phase shift. Then came oversampling and digital filters.
So a regular CD can sound just fine if the recording process is well managed and an impressive DAC can not fix a bad recording.
OBTW, be sure you don't confuse 192Kbps with 192KHz. Not the same thing at all.
Another point is that the human ear has limitations that certain "distortion" can aid. Do not assume that greater fidelity will sound better. I have heard familiar music become clearer when processed by a modest mp3 compression because the compression removed ~noise. Some communications equipment deliberately adds distortion to make speech clearer.
The Aural Exciter And The Audio Enhancer - A Guide
Last edited:
Have you ever compared them yourself? You'll hear the difference. The digital processing, such as interpolation, dither and FIR at 44.1K/16 bit affect the digital data way below 15K.
I believe there is q missanderstanding about decibel and dynamic here. The most common is to mix transcient or short phrases in a big classical orchestra and the continuous decibel of a rocketship at Cap Canaveral. (around 120 dB contnuous)
The noise floor when silence is made by the director is around 40 dB and the highest pressure measured on some programx is around 125 db to 130 dB but are very short or made of transcients. Ears tolerate thosebtranscient peaks. However the rocketship during its 1 minute will cause damages t¥ your ears cause it's continuous (understand too long for our hearing system at such dB levels).
If you want serious classical'you shoumd have a system anable to such transcients, the room for such a job and of courqe the reccordi g allowing to it. Good luck ! Cinema venues is your best chance as the r3ccording chain and reproducing systeme is not the same than home Fi.
Luckilly for us, we don't only liwten transcient when listening musics...but for the moment no system is able to copy the feeling of a big orchestra yet. While sometimes you may listen to more details on some reccordingx vs same live event... That is because magazine rzviewer always focus on speakers clearness and troncated bandwith and particular listening roomw and so on...
So trade offs must be made for us poor hifists.... I only focus with jazz and classical rzccordingq having soul in the way it's captured and mixed, and deqpite laka of dynamics I didn t find better trade off than a 16 bits tda1541A S1 or a an 20 BITS AD1862, cause texture and tonal balance are things ears arz very foccused on... and it's also highlighted with our bad hifi habits as foccusing on the samd things as details when testing hifi devices...
One has to reset his ears in venues time to time and bits or sample rates becomes'less important. Reccordings is tne thing to focus on.
I also listen to a lot of rock and worldwide music and I am verry happy wirh my old pcm dacs that I find more accurate in all the trade offs related to music reproduction.
Of course I may be biased....Just for what it worths.
How about 29 kHz and 30 kHz applied simultaneously, or 14 kHz and 22 kHz simultaneously? If that low-pass filter is not entirely linear (and even the air is not entirely linear - there is always an asymmetry because pressure can't go negative), you may get an intermodulation product at 1 kHz or 8 kHz, respectively. I haven't a clue if this effect is of any significance at the ultrasound levels in normal (gamelan) music, though.
This!
And as I see it, using something like 96khz samplerate is not mostly for the benefit of the quality of the file. But more to try and ensure proper samplerate during the rest of the signal chain.
If it's been upsampled from 48khz to 96khz, then I do not see that as a humongous problem, easy to just double. Worse to downsample from 96khz to 44.1khz, not all the points fit easily.
Hardly any recording studio that does NOT use 96khz or higher these days. I am in process with a record company and they demanded minimum 16bit/44.1khz, but I gave them 24bit 96khz. Most artists, no matter how well known, work with 24bit. The relatively few I've talked to (and I share this conclusion after a lot of trial) claim there is no real benefit to using more than 48khz on the files.
Hardly any recording studio that does NOT use 96khz or higher these days. I am in process with a record company and they demanded minimum 16bit/44.1khz, but I gave them 24bit 96khz. Most artists, no matter how well known, work with 24bit. The relatively few I've talked to (and I share this conclusion after a lot of trial) claim there is no real benefit to using more than 48khz on the files.
I bet you've never heard a song with more than 44kHz / 16bit. There is none. From the studios come only 44kHz / 16bit. Why? Because humans aren't bats. Even if you've ever heard a 192kHz / 24bit song, that was none. That was just an upsampled 44.1k wolf in sheep's clothing.
Let us summarize please, would like to understand if I get it please...
Assuming FauxFrench has a room system or headphones chain that is good enough in dB/noise ratio, able to handle full scale dynamic of the reccordings, and the High Res preserve such dynamic in the reccording process. And the volume control allows also to proof that (means 32 bits dac to hear 24 bits reccording with the difference for digital volume control?)
He indeed should check than the 24/192 reccordings is not truncated in term of mastering process and of course reccording. It litteraly means two master process : one for the main markett with all the complex mastering made by the the sound engineer. And another one where there are less artifact, less psycho acoustic artifacts, higher rejection of preringing because the 192 Khz sampling process, mics able to capture such band pass and pressures a,d the garanty the ADC converter is 24/192 or more.
So it means recent reccording and high pressure and dynamic event (classical grand orchestra for instance) and the garanty when you buy it it's a true 24/192 proof chain capture and mixing.
A dac and a streamer that have no automatic oversampling process such some of the todays DAC chips.
Then it could be great to buy the two reccordings : redbook and HiRes to check in a DBX process (wife or soons to lauch randomly the reccords) if and what are the differences ! And then check what are able to produce guys like Chesky who cares of the full chain of the reccording process (remember : mics on heavy sand bags and so on...).
Am I right with that ? Are the web main platforms transparent enough with what we buy in HighRes format (ADC rate, sampling process, way it's mixed...?)
What could be the hearable improvments ?
24 bits : so higher noise ration and higher dynamic scale ! I read the theorical limits of our ears is 21 dB ?! More informations due to more samples and shorter time windows ?
192 K Hz : less pre ringings and choice of more natural filter for psycho acoustic and ears (phase) and less mixing artifacts from the enginner in the band below 20 K Hz ?
I really would like reccording engineers testimonie about that and assuming I'm not missunderstanding what we talk about. Sorry if too much simplification.
Assuming FauxFrench has a room system or headphones chain that is good enough in dB/noise ratio, able to handle full scale dynamic of the reccordings, and the High Res preserve such dynamic in the reccording process. And the volume control allows also to proof that (means 32 bits dac to hear 24 bits reccording with the difference for digital volume control?)
He indeed should check than the 24/192 reccordings is not truncated in term of mastering process and of course reccording. It litteraly means two master process : one for the main markett with all the complex mastering made by the the sound engineer. And another one where there are less artifact, less psycho acoustic artifacts, higher rejection of preringing because the 192 Khz sampling process, mics able to capture such band pass and pressures a,d the garanty the ADC converter is 24/192 or more.
So it means recent reccording and high pressure and dynamic event (classical grand orchestra for instance) and the garanty when you buy it it's a true 24/192 proof chain capture and mixing.
A dac and a streamer that have no automatic oversampling process such some of the todays DAC chips.
Then it could be great to buy the two reccordings : redbook and HiRes to check in a DBX process (wife or soons to lauch randomly the reccords) if and what are the differences ! And then check what are able to produce guys like Chesky who cares of the full chain of the reccording process (remember : mics on heavy sand bags and so on...).
Am I right with that ? Are the web main platforms transparent enough with what we buy in HighRes format (ADC rate, sampling process, way it's mixed...?)
What could be the hearable improvments ?
24 bits : so higher noise ration and higher dynamic scale ! I read the theorical limits of our ears is 21 dB ?! More informations due to more samples and shorter time windows ?
192 K Hz : less pre ringings and choice of more natural filter for psycho acoustic and ears (phase) and less mixing artifacts from the enginner in the band below 20 K Hz ?
I really would like reccording engineers testimonie about that and assuming I'm not missunderstanding what we talk about. Sorry if too much simplification.
Are we asking the right questions and answering the right things?
16 bit only allows for 65536 quantisation levels for the sound.
Human hearing has a safe dynamic range of roughly 100dB between 20Hz and 20KHz
That 100dB is frequency dependant of course and we’re far more sensitive at certain frequencies, around 2KHz we’re most sensitive to sound.
So, at 2KHz, is 65,536 steps enough to represent 100dB of dynamic range?
In easy ratio numbers that’s roughly 0.66 bits per step of sound
I’d say that 24Bit is a worthwhile increase....(16.7 million levels)
192Khz....
Less of a valid argument there unless you’re talking 4th and 5th order harmonica and shifting those away from the hearing limit for recording. No real difference for playback though as our ears aren’t capable of registering the difference
16 bit only allows for 65536 quantisation levels for the sound.
Human hearing has a safe dynamic range of roughly 100dB between 20Hz and 20KHz
That 100dB is frequency dependant of course and we’re far more sensitive at certain frequencies, around 2KHz we’re most sensitive to sound.
So, at 2KHz, is 65,536 steps enough to represent 100dB of dynamic range?
In easy ratio numbers that’s roughly 0.66 bits per step of sound
I’d say that 24Bit is a worthwhile increase....(16.7 million levels)
192Khz....
Less of a valid argument there unless you’re talking 4th and 5th order harmonica and shifting those away from the hearing limit for recording. No real difference for playback though as our ears aren’t capable of registering the difference
It seems the more bandwidth the better https://www.diyaudio.com/forums/eve...difference-2-solid-preamps-2.html#post5966475
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.