Octave Music Don Grusin High Resolution Music Analysis (Video)

[whazzup]

You win, everyone record, mix and edit like Michael Jacksons. Chesky Record, Stereophile, Reference Recording minimum mic, editing and mixing are scam. You need to pull off 48 mulit track, put dozen of eq, compression, editing for 48 hours. It is fairytale to press record and direct mix perfectly with good mic position and console control. A-B micing, X-Y micing ....etc are all gimmicks .

Mixing console control? Still the same behavior of (eventually) smashing multiple tracks into 1 waveform?? With multiple tracks combined now your 'noise' is multiplied correct? There's only 1 disc or music track playing on your home stereo right??

Orchestra / ensemble music are mixed exactly like what you see in the multi-track DAW in the Micheal Jackson example. Maybe your fight is with the sound editors and mixers. Tell them not to use DAW, or any software, in their production workflow so as to maintain the purity of your waveforms. Or tell your recording engineers to use only 1 microphone.

 

[kongwee]

Orchestra / ensemble music are mixed exactly like what you see in the multi-track DAW in the Micheal Jackson example. Maybe your fight is with the sound editors and mixers. Tell them not to use DAW, or any software, in their production workflow so as to maintain the purity of your waveforms.

In real world recording yes and no. Orchestra and ensemble minimum micing only use 4 mic. You can have more. Stereophile and Chesky Record even list their rig on their CD pamphlets and even mic position. Don't even need 24 channel console. A small 4 track rig. Record, edit and master. Don't even need the so call mixing process. Many of these type of recording engineers around. Many just need 2 mic for on site recording for orchestra. I do DAW but I am not that kind of engineer. I have a friend that does it. You have to meet someone like this to believe it.

 

[whazzup]

In real world recording yes and no. Orchestra and ensemble minimum micing only use 4 mic. You can have more. Stereophile and Chesky Record even list their rig on their CD pamphlets and even mic position. Don't even need 24 channel console. A small 4 track rig. Record, edit and master. Don't even need the so call mixing process. Many of these type of recording engineers around. Many just need 2 mic for on site recording for orchestra. I do DAW but I am not that kind of engineer. I have a friend that does it. You have to meet someone like this to believe it.

4 mic or 24 mics, the fact that there's more than 1 mic means your final waveform in your CD or music track has been edited on a multi track device (mixer, daw etc). No such thing as a pure unmolested waveform unless there's only a single microphone.

But sure, have fun pretending that your 4-miked orchestra track has a pure audio waveform...

 

[kongwee]

4 mic or 24 mics, the fact that there's more than 1 mic means your final waveform in your CD or music track has been edited on a multi track device (mixer, daw etc). No such thing as a pure unmolested waveform unless there's only a single microphone.

Remove noise from such waveform is not, coming back to the noise subject. You can derail off the subject whatever you want. Orchestra or ensemble, they won't care to remove noise. This is what they heard before they press the record button. Mic noise, preamp noise so be it. Many of the process is edit and master right away, Totally different metrology from pop songs. I do voice over, one mic is enough. Send to my client raw.

 

[amirm]

Isn't that DAC noise or rather output noise than the audio file?

No, it is encoded in the content. See my OP in this thread.

 

[kongwee]

No, it is encoded in the content. See my OP in this thread.

Over the arrow it say at XLR output. How the noise shape like digitally, I mean the audio wave file itself?

 

[Miska]

I showed the real-time spectrum of the analog output of Marantz SACD player using my Audio Precision:

Yes, that's just Marantz' take on how to do the D/A conversion.

Nope. To pretend there is spectrum above CD in SACD/DSD64 encoding and playback, all or most of that noise is output. From marketing point of view they simply can't filter it all out or they would have nothing left of the ultrasonic response.

Yes, that's the case for DSD64. As I said before, you can consider it bandwidth wise equivalent to 48/24 PCM. It just lacks the brickwall filtering artifacts.

DSD64 has ~25 kHz of noise free bandwidth.
DSD128 has ~50 kHz of noise free bandwidth.
DSD256 has ~100 kHz of noise free bandwidth.
DSD512 has ~200 kHz of noise free banwdidth.
DSD1024 has ~400 kHz of noise free banwidth.

Pick the rate based on how much noise free bandwidth you need. But none of these are plagued by artifacts of the brickwall decimation anti-alias filters. So your time domain performance of DSD64 will beat 176.4k PCM.

PCM doesn't have that issue. You get 96 kHz content and it will have 48 kHz of encoded content.

Problem is that 96 kHz is useless for D/A conversion without subjecting it to heavy DSP processing first. If you are using any delta-sigma DAC for playing it, it will go through all the oversampling and delta-sigma modulation to put the noise back there before it can be converted to analog. If you are playing it through NOS R2R ladder at that rate, you are producing horrendous amount of correlated ultrasonic distortion.

And you won't be able to tell if that is recorded in DSD128 or not. You can even take that, convert it to DSD128 and back, and still you wouldn't be able to tell the difference.

Same way with DSD, you get DSD128 and you get ~50 kHz of encoded content. And with a good modulator you get more dynamic range than 24-bit PCM can provide.

I once made experiment just for fun. I took RedBook file, converted it to DSD64 and then back, and was able to obtain bit-perfect copy of the original, despite the conversion.

Here is a track from Diana Krall 96 kHz sampling album

And here's 96k PCM conversion I just made today using my tools from multichannel DSD256:

But I rather play the original multichannel DSD256 than some PCM conversion of it.

 

[Blumlein 88]

4 mic or 24 mics, the fact that there's more than 1 mic means your final waveform in your CD or music track has been edited on a multi track device (mixer, daw etc). No such thing as a pure unmolested waveform unless there's only a single microphone.

But sure, have fun pretending that your 4-miked orchestra track has a pure audio waveform...

You do understand you can have two channel recordings with one microphone per channel and no processing done at all? No mixing, no mastering, no manipulation. Same for any channel count actually with one microphone per channel.

 

[amirm]

Over the arrow it say at XLR output. How the noise shape like digitally, I mean the audio wave file itself?

Once again, in the OP I show what is inside the file itself, analyzed digitally.

 

[amirm]

Yes, that's the case for DSD64. As I said before, you can consider it bandwidth wise equivalent to 48/24 PCM. It just lacks the brickwall filtering artifacts.

99% of DSD content out there is DSD64. And that is the topic of this review and thread: content released by Octave records in DSD64.

As to "artifacts" you keep repeating that line as if doing so makes it right. The world is PCM. It is a powerful format with multiple sample rates, all supported in even cheapest devices.

Any format like DSD which causes trips to analog for editing completely defeats the purpose of digital (i.e. lossless transformation). Conversion to PCM as an alternative demonstrates that we are far better off starting with PCM.

 

[amirm]

But I rather play the original multichannel DSD256 than some PCM conversion of it.

And I rather have PCM to start, than DSD of any flavor. The wider you make DSD, the more massive the files get with vast amount of noise stuffed in it. Here is an example:

1.2 Gigabytes for just one track! That is 2X the size of a CD! And for what? For this?

Look at how much spectrum is wasted to encode noise. Definition of insanity.

What we need out of "high-res" is 24 bit encoding at something better than 44.1 kHz sample rate. Nothing more. No need to clock something at 11 MHz to capture/create a bunch of noise.

I don't get the impression that you have any feeling or understanding of what is going on here.

 

[amirm]

Same way with DSD, you get DSD128 and you get ~50 kHz of encoded content. And with a good modulator you get more dynamic range than 24-bit PCM can provide.

Best case scenario for music that is available is 18 bits based on detailed statistical surveys. 24 bits provides plenty of headroom for that. You live in a fantasy if you think you can capture or create music at 24 bits let alone needing more.

 

[Robbo99999]

Been listening to the Tidal copy in 16/44. Delightful, relaxing, contemplative. Does anyone here actually like music?

I don't know why this thread is so popular in terms of number of replies, just a load of people spinning their wheels over stuff that doesn't matter or make a difference.

 

[firedog]

If nearly all DSD went DXD for at least one step, and some still claim DSD is better, then why wouldn't you be better off doing everything in DXD with all the processing convenience of PCM? DSD is a nuisance format that doesn't make any sense.

Because DSD sounds/can sound slightly different, and some people prefer that sound. Some devices work best (lowest distortion) at high DSD rates (Miska has demonstrated this). so some like that for playback in their systems.
And, it works for hi-res multichannel.

 

[KSTR]

Once again, in the OP I show what is inside the file itself, analyzed digitally.

... which is as useful as looking at the raw modulator output (on/off) bit stream of a PWM amp, or looking at the RF input signal of an AM radio transmitting antenna and then complaining about "too much RF". Your're not looking at the actual content, you look at a modulated carrier and the carrier obviously dominates the wide-band characteristics of the total signal.

To remove the carrier, proper DSD to analog conversion relies to 100% on a steep analog filter right above 20kHz. All the "general purpose" chip-based DAC implementations don't have those as these devices are also (actually primarily) designed for higher rate PCM content and there a fixed frequency higher frequency analog lowpass filter is preferred, typically not any lower than at 100kHz. DSD has the same problem as NOS PCM, the analog post-filter is paramount and you need a different one for each sample rate.

I think we all agree that a DSD bit stream in form of .DSF files is an impractical and wasteful format but it was never intended to be an external, user-accessible format. Aside from that, when used correctly (which starts at the encoding modulator during recording), DSD is a very nice and competent rendering mechanism. @Miska has explained this, among other main aspects of DSD, in great detail and dependable correctness (as far as I can judge).

 

[thorvat]

I think we all agree that a DSD bit stream in form of .DSF files is an impractical and wasteful format but it was never intended to be an external, user-accessible format.

And what was DSD format intended for? What problem is it trying to solve and/or what solution to optimise?

 

[amirm]

.. which is as useful as looking at the raw modulator output (on/off) bit stream of a PWM amp, or looking at the RF input signal of an AM radio transmitting antenna and then complaining about "too much RF".

No it isn't. We don't store and transmit those streams. Nor do we need to edit them. Regardless I have shown two dacs and hardware sacd player all leaving ton of that ultrasonic noise intact. I have not seen a single Dsd dac which filters to 25 Khz. It is a fantasy that gets repeated without any backup.

 

[amirm]

@Miska has explained this, among other main aspects of DSD, in great detail and dependable correctness (as far as I can judge).

Your judgment is wrong. He is an obfuscation and advertising machine for DSD. you honestly think 24/96 Khz broken as he claims?

 

[kongwee]

Won't doubt Miska knowledge on ADC and DAC whether it is PCM or DSD. DSD doesn't have full commercial chain like PCM. There is no end to debate about it.

 

[Geert]

There is no end to debate about it.

Not if you want to cover all aspects. But the scope of this review is what's in it for the customer. What are the benefits of DSD as an end user distribution format? The drawbacks are clear: gigantic file sizes, can't do DRC without conversion to PCM (in most cases), potentially loads of noise being output by the DAC.