Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Neil Young’s new audio format (Read 130333 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Neil Young’s new audio format

Reply #300
Another technical writer reveals his disappointment at Young's "revolutionary " promise not being after all, revolutionary at all.
I was going to question this part...
Quote
“In order to hear a difference between FLAC and AAC, you’d need to generate a special test tone. For any sounds you’d ever actually want to listen to, you’ll hear no difference.”
...but since he's talking about 256kbps AAC in a non-scientific article, I think it's fine.

Are there many 256kbps AAC problem samples?

Cheers,
David.


Neil Young’s new audio format

Reply #302
Another technical writer reveals his disappointment at Young's "revolutionary " promise not being after all, revolutionary at all.
I was going to question this part...
Quote
“In order to hear a difference between FLAC and AAC, you’d need to generate a special test tone. For any sounds you’d ever actually want to listen to, you’ll hear no difference.”
...but since he's talking about 256kbps AAC in a non-scientific article, I think it's fine.

Are there many 256kbps AAC problem samples?



Possibly more to the point, are there any samples that are a problem when supplied to the coder in a hi rez format, but not a problem when supplied in a standard (16/44, 16/48) format, or vice versa.

I'm under the impression that even the best coders decode audio with that fits perfectly into a 16/44 linear PCM format. Is that really true?


Neil Young’s new audio format

Reply #304
You can encode (HE)AAC at 96kHz and get some ultrasonic content back out (google it - it's not common, but it's been done). However, I don't think that's the point at all.

Cheers,
David.

Neil Young’s new audio format

Reply #305
You can encode (HE)AAC at 96kHz and get some ultrasonic content back out (google it - it's not common, but it's been done). However, I don't think that's the point at all.


When dealing with many of the people we deal with, it is IME good to know the exceptions to the rule, no matter how questionable their relevance.

Thanks.

BTW in the heretofore unknown (to me) Pono (and HiRez)  supporting bafflegab department, let me offer this:

Yet another hi rez advocate speaks

"
Working in the digital audio domain, there are quite a number of reasons to increase the underlying sampling rate of the system. First of all, most feedback based algorithms are taking advantage of a higher SR resulting in a better perceived audio quality. Most prominently, all types of IIR filters are affected especially but not limited to the case where very steep slopes are computed. As an added sugar, the curve warping gets moved outside the hearing spectrum at higher SR.

Whenever modulation at audio rate occurs, there will be serious distortion. This can be minimized to some extend at higher SR and so a wide range of digital audio applications is capitalizing from it: FM oscillators, ring modulation, compression and limiting – just to name a few. The so-called IMD (inter modulation distortion) is even worse in the digital domain since also aliasing gets introduced. This occurs when the additionally created distortion content exceeds the Nyquist frequency and gets folded back into the spectrum below. Increasing the SR, immediately relieves this effect. In general, this applies to all kind of non-linear processing and therefore all types of saturation and distortion algorithms are benefiting from higher SR.
"

Neil Young’s new audio format

Reply #306
...but that's about production. If people want to use higher sample rates in production, so be it. It's not the panacea that writer claims. If it's the only trick you use to avoid aliasing when doing the kind of processing mentioned, you need to MHz sampling rate for audio. Quite obviously, there are other, better tricks that good DSP designers know only too well. However, a higher sample rate helps a bit as well.

That is irrelevant to home delivery formats.


It gets like Groundhog day, doesn't it? We must have had these discussions 10 or 20 times, and still the same misleading stuff gets written. Anyone would think there was money to be made from misleading people...

Cheers,
David.

Neil Young’s new audio format

Reply #307
...but that's about production.


I think that may be an overly-generous opinion.

For example:

"
First of all, most feedback based algorithms are taking advantage of a higher SR resulting in a better perceived audio quality. Most prominently, all types of IIR filters are affected especially but not limited to the case where very steep slopes are computed. As an added sugar, the curve warping gets moved outside the hearing spectrum at higher SR.
"

From the second sentence we can discern that by "feedback based algorithms" the author is describing IIR filters.  The first sentence thus seems to claim that  "Most (IIR filters) are taking advantage of a higher SR resulting in a better perceived audio quality. 

After looking at a ton real world of IIR filter designs and design guides I see that there is no such general rule.  Have I missing something?

It is true that a lot of filters that are used alonq with oversampling are IIR filters, but that seems to be putting the cart before the horse.  AFAIK the oversampling was done for a different purpose than facilitating the use of IIR filters.  This seems to have been true for a long time. For example the  DAC used in the first generation Philips CD 100 was over sampled, but history shows that was done  to improve resolution.

Another example:

"Whenever modulation at audio rate occurs, there will be serious distortion. This can be minimized to some extend at higher SR and so a wide range of digital audio applications is capitalizing from it: FM oscillators, ring modulation, compression and limiting – just to name a few."

Again this is AFAIK just not a general rule of digital audio.  IME, in general modulation of audio frequencies by audio frequencies in the digital domain yields near-ideal results without upsampling.  There are some exceptional conditions where it might be true, but this guy seems to want to represent that this problem happens every time. It doesn't!

BTW, both of these situations relate to reproduction at least as much as they relate to production.

Neil Young’s new audio format

Reply #308
...but that's about production.


I think that may be an overly-generous opinion.
I thought the clue was in the first line of the article, not to mention the blog's subtitle.

Quote
Again this is AFAIK just not a general rule of digital audio.  IME, in general modulation of audio frequencies by audio frequencies in the digital domain yields near-ideal results without upsampling.  There are some exceptional conditions where it might be true, but this guy seems to want to represent that this problem happens every time. It doesn't!
As you have written it (modulation of one frequency by another), you get sum+difference products. For "audio" signals, doesn't that get you to 20kHz+20kHz=40kHz signal i.e. double sampling rate required to avoid aliasing?

IIRC simply kinking the transfer function a little in the digital domain (e.g. passive soft clipping, or passive low level grunge, or ...) gives horrible aliasing if you do it in a straightforward manner at the base sample rate.

I'm not going to embarrass myself by trying to remember the details of frequency warping and accuracy near Nyquist wrt IIR filters. I've not addressed that topic properly for 15 years. Maybe someone who still does it knows whether there's any truth to what's written in that blog.

Cheers,
David.

Neil Young’s new audio format

Reply #309
...but that's about production.


I think that may be an overly-generous opinion.
I thought the clue was in the first line of the article, not to mention the blog's subtitle.

Quote
Again this is AFAIK just not a general rule of digital audio. 

IME, in general modulation of audio frequencies by audio frequencies in the digital domain yields near-ideal results without upsampling.  There are some exceptional conditions where it might be true, but this guy seems to want to represent that this problem happens every time. It doesn't!
As you have written it (modulation of one frequency by another), you get sum+difference products. For "audio" signals, doesn't that get you to 20kHz+20kHz=40kHz signal i.e. double sampling rate required to avoid aliasing?

IIRC simply kinking the transfer function a little in the digital domain (e.g. passive soft clipping, or passive low level grunge, or ...) gives horrible aliasing if you do it in a straightforward manner at the base sample rate.


I don't think you don't recall correctly.

If you want "Horrible aliasing"  you need source material with exceptional amounts of high frequency content and a large  nonlinearity.  "A little"  e.g. less than 0.1% won't do it with typical music.

For example, if the clipping is serious, the unaliased artifacts due to the clipping itself are generally far more audible.

One of the more instructional facts of life is how good 44 KHz DACs whose brick wall filters have minimal attenuation (as little as - 3 dB) at Nyquist actually sound.  Heck 44 Khz DACs with no brick wall filters at all generally don't sound all that bad.  For one thing, any signal in the music that might be present at those frequencies > 20 KHz is usually 40 dB or more below FS.