• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

KeithPhantom

Addicted to Fun and Learning
Forum Donor
Joined
May 8, 2020
Messages
641
Likes
658
Also, I'm wondering. How one would differentiate a 44.1 16 bit flac, vs a 44.1 mp3 as another thing that crossed my mind. Of course this assumed MP3 V0 or 320kbps.
I do not even bother, especially with the music I listen the most often. Only $4.99 for Spotify Premium is enough and I do not bother for going to lossless anymore. If I even bother with the quality of the music, I bother because of the mixing, mastering, mic placement, and those kinds of things where you have a difference worthy to be discussed. But most of the time, I close my eyes and forget the quality of the gear and all of that because is all taken care of already.
 

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
I do not even bother, especially with the music I listen the most often. Only $4.99 for Spotify Premium is enough and I do not bother for going to lossless anymore. If I even bother with the quality of the music, I bother because of the mixing, mastering, mic placement, and those kinds of things where you have a difference worthy to be discussed. But most of the time, I close my eyes and forget the quality of the gear and all of that because is all taken care of already.

Don't really stream because large portions of my music aren't all that common on major vendor offerings. Also like using JRiver for example to manage DSP and such.

I actually don't even care about quality much either. I just have lossless for archival and home listening (and also because JRiver for some stupid reason still doesn't support Opus metadata support, but it can play the files). I actually convert all my music to 96kbps Opus files (not 192 or anything like that, but actually 96), and they all sound great (mostly used for mobile phone listening on the go).
 

krabapple

Major Contributor
Forum Donor
Joined
Apr 15, 2016
Messages
3,169
Likes
3,717
Some DVD-Audio had Watermark protection that would only allow 3 to 5 seconds of playback in players that supported watermark detction. And that watermark signal is somehwere in the ultrasonic, forgot which frequency exactly. It is hard to remove without cutting out that frequency band. There is a process explained somehwere. Anyway, i found it was much easier to convert it to DSD.
DVD-Audio are now hard to play in foobar. They are messy. SACD-R works like a charm. You have your 2ch and your mch stream and tags.... You can upsample everything to DSD128 or 256... or leave it as is...


Ripping your DVDAs to 'backup' files is simple, and has been for many years now, with relatively inexpensive, legal software like DVD Audio Extractor. There is no sonic degradation.

SACD ripping has become more feasible in recent years too but still requires special hardware and software.

What is so silly about the SACD vs PCM debates is that DSD was literally intended to be converted to PCM. Its sample rates are multiples of 44.1.
DSD was supposed to be the archival format, that would be seamlessly rendered to PCM for consumer media.

But then 'someone' got the bright idea instead to market a highly copy protected consumer medium for DSD itself...SACD.
 

polmuaddib

Senior Member
Forum Donor
Joined
Mar 1, 2020
Messages
479
Likes
852
Ripping your DVDAs to 'backup' files is simple, and has been for many years now, with relatively inexpensive, legal software like DVD Audio Extractor. There is no sonic degradation.

SACD ripping has become more feasible in recent years too but still requires special hardware and software.

What is so silly about the SACD vs PCM debates is that DSD was literally intended to be converted to PCM. Its sample rates are multiples of 44.1.
DSD was supposed to be the archival format, that would be seamlessly rendered to PCM for consumer media.

But then 'someone' got the bright idea instead to market a highly copy protected consumer medium for DSD itself...SACD.
You are right. But if you want to keep DVD-Audio structure to play in a DVD-Audio player, you are not removing watermark by ripping it to PCM format of any kind, be it FLAC or WAV. As soon as you make DVD-Audio again from those FLACs or whatever, watermark is active again. The only way to remove it is to "cut" it out of spectrum, which is also cutting out possible music material. The other way is converting to DSD and making SACD-R.
My point is that DSD can be useful sometime. Not much, of course.
BTW, those DSF and DFF files are DSD uncompressed, right? There are no DST compressed files? Other then DST in SACD structure... Seems a waste of space, if there is a lossless compression for DSD.
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,700
Location
Hampshire
Okay so, I have the Passion of The Christ album by John Debney. It shows a cutoff which would indicate an upsample, but this is from supposedly a 44.1Hz release. That would indicate it was created at even lower sample rates than 44.1 (I know this is a thing in video games for exmaple, didn't expect it from soundtracks). Also most music that throws me off doesn't actually have a true "abrupt" end.
Could you post a picture of that spectrum?

What my claim is, you can't actually tell all the time.
I thought you were asking a question. What can't one tell all the time?

As for bit depth, yeah how would I figure that out, if something was converted up or down for example (more interested in down obviously just like in the sample rate ordeal).
If a 16-bit recording is placed in a 24-bit container, the low 8 bits will be all zeros.

Also, I'm wondering. How one would differentiate a 44.1 16 bit flac, vs a 44.1 mp3 as another thing that crossed my mind. Of course this assumed MP3 V0 or 320kbps.
MP3 (or some other perceptual coding) can be suspected if there are holes in the spectrum, i.e. black patches on the spectrogram. It's not a very reliable test, though.

I've seen releases of the same song vary spectrally even with bit depth and sample rates being the same (both FLAC of course). And I was wondering how that sort of thing occurs.
It's common to release several masterings for different targets. Sometimes there are alternate mixes (often as "bonus" tracks). Re-releases of old material frequently differs in one way or another. Lots of possible reasons.
 
Last edited:

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
Could post a picture of that spectrum?


I thought you were asking a question. What can't one tell all the time?


If a 16-bit recording is placed in a 24-bit container, the low 8 bits will be all zeros.


MP3 (or some other perceptual coding) can be suspected if there are holes in the spectrum, i.e. black patches on the spectrogram. It's not a very reliable test, though.


It's common to release several masterings for different targets. Sometimes there are alternate mixes (often as "bonus" tracks). Re-releases of old material frequently differs in one way or another. Lots of possible reasons.

Yeah, so something like this.

As for bit-depth, I specifically said I was more interested in knowing how would one know if it was 24 down to 16.

With respect to masterings for different targets, these are all "audiophiley" tracks (from Amber Rubarth). They don't really sound like different masterings, just seems as if dynamic range is messed with, and that's about it. You say re-releases of old material differs, I can see that. Can't see why especially since it's not explaining if and why a certain mastering was done, and especially can't understand how one would know it's even occurring without listening to multiple. Going back to the whole audio industry's garbage standards all around (or their non-existence) in many respects.
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,700
Location
Hampshire
That's looks like normal music plus some noise around 18 kHz, most likely picked up from some electrical equipment in the vicinity. In the last third or so of the track, the background noise is a little higher. They probably made some change to the recording equipment there, perhaps enabled another microphone feed.

As for bit-depth, I specifically said I was more interested in knowing how would one know if it was 24 down to 16.
If you have a 16-bit file, there is no way to tell if some earlier production stage used a higher resolution.
 

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
That's looks like normal music plus some noise around 18 kHz, most likely picked up from some electrical equipment in the vicinity. In the last third or so of the track, the background noise is a little higher. They probably made some change to the recording equipment there, perhaps enabled another microphone feed.


If you have a 16-bit file, there is no way to tell if some earlier production stage used a higher resolution.

Ah okay, thanks mansr. I had no idea what any of that stuff was. I've seen that sort of line sometimes when looking at MP3's in the past as well (but it's not as thick, instead it's like a thin but very distinct line), so I thought it had something to do with format types (but was confused seeing as how the spectrum I showed was from native FLAC).

Do you know of any approximations of what a lossy to lossless spectrum anomaly's look like with example? (I know you said "holes" and such, but I don't know if I'd be looking at it, even if it was in front of my face).
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,700
Location
Hampshire
Ah okay, thanks mansr. I had no idea what any of that stuff was. I've seen that sort of line sometimes when looking at MP3's in the past as well (but it's not as thick, instead it's like a thin but very distinct line), so I thought it had something to do with format types (but was confused seeing as how the spectrum I showed was from native FLAC).
In somewhat old recordings, it's common to see a tone caused by the horizontal sync of some nearby CRT monitor. Switching power supplies are another possible source of such noise. In your picture, it's probably something using spread spectrum switching, hence the multiple parallel lines.

Do you know of any approximations of what a lossy to lossless spectrum anomaly's look like with example? (I know you said "holes" and such, but I don't know if I'd be looking at it, even if it was in front of my face).
Sure, compare these two:
orig.png


mp3.png
 

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
Just wondering, what are the formats being shown here? And what is technically going on when this is being shown (like what portions are being cut out from the lossy?) Or is this something one would only know if being told what the encoder properties were (like format, bit-rate etc..?).
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,636
Likes
7,497
In somewhat old recordings, it's common to see a tone caused by the horizontal sync of some nearby CRT monitor.

When I first started loading up ripped CDs and other digital sources into audio editors, I was surprised at how many had telltale lines across the spectrograms around 15.7kHz - CRT monitors in-studio somewhere. The tone is all over a lot of '80s music in particular, which makes sense because a number of synthesizer setups had CRT monitors as part of them. It's also interesting to see the slight variations in the precise frequency where the tone shows up in sources that were originally recorded to analogue tape - gives a clue to tiny speed variations in the original recorder and/or the playback deck used to make the digital transfer.
 
  • Like
Reactions: Tks

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,700
Location
Hampshire
Just wondering, what are the formats being shown here? And what is technically going on when this is being shown (like what portions are being cut out from the lossy?) Or is this something one would only know if being told what the encoder properties were (like format, bit-rate etc..?).
Top is 30 seconds of a FLAC from Mark Waldrep downsampled to 44.1 kHz. Bottom is the same encoded as mp3. The missing (black) bits are whatever the encoder deemed inaudible, including everything above 20 kHz.
 

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
Top is 30 seconds of a FLAC from Mark Waldrep downsampled to 44.1 kHz. Bottom is the same encoded as mp3. The missing (black) bits are whatever the encoder deemed inaudible, including everything above 20 kHz.

Ah okay, so in the file I presented, because there is some content above the 20kHz range (in the form of hazy signal clouds above 20kHz in the latter part of the song especially) I can rest assured I am not getting something resampled from lossy or some oddity like that?

One final thing as I overstay my questionaire welcome. If you encode a 24-bit file, to 16 bit non-dithered. What sorts of reprocussions (if any) does that have with respect to audibility or even visually, spectrum-wise? I know you said there's no way to tell this occured, but if it you had side-by-side images of each, would differences be visible with any particular setting in the spectrum generation?
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,636
Likes
7,497
Absolutely. PAL line whistle, 15.625kHz!

Oh, right, forgot about that! NTSC is 15.734kHz - so you can also get a clue as to where the recording was made based on what the CRT flyback frequency is on the source! :)
 

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
https://hydrogenaud.io/index.php?topic=108864.msg948350#msg948350
My attempt to fake a hi-res file using a 44k source, with improvised >22k content.

Could've sworn I seen some of your stuff before. An exchange ages ago where someone claiming they can tell something was sampled a certain way or something and could tell just by looking at a spectrum, and then you put him to the test and he kinda failed after you gave the keys to unlock which files were which? Was that you or am I thinking of something/someone totally different?
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
Could've sworn I seen some of your stuff before. An exchange ages ago where someone claiming they can tell something was sampled a certain way or something and could tell just by looking at a spectrum, and then you put him to the test and he kinda failed after you gave the keys to unlock which files were which? Was that you or am I thinking of something/someone totally different?
Yes it was me. Asked someone to sort the bitrate of mp3 files based on spectral content. I tweaked the encoding parameters so that low bitrate files looked nice on spectrograms.
 
  • Like
Reactions: Tks

T3RIAD

Member
Joined
Sep 11, 2019
Messages
38
Likes
23
If DSD is better quality then why do audiophiles prefer R2R DACs, which look like PCM internally, as opposed to the sigma-delta DACs that look like DSD internally?

https://hydrogenaud.io/index.php?topic=108864.msg948350#msg948350
My attempt to fake a hi-res file using a 44k source, with improvised >22k content.

My Samsung phone has a function that does this. You turn on "bandwidth upscaling" in the sound settings and it brings the sample rate up to 192 kHz, adding a bunch of random UHF noise to make your music "hi-res."
 
Top Bottom