• 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!

Minimum Phase vs Linear Phase

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,383
Likes
24,749
Location
Alfred, NY

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,029
Likes
10,797
Location
São Paulo, Brazil
OP
Ron Texas

Ron Texas

Master Contributor
Joined
Jun 10, 2018
Messages
6,078
Likes
8,914

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,029
Likes
10,797
Location
São Paulo, Brazil
OP
Ron Texas

Ron Texas

Master Contributor
Joined
Jun 10, 2018
Messages
6,078
Likes
8,914

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,029
Likes
10,797
Location
São Paulo, Brazil
What I like about the free Resampler-V plugin for foobar is that it has these graphs that show both frequency and time domains while I set up the parameters.

Through some trial and error I came to this setting below, which attenuates to -144 db (24 bit noise floor), is leaky on high frequency but has almost perfect no ringing impulse response. I use this to upsample my files offline, and then playback them back in high res. Never did blind test but for me snares sound much cleaner with this.

Would have been soooo much easier to have these on JRiver for on-the-fly upsampling...

SoX Resampler-V settings.png


https://sourceforge.net/projects/resamplerv/
 
Last edited:

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
I'd trade less perfect impulse response for lower image amplitude but it's a preference thing. I cannot hear that high anyway and don't want the images doing something wonky down the line.
 

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
Some words about SoX and others resampling. If the vinyl ripping on 24/192 does not have an exceptional sound I convert it to 24/96 FLAC 0 from foobar2000 (I do not like as a player but yes like a tool, a lot) and also save space, on the internal disk and on the external backup. Usually, with 24/192 the sound is something unreal, too much soundstage? when I played them in my first system. Why? I do not know.

I do not like the sound with the SoX implemented in JRMC since ... always.

BTW, interesting your list of DAC with minimum phase and slow (my choice too). The logical thing is to assume that any modern DAC has the minimum and linear implemented in its two variants but...

https://yabb.jriver.com/interact/in...HPSESSID=d92m3jk4nn3sf3s1krtru594f5#msg785295
 
OP
Ron Texas

Ron Texas

Master Contributor
Joined
Jun 10, 2018
Messages
6,078
Likes
8,914
@Matias my experience with Resampler-V is it used a lot of CPU resources as compared to the Foobar plugin or JRiver implementations.
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
Some words about SoX and others resampling. If the vinyl ripping on 24/192 does not have an exceptional sound I convert it to 24/96 FLAC

I am sure you are aware that 12 bits at 44.1 kHz is enough for vinyl....
 

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,029
Likes
10,797
Location
São Paulo, Brazil
@Matias my experience with Resampler-V is it used a lot of CPU resources as compared to the Foobar plugin or JRiver implementations.
I use it offline. That is, I convert on foobar the original files to new upsampled files, so that afterwards during playback JRiver doesn't do anything. Waste of drive space though... Thanks, JRiver.
 
Last edited:

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,424
Likes
4,030
Location
Pacific Northwest
I am sure you are aware that 12 bits at 44.1 kHz is enough for vinyl....
Not necessarily true for all vinyl. My Cardas test LP had a frequency sweep with sound out to about 30 kHz, when I recorded it at 96 k (I used a low output MC whose FR extended well beyond 20 kHz). Also, quadraphonic LPs encoded information into the above 20 kHz range. So if you want to digitally capture ALL the info on an LP you may need to sample faster than 44.1 k.

This would be an interesting test: find an old quadraphonic LP in mint condition, and the decoder preamp. Play the LP on a top quality turntable having usable response beyond 20kHz. Make 2 digital recordings: one at 44-16, another at 96-24. When you play back these digital versions and feed them into the line-level input of the decoder, the 1st should sound like stereo, the 2nd should sound quadraphonic.

That said, of course only a tiny % of LP have anything above 15 kHz, and one can question whether anything higher is useful. So 16-44 would be more than sufficient for most LPs.
 
  • Like
Reactions: MWC

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
That said, of course only a tiny % of LP have anything above 15 kHz, and one can question whether anything higher is useful. So 16-44 would be more than sufficient for most LPs.

Indeed. Of course you could also record vinyl in 24 bits to capture all possible information - it is just that half of that is just noise. :)
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,424
Likes
4,030
Location
Pacific Northwest
Sure, the S/N of vinyl above 20 kHz is low. But it's not all noise. Quadraphonic LPs encoded information in that range and it actually worked. So for some LPs, it would be reasonable to use higher sampling frequencies when digitally recording.
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
Sure, the S/N of vinyl above 20 kHz is low. But it's not all noise. Quadraphonic LPs encoded information in that range and it actually worked. So for some LPs, it would be reasonable to use higher sampling frequencies when digitally recording.

But what is the point if you can't hear it?
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,424
Likes
4,030
Location
Pacific Northwest
My example answers that question. Maybe you have one of those vintage quadraphonic preamps and you want to digitally record that old quadraphonic Pink Floyd album lest you play it too many times and the needle ploughs flat the delicate supersonic frequencies encoding the quadraphonics. So it's digitally captured forever and you can listen at your leisure to its vintage 4-channel sonic glory.

Or, even if you don't have the vintage quadraphonic decoder, you want to ensure you're archiving all the information from that LP just in case you ever got one in the future. Then if your old uncle Joe dies, who was an audiophile in the 1970s, and you inherit his quadraphonic preamp, you're all set.

It's all in fun of course, but the point is that some LPs may have some non-noise content that 44-16 can't capture.
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
It's all in fun of course, but the point is that some LPs may have some non-noise content that 44-16 can't capture.

Sure, quadraphonic is a special case, but you wouldn't record every album in 192/24 just because a few of them are quadraphonic.
 

Guermantes

Senior Member
Joined
Feb 19, 2018
Messages
484
Likes
561
Location
Brisbane, Australia
All of this debate seems focused on the DAC reconstruction filter, but surely the "damage" is already done in the choice of anti-aliasing and/or oversampling filters used in the ADC. A simple analogue RC filter is not linear phase but a subsequent digital filter used in oversampling probably will be.

For those concerned about linear vs minimum reconstruction filters, are you also concerned about the type of filters used to capture acoustic performances? If not, why not?
 

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
http://archimago.blogspot.com/2015/07/the-linear-vs-minimum-phase-upsampling_10.html
2. There was a general tendency towards the minimum phase setting with those listening with speakers whereas headphone users seemed to skew more towards the linear phase setting. This brings up interesting questions about the differences between the sound presented through speakers (especially soundstage and imaging qualities) versus how sound is perceived through headphones (free of room interactions, lower channel crosstalk, mental integration of the stereo image). Perhaps the digital filter settings should be taylored to the type of listening...

Here: usually speakers in near field.

My two cents

We are facing a complex system in which many variants are involved. The tests must be in the same equipment and room to reduce the error. If the listening is also with speakers and far field, more complicated will be, and if we use Room EQ more too.

Therefore, the logical thing is to have both types of filters and other intermediates, letting the listener decide in each situation and type of music played. That is, there is no clear choice, it will depend on our circumstances.

And, in case of retouching the frequency response, it is better that it is not aggressively and not in the entire audible band. Better to have speakers with a fairly flat response and adjust without abusing, without looking for an impeccable graphic to show off in the forums.

Update: I am ALWAYS writing about music. Cinema is other world, where the sound is very manipulated and spectacularity is sought, so the use of Room EQ is usually the best in that case, and in all audible band. BTW, to listen to music (not musical PRODUCTS), better with Pure Direct in the AVR but, again, you need speakers with a reasonably flat frequency response and a not very problematic room.
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
Please maty, just hit the "edit" button instead of double posting a new sentence starting with "And". It just clutters things up.
 
Top Bottom