Parks Audio Waxwing measurements

[Balle Clorin]

Overload happens in the input stage governed by the gain. Your ADC level "A" is at -7.0/-6.6 (good).
Your digital level "D" after processing (eq and volume) is a -0.7 (fine but close).
Audio Technica MC OC9MLii measured to give 0.35mV at 200 Ohm . Using either 61db gain og 41+20 from SUT. Or 39db on Shure v15 IV Jico SAS-B. Original Shure spec was 4mV

Strongly agree with this post. Looked at it again. Can you confirm the gain/cart for this reading?

Audio Technica MC OC9MLii measured to give 0.35mV at 200 Ohm . Using either 61db gain or 41+20 from SUT. Or 39db on Shure v15 IV Jico SAS-B. Original Shure spec was 4mV. Should maybe reduce MM to 36? Since my Jico seems hotter than 4mV .. need to check that
EDIT:My Shure+jico gives 4.5 mV.
So at 38db gain I have 15 db headroom verified by measurement. With 38db gain om Shure I get same output as with 60dB OC9( 0.35mV measured)

 

[undeakyetonde]

EDIT: I am still looking into how gain affects the Puffin noise floor..

Looked at that just recently (for Waxwing)

https://www.audiosciencereview.com/...un-with-vinyl-measurements.20278/post-2070778

tl;dr for MM cartridges lowering gain does raise noise floor, amount depends on frequency region
e.g. for -10 dB gain we get 10 dB higher noise floor <1 kHz, but the difference diminishes as we go to 20 kHz

 

[mppix]

Audio Technica MC OC9MLii measured to give 0.35mV at 200 Ohm . Using either 61db gain or 41+20 from SUT. Or 39db on Shure v15 IV Jico SAS-B. Original Shure spec was 4mV. Should maybe reduce MM to 36? Since my Jico seems hotter than 4mV .. need to check that

I'm running 36 with the VM740ML (4mV MM). It is rare to get an actual clip and only with problematic records.

I was asking because your MC analog noise (A) was ~10dB below (edit: read "better than") mine (I take the measurements with turntable spinning and stylus over the record without contact, Pufffin bandwidth fully open 0 to 48kHz).

I think I understand now that this is due to the 200Ohm (MC) vs the 47k (MM) impedance. However this does not immediately translate into noise levels at the output (D) due to the higher MC gain. I am now very tempted to get a MC cartridge to play with

 

[mppix]

Looked at that just recently (for Waxwing)

https://www.audiosciencereview.com/...un-with-vinyl-measurements.20278/post-2070778

tl;dr for MM cartridges lowering gain does raise noise floor, amount depends on frequency region
e.g. for -10 dB gain we get 10 dB higher noise floor <1 kHz, but the difference diminishes as we go to 20 kHz

Seems to confirm the strategy: choose max gain that does not clip -> adjust volume to match desired line level
Do I read this correctly, the measurement reads > 80dB SNR without using trickery like mono-bass?

 

[undeakyetonde]

Seems to confirm the strategy: choose max gain that does not clip -> adjust volume to match desired line level
Do I read this correctly, the measurement reads > 80dB SNR without using trickery like mono-bass?

Depends on what you mean by SNR

For Waxwing with 38 dB gain, connected to a Audio-Technica VM95ML cartridge, RMS of the noise floor measures approximately:

When bandlimited to 25 Hz - 48 kHz (4th order Butterworth lowpass filter at 25 Hz from Waxwing, a brickwall 48kHz highpass from its ADC?):
61 dB relative to 5 cm/s rms 1 kHz sine
69 dB relative to max RMS level of a 45 RPM 12" record I have (post-RIAA)
79 dB relative to peak of the same record

When bandlimited to 220 Hz - 22 kHz (like Michael Fidler suggests) using a brickwall filter:
69 dB relative to 5 cm/s rms 1 kHz sine
75 dB relative to max RMS level of a 45 RPM 12" record I have (post-RIAA)
86 dB relative to peak of the same record

Max RMS level - 'peak' of RMS computed for 50 ms segments
Kind of "peak instantaneous power" but less instantaneous than peak sample value, so more appropriate for estimating 'common sense' SNR

For regular 33.3 rpm 12" albums SNR would be lower

 

[Thomas_A]

With the previous posts using the CA-TRS1007 record, the 1 kHz signal is 3.54 cm/s peak (L or R). Giving around -69 dBA for the L+R track for L channel and -75 dBA when combined to mono (cancelling vertical signal). It is 5 cm/s peak for the L+R but I just guess this translates to 3.54 cm/s L or R.

TRS-1007

 

[morillon]

from denon doc
;-)

 

[Thomas_A]

Depends on what you mean by SNR

For Waxwing with 38 dB gain, connected to a Audio-Technica VM95ML cartridge, RMS of the noise floor measures approximately:

When bandlimited to 25 Hz - 48 kHz (4th order Butterworth lowpass filter at 25 Hz from Waxwing, a brickwall 48kHz highpass from its ADC?):
61 dB relative to 5cm/s rms 1 kHz sine
69 dB relative to max RMS level of a 45 RPM 12" record I have (post-RIAA)
79 dB relative to peak of the same record

When bandlimited to 220 Hz - 22 kHz (like Michael Fidler suggests) using a brickwall filter:
69 dB relative to 5cm/s rms 1 kHz sine
75 dB relative to max RMS level of a 45 RPM 12" record I have (post-RIAA)
86 dB relative to peak of the same record

Max RMS level - 'peak' of RMS computed for 50 ms segments
Kind of "peak instantaneous power" but less instantaneous than peak sample value, so more appropriate for estimating 'common sense' SNR

For regular 33.3 rpm 12" albums SNR would be lower

See also

Phono preamp headroom - why?

Not for me. Until recently, the pre I was using was the cheapest I'd been able to get at a local store - Project Phono Box. In July I got the waxwing - and even with the "Magic" feature turned off noise (particularly pop/crackle) is (seems) much lower - or at least much less intrusive. To the...

www.audiosciencereview.com

RMS level of the 33 RPM music record was about the same as the RMS 3.54 cm/s (peak) reference; -22 to -23 dB.
Max RMS 10 ms window for the music was at -12 to -13 dB (10 dB higher the the reference)
Peak for the music was at about -3 dB (20 dB higher than the reference)

 

[Thomas_A]

from denon doc
;-)

Yep.

 

[Balle Clorin]

I'm running 36 with the VM740ML (4mV MM). It is rare to get an actual clip and only with problematic records.

I was asking because your MC analog noise (A) was ~10dB blow mine (I take the measurements with turntable spinning and stylus over the record without contact, Pufffin bandwidth fully open 0 to 48kHz).
View attachment 389936
I think I understand now that this is due to the 200Ohm (MC) vs the 47k (MM) impedance. However this does not immediately translate into noise levels at the output (D) due to the higher MC gain. I am now very tempted to get a MC cartridge to play with

Not a problem to use MC vs MM in my Puffin.
With lo cut 20hz high 30k,derumble on I have this evening -85dB noise floor with MM at 38db, and -82-83dB at MC with 60db gain /200ohm. With a 20dB SUT+40db gain I get -84.5 db noise floor
D :noise
With 38 and 60db I get same signal level with MM and MC. 3.54cm/s= MM 4.5mV gives -19.5db….

At 36dB and MM have -87 on both A and D, never seen -99 on anything

 

[mppix]

Not a problem to use MC vs MM in my Puffin.
With lo cut 20hz high 30k,derumble on I have this evening -85dB noise floor with MM at 38db, and -82-83dB at MC with 60db gain /200ohm. With a 20dB SUT+40db gain I get -84.5 db noise floor
D :noise
With 38 and 60db I get same signal level with MM and MC. 3.54cm/s= MM 4.5mV gives -19.5db….

At 36dB and MM have -87 on both A and D, never seen -99 on anything

~10dB below meant "better than" mine.
I regularly get about A=85dB with 36dB 47k MM (turntable rotating, no filters). I agree that a 200Ohm MC should result in a similar (slightly worse) noise floor.
However, your original post had A at -95dB (?) .. if true with 60dB gain, I'm very tempted to try this MC cartridge.

The D=-99.9dB is because I run all of my system at 24/96 with -24 LUFS approximate reference level. I currently adjust the Puffin digital volume but could as well reduce gain.
Probably not relevant to most people.

 

[AaronJ]

I suggested in the Waxwing thread as a test case to take an unedited recording of a track with known clicks that the magic function suppresses. Feed that digital track through a DAC and into the Waxwing with gain set to 0 or -6, whatever, and no RIAA. See if the device can detect the difference between an actual click and a recorded click.

 

[Balle Clorin]

A Recorded click will Per definition not be detected I think. But I do not know how the algorithm works, maybe a higher level of a click will cause it to be classified as a click to correct

 

[antcollinet]

A Recorded click will Per definition not be detected I think.

Since "magic" takes place in the digital domain - anything it detects has effectively already been recorded.

 

[mppix]

I suggested in the Waxwing thread as a test case to take an unedited recording of a track with known clicks that the magic function suppresses. Feed that digital track through a DAC and into the Waxwing with gain set to 0 or -6, whatever, and no RIAA. See if the device can detect the difference between an actual click and a recorded click.

It is probably easier to detect (and remove) clicks that saturate one or a few ADC samples. It is probably more complex but I'm curious if that is part of it.

 

[pogo]

Magic is probably not always magic.
I just listened to my JMJ with Magic and I immediately noticed the lack of higher frequencies in the bass drum. By switching Magic on/off directly, the difference was clearly audible. Attached are the recordings I made for my observation, including the delta audio, which is suppressed by Magic. In this case, Magic off is the better choice.

Waxwing settings:

I have heard title/track 3.
Blue = Magic off
White = Magic on

 

[AaronJ]

That is indeed interesting. Could you send that to Shannon so he can investigate? Might only be so much that can be done.

 

[pogo]

That's what I'm going to do

 

[pogo]

Shannon looked at my observation and has already replied:
I plan to spend most of 2025 doing R&D to improve Magic and take it to the next level. It’s my number one priority going forward.

Thanks to Shannon and I'm looking forward to the improvements and the new functions like Resonance, ...
I am already enthusiastic about Waxwing and if there is a further improvement, ...

 

[WDeranged]

Shannon looked at my observation and has already replied:
I plan to spend most of 2025 doing R&D to improve Magic and take it to the next level. It’s my number one priority going forward.

Thanks to Shannon and I'm looking forward to the improvements and the new functions like Resonance, ...
I am already enthusiastic about Waxwing and if there is a further improvement, ...

That is very good news. If Magic ever approaches SugarCube quality there will be a good amount of people ready to jump ship.

It's a fantastic device but support from the manufacturer is sketchy.