Topic: Tascam HD-P2 - Oade ACM v. Stock: gain / noise measurements

[Brian Skalinder]

I did some testing this afternoon for kicks.  FWIW...


TEST INFO

I started off with a simple test to identify the gain differences between the HD-P2 stock and Oade Advanced Concert Mod.  I also performed frequency response analysis to measure noise in a few spots along the way.  Here's how I did it:

• Generated a 1 kHz tone in Samplitude SE at 32bfp / 48 kHz
• Looped playback, Foobar2000 / M-Audio Audiophile 24/96 S/PDIF-out > Bel Canto DAC 1.1 > HD-P2s
• Recorded mic-in, no pad, with gain knobs all the way down
• Recorded mic-in, no pad, with various "numbered" (0-10) gain settings
• Repeated above two bullets with the -20 dB pad engaged

I may do the same for line-in, but I'm working without a preamp right now and couldn't generate a strong enough signal to perform the test line-in.


GAIN MAP

For each test, I played back the same 1 kHz tone in Foobar2000 on my PC.  As a starting point, I applied max gain on both units (10 on the dial) and adjusted Foobar's output volume until each unit hit 0 dBFS.  I then turned the gain all the way down on the each HD-P2 (0 on the dial).  Finally, I started recording, adjusting the gain on each unit in steps, stopping at each number on the dial for ~10 seconds.  I then measured the gain in Samplitude.  The results:

====================================================================
       STOCK / OADE ADVANCED CONCERT MOD HD-P2, MIC-IN, NO PAD
--------------------------------------------------------------------
HD-P2   |           STOCK              |            ACM
--------------------------------------------------------------------
Foobar  |           -43.5 dB           |          -32.5 dB
--------------------------------------------------------------------
Dial #  |  +dB Total  |  +dB / Dial #  |  +dB Total  |  +dB / Dial #
====================================================================
     0  |       0.00  |             -  |       0.00  |             -
     1  |       0.62  |          0.62  |       0.72  |          0.72
     2  |       3.22  |          2.60  |       3.14  |          2.42
     3  |       7.12  |          3.90  |       7.12  |          3.98
     4  |      13.49  |          6.37  |      13.11  |          5.99
     5  |      17.96  |          4.47  |      17.64  |          4.53
     6  |      21.75  |          3.79  |      21.27  |          3.63
     7  |      26.46  |          4.71  |      25.50  |          4.23
     8  |      31.02  |          4.56  |      30.47  |          4.97
     9  |      36.79  |          5.77  |      36.80  |          6.33
    10  |      47.76  |         10.97  |      46.47  |          9.67

Now with the -20 dB pad engaged (it seems the -20 dB pad is actually about -22 dB):

====================================================================
  STOCK / OADE ADVANCED CONCERT MOD HD-P2, MIC-IN, -20 PAD ENGAGED
--------------------------------------------------------------------
HD-P2   |           STOCK              |            ACM
--------------------------------------------------------------------
Foobar  |           -21.5 dB           |          -11.0 dB
--------------------------------------------------------------------
Dial #  |  +dB Total  |  +dB / Dial #  |  +dB Total  |  +dB / Dial #
====================================================================
     0  |       0.00  |             -  |       0.00  |             -
     1  |       0.64  |          0.64  |       0.74  |          0.74
     2  |       3.28  |          2.64  |       3.28  |          2.54
     3  |       7.54  |          4.26  |       7.06  |          3.78
     4  |      13.88  |          6.34  |      13.20  |          6.14
     5  |      18.38  |          4.50  |      17.59  |          4.39
     6  |      22.26  |          3.88  |      21.32  |          3.73
     7  |      26.98  |          4.72  |      25.40  |          4.08
     8  |      31.55  |          4.57  |      30.46  |          5.06
     9  |      37.86  |          6.31  |      36.81  |          6.35
    10  |      47.14  |          9.28  |      46.37  |          9.56


NOISE MEASUREMENTS

I also measured the frequency response (with Adobe Audition 1.5 using 65,536 FFT Blackmann-Harris) for both the stock and ACM HD-P2 mic-in, linear and log, with and without the pad.  I measured at no gain, about 1/2 gain (in dB, or 6 on the dial), and full gain (10 on the dial).

UNIT  | COLOR
=======================
ACM   | GREEN
STOCK | RED

Mic-in, no pad

Gain at 0

• linear
  
• log
  
  

Gain at 6

• linear
  
• log
  
  

Gain at 10

• linear
  
• log
  
  

Mic-in, -20 dB pad

Gain at 0

• linear
  
• log
  
  

Gain at 6

• linear
  
• log
  
  

Gain at 10

• linear
  
• log
  
  

CONCLUSION

So what does all this mean?  Well, unless my testing methodology is fundamentally flawed (which is entirely possible), I think it means:

<01>  Both the stock and ACM HD-P2 offer roughly the same gain range:  about 0 to +47 dB.

<02>  The ACM HD-P2 requires a stronger signal (about +11 dB relative to stock) to achieve the same dBFS level; in other words, applying the same strength signal to both the stock and ACM HD-P2 will produce levels ~11 dB lower on the ACM HD-P2.

<03>  The ACM HD-P2 offers a lower noise floor with no gain, some gain, and especially at high gain.

<04>  I can more easily adjust gain on the fly now, knowing how the gain knob dial #s correspond to dB.

[jackmf]

Thank you Brian...that was most informative and helpful! No suprise there!

[macdaddy]

thanks brian.

[twatts (pants are so over-rated...)]

Freekin' A Right!!!  Way to go Brian!!!

+Ts for the next month or so... 

T

[Roving Sign]

Just so I understand - these are the Foobar settings that produced 0dbFS?


Foobar  |           -43.5 dB           |          -32.5 dB


adjusted Foobar's output volume until each unit hit 0 dBFS

[Brian Skalinder]

Just so I understand - these are the Foobar settings that produced 0dbFS?

Yes, that's correct.  Same signal played back in Foobar for both units.  The stock unit needed a volume setting of -43.5 dB to achieve 0 dBFS, while the ACM unit needed a volume setting of -32.5 dB (~11 dB louder) to achieve 0 dBFS.

[spzkt]

Hi Brian

I realize you have moved on from the HD-P2 but I have a query your testing:
 
In the noise graphs do the 1000hz tones peak at 0dBfs for both units? It's a bit hard to see this from the graphs, and I'm wondering if the 11dB difference in noise floors is primarily due to the 11dB difference in gain between the units?

cheers
Paul

Just so I understand - these are the Foobar settings that produced 0dbFS?

Yes, that's correct.  Same signal played back in Foobar for both units.  The stock unit needed a volume setting of -43.5 dB to achieve 0 dBFS, while the ACM unit needed a volume setting of -32.5 dB (~11 dB louder) to achieve 0 dBFS.

[Brian Skalinder]

I realize you have moved on from the HD-P2 but I have a query your testing:
 
In the noise graphs do the 1000hz tones peak at 0dBfs for both units? It's a bit hard to see this from the graphs, and I'm wondering if the 11dB difference in noise floors is primarily due to the 11dB difference in gain between the units?

Good question, Paul.  While the dBFS value differed depending on the amount of gain applied in each of the noise graphs, the gain applied was basically the same (±1 dB) and the dBFS level achieved was basically the same (±1 dBFS) for both stock and ACM (stock first / ACM second):

MIC-IN, NO PAD

Dial #   |  Gain (dB)        |  dBFS
========================================================
 0 /  0  |  +00.00 / +00.00  |  -46.95 / -48.50
 6 /  6  |  +21.75 / +21.27  |  -26.75 / -25.68
10 / 10  |  +47.76 / +46.47  |  -00.74 / -00.66

MIC-IN, -20 dB PAD

Dial #   |  Gain (dB)        |  dBFS
========================================================
Dial #   |  Gain (db)        |  dBFS
========================================================
 0 /  0  |  +00.00 / +00.00  |  -47.80 / -46.75
 6 /  6  |  +22.26 / +21.32  |  -25.54 / -25.43
10 / 10  |  +47.14 / +46.37  |  -00.66 / -00.38

I think this doesn't answer your question, though hopefully it provides more insight into how I performed the tests.  As my notes below indicate, the starting 1kHZ signal was ~11dB louder for the ACM test, even though I applied the same amount of gain, and achieved the same dBFS levels for both units.  In hindsight, I also should've tested with a fixed 1kHz signal, in which case, for example, the stock HD-P2 would've required adding +22.26 dB of gain to achieve -25.54 dBFS, while the ACM HD-P2 would've required +33.26 dB of gain to achieve the same dBFS.

[spzkt]

Does this answer your question?

Hi Brian,

thanks for taking time to respond! 

I think I've worked out why I was confused by the graphs.  The 11dB reduction in overall gain on the ACM gives the appearance of the mic pre having an 11dB lower noise floor, when in fact the noisefloor is probably altered by a very small amount.  If you feed an 11dB higher signal into the ACM this increases signal to noise by 11dB.

A more accurate comparison of noise floors might be to feed both units the same level test signal, then boost the resulting file/output from the ACM by 11dB?

Please correct me if I'm making false assumptions but this would seem to explain the 11dB difference in noise floors shown in your graph.

I'm actually really interested in this stuff as I'm waiting on delivery of an Oade modded HDP2 and I'm curious as to what kind of improvements the mods actually bring.

cheers
Paul

[Brian Skalinder]

A more accurate comparison of noise floors might be to feed both units the same level test signal, then boost the resulting file/output from the ACM by 11dB?

Please correct me if I'm making false assumptions but this would seem to explain the 11dB difference in noise floors shown in your graph.

I'm not sure that would be accurate.  I think that assumes every 1 dB of gain = 1 dB of increased noise, which I'm not sure holds true.  In hindsight, what I should have done is fed both units the same level test signal (i.e. a 1 kHz tone at a fixed amplitude) and applied gain on both stock and ACM to achieve the same dBFS.  This would have required, for example, +20 dB of gain on the stock unit, and +31 dB of gain on the ACM unit.  I think that would most closely represent a real-world test of the noise floors, as it also takes into account the gain structure.

[spzkt]

Brian you are right.

While I think using 11db would be a reasonable approximation, setting gain as you suggest is going to be far more accurate. The gain adjustments on the HDP2 are done in the digital domain (except the -20dB pad which is analog) so the resulting files would reflect signal to noise of the analog stages.

cheers
Paul