there is no "PURE" way to A/B compare, but what about active splits?

[Plush]

You "engineers" have really got my dander up now. I have proposed a legitimate traditional test and some people continue to see how it cannot work or cannot be fair.

In order to see how the traditional test might indeed provide a useable comparison, we have to assume certain things. It you are testing a Nagra VI and a Sound Devices recorder you can assume that both laudable and respected manufacturers have made sure that their machines have:

stable phantom power
stable clock
low noise
good mic amps
input circuitry that does not load the microphone--high input impedance mic amp

among other things

Let's say we are splitting two Schoeps CMC62 omni mics. Both are powered from an external 48 volt supply. We don't need to test the 48 volt supply of the machines because the manufacturer has already done this for us.

So we split the mics and we listen. This gives us a good idea of the performance of the mic and the individual machines.

IF we hear a gross frequency response difference between the machines, we begin to test our set up. We plug the individual mics in to each machine and hear a "reference"
pick up of the music. We audition each split to see if that is the problem.

WE DO NOT start out questioning a simple and relatively BENIGN splitter whose job and whose characteristics are well known. The BIAS against the transformer is specious. The BIAS against splitters is ignorant and specious.

I'm not angry about the objections, Peter. I see them as ignorant. Why don't you stop dumping on it and just give it a try?

I am using Wireworks splitters with Jensen transformers. They are a high quality unit that doesn't change the sound. Smart people designed them and smart testers easily
overcome your objections.

[ISedlacek]

Slightly OT but I cannot help:

Peter, do you sometimes make (or even enjoy) some real music ??

If yes (and if you are not just an enthusiastic theoretician), have you even selected some equipment for it that you like ? But it is very hard to imagine, because how can you ever prefer something over something else, when it is next to impossible to compare things ... Whatever difference you hear is just an illusion and imperfection and is caused by 100 000 very serious factors that should be very scientifically checked and considered. Seems that in ideal test conditions, everything should sound exactly the same. If we still hear a difference, there must be some test error somewhere ... So in the end, it must be very easy to select the gear for you - since it really does not matter what you take ... any difference is just illusory and is always a result of imperfect conditions ... Right ?

But if you really make music, please, send us some examples ... It must sound quite incredible with all this unmatched discrimination and setting the conditions that are more perfect than perfection personified )

[MichaelPatrick]

Originally Posted by John Willett

Not quite - as the question arises "how do you power the mics"?

If you power the mics from one recorder and not the other you are skewing the results as one recorder does not have to supply the extra power.

Alternatively - if you use a 3-way split using the third way for power only and no machine supplying phantom power, then you are not testing to see if the extra current drain of providing phantom power affects the sound (although the two machines would be operating under identical conditions).

Any recorder that changes sonically because it IS or IS NOT supplying phantom power is not worth having. Do you really think Sound Devices or Nagra will be found in such a category of suspects?

P.S. I agree with Hudson and Ivo. Imposing laboratory-grade controls on recording engineers who just want to see how a recorder's preamps and converters sound is impractical at best and foolish at worst. Controls need to be adequate for the purpose, and the goal here is not to sustain human life on the moon or calibrate a neurosurgical instrument. Please, this silliness is all out of proportion.

[Audiop]

Originally Posted by MichaelPatrick

Any recorder that changes sonically because it IS or IS NOT supplying phantom power is not worth having. Do you really think Sound Devices or Nagra will be found in such a category of suspects?

That is something that is investigated by a proper test. Not guesswork or relying on reputation.

P.S. I agree with Hudson and Ivo. Imposing laboratory-grade controls on recording engineers who just want to see how a recorder's preamps and converters sound is impractical at best and foolish at worst. Controls need to be adequate for the purpose, and the goal here is not to sustain human life on the moon or calibrate a neurosurgical instrument. Please, this silliness is all out of proportion.

What's silly about making a proper test? I think it's professional.

Yes, controls need to be adequate for the purpose and I assume the purpose is to learn something about the gear, no?

Time and time again flawed tests are set up which leads to false results and conclusions.


/Peter

[Audiop]

Originally Posted by ISedlacek

Slightly OT but I cannot help:

Peter, do you sometimes make (or even enjoy) some real music ??

Are you serious with that question or just feeling good being rude?

I've sensed bad attitudes from you for some time and I ask you to stop it and focus on the things being discussed.

If yes (and if you are not just an enthusiastic theoretician), have you even selected some equipment for it that you like ?

Of course. Why would I pick equipment that I dislike??

But it is very hard to imagine, because how can you ever prefer something over something else, when it is next to impossible to compare things

That is your words. I think it's very possible most of the time to compare things in a proper way. We need to be a little more specific here Ivo.

... Whatever difference you hear is just an illusion

Which is easily ruled out with blind testing.

and imperfection and is caused by 100 000 very serious factors that should be very scientifically checked and considered.

When doing a proper test you go thru the method and make sure as few unkown variables as possible interfer with the test. Why wouldn't you? Lazyness? Ignorance? Fear?

Seems that in ideal test conditions, everything should sound exactly the same.

Are you serious or just provoking? A roaring tiger and a frog sounds the same because the conditions is "ideal" or controlled?

If we still hear a difference, there must be some test error somewhere ...

What makes you think that? Sounds very confused to me.

So in the end, it must be very easy to select the gear for you - since it really does not matter what you take ... any difference is just illusory and is always a result of imperfect conditions ... Right ?

I don't know if you had a bad day but you don't make any sense at all Ivo.

Sometimes it's easy to select gear and sometimes not, of course! Of course the gear matter what makes you think I was of another opinion? Or just making trouble and being provocative again?

But if you really make music, please, send us some examples ... It must sound quite incredible with all this unmatched discrimination and setting the conditions that are more perfect than perfection personified )

Nice Ivo, nice!

So if you approve the recordings I've made my statements and knowledge that I share have some merit but if you feel it sounds so-so or outright bad I'm a lost case?

The performers, the room and the mic's and the positioning of them is what makes a nice sounding record.

As I've told you earlier I looped your excellent sounding "Agnus dei" thru my main recording gear and the quality was identical to me after a loop thru DA-micpre-AD.
If the gear of the type "transmission links" can pass on a signal without (significant) audible alteration the gear gets thumbs up from me.

You didn't wanted me to post it remember? I can send it you you privately so you can hear for yourself.

If you want to take this further I ask you to leave "me" out of the discussion and focus at the things being discussed and please also be a little more specific.

Learning the difference between subjective and objective, being specific, being polite, not putting words in others mouth and skipping personal attacks is good things for contstructive discussions.


/Peter

[MichaelPatrick]

Peter, of course tests are necessary, but the confusion here is between what recording engineers do and what manufacturers like Nagra and Sound Devices to in their R&D and QA labs to design and build quality products. When I buy an expensive recorder I don't expect it to have defective mic power. No one should.

This kind of testing surely has value, but it may be more useful to people in the geekslutz forum. This is a forum where people discuss recording on location. I, for one, don't want to retest the phantom power circuits in my recorders. You should consider taking this testing crusade to manufacturers who don't meet your R&D standards. It may do some real good there.

[Audiop]

I'm a location recordist and I test my gear because I want to know my gear and I want to learn about correlation between audible and measured performance.

Gear that do not audibly add or subtract anything to the signal is what I want and the only way to get such gear is via well performed tests.

Also people at this forum and other do test their gear or propose tests.. that's when I help with an analyze. I don't tell people to do any tests. I do recommend that IF you put in the effort to do a test, than do it right.

Why some individuals feel they need to ridicule me on the basis of the above is beyond me.


/Peter

[mixedupsteve]

I think all of the tests proposed are all valid meaningful tests.
Passive splitter
Active splitter
Matched pair direct.
I would be willing to bet that all of these tests would result in similar results as to which recorder is favored.

Anybody want to test the tests??

[didier.brest]

Originally Posted by mixedupsteve

Anybody want to test the tests??

Here a test of the LA Audio MS424 active mic splitter.

2 x Schoeps MK21 -->2 x 3m Monster Performer 500 cables--> MS424, 2 x 2 x transformer balanced outputs ---> 4 x 5m Sommer Galileo cables

then

1) DAVBG1 35 dB gain --> 2 x 2m Monster Performer 500 cables --> MR816X with pad and minimum gain

2) MR816X gain setting at 12:00

Gain adjustment in DAW between -2 and +2 dB.

[James Lehmann]

I'm very interested in this thread - particularly as it pertains to my personal crusade about establishing the way different mic cables present sound differently.

I've already done some shoot-outs with matched Gefell mics and matched pre-amps, swapping only cables and posting the WAVs.

The results of that were perfectly conclusive to me personally, but I've purchased a passive mic splitter for my next round of tests in order to address the next row of sceptics.

Anticipating the arguments about phantom power (see various posts above), I'd planned to conduct my test with an SM7B which doesn't require 48v. I've even purchased a special Neutrik connector that will allow me to connect the SM7B directly to the splitter box, removing any influence a 'head-cable' might have on the test (although I can't see why the effect wouldn't be identical for the two test cables further down the line).

Can anyone (apart from Peter, whose predictably contrarian stance is usually guaranteed to grind any interest in threads like this into shards of de-contextualised multi-quotes) present any reasonable objections to this methodology for an aural shoot-out of cables:

Dynamic mic >> Passive splitter >> Trim-linked dual pre-amps

[didier.brest]

Yes there is one that was exposed (as an objection to my proposal identical to your one about using a passive splitter) in the thread from which this one was created by the admin: a passive splitter doest not ensure a convenient isolation of both preamps, that is connecting a preamp B to the splitter output may affect the sound of another connected preamp A. What is untrue with an active splitter.

[roonsbane]

I have tested preamps many times with three way splitters and it just works. Put a decent Phantom supply on the primary and the two preamps for testing on each of the isolated legs of the splitter. A mic splitter's isolated output is designed to be loaded into a microphone preamp. Many times in a busy recording career you will be asked to take a feed from a transformer isolated output. If it's a decent transformer, and well implemented, and you see a decent preamp (or whole console) on the primary, it will be the least of your worries. Move on and do your job. You have to assume with good to great gear, that it can supply needed current.

This is by far the most reliable, most transparent way to A-B mic preamps.

To tell if your preamp is not supplying enough current, listen to the preamp individually under extreme full range signal conditions. Luckily I haven't used gear that crappy in a very long time.

Cameron

[didier.brest]

I compared two cables:

Pure Sound. 10 m length. It's a brand from Stagg. Low cost (a few euros). It seems there is now only a 15m length microphone cable available from Stagg.

Vovox unshielded. 5 m length.

Microphones: Schoeps CMC6-MK21. Cables to the mic splitter: Monster Performer 500, 3 m length. Preamp and ADC: MR816X.

The Pure Sound track was about 1 dB quieter, which I compensated digitally, separately on left (+ 1.15 dB) and right channel (+ 0.9 dB).

[Brent Hahn]

Originally Posted by didier.brest

I compared two cables:

Pure Sound. 10 m length. It's a brand from Stagg. Low cost (a few euros). It seems there is now only a 15m length microphone cable available from Stagg.

Vovox unshielded. 5 m length.

Microphones: Schoeps CMC6-MK21. Cables to the mic splitter: Monster Performer 500, 3 m length. Preamp and ADC: MR816X.

The Pure Sound track was about 1 dB quieter, which I compensated digitally, separately on left (+ 1.15 dB) and right channel (+ 0.9 dB).

I do hear a difference. The thing the two recordings have in common, though, is a bunch of harmonic distortion. Seems to me that fussing over mic cables in this context is like treating a hangnail on a guy with a bullet wound.

[mixedupsteve]

Originally Posted by didier.brest

Here a test of the LA Audio MS424 active mic splitter.

2 x Schoeps MK21 -->2 x 3m Monster Performer 500 cables--> MS424, 2 x 2 x transformer balanced outputs ---> 4 x 5m Sommer Galileo cables

then

1) DAVBG1 35 dB gain --> 2 x 2m Monster Performer 500 cables --> MR816X with pad and minimum gain

2) MR816X gain setting at 12:00

Gain adjustment in DAW between -2 and +2 dB.

Interesting, I like them both. It's not exactly what was being proposed though.

[Mark D.]

The DAV sounds better. Any shootouts with DAV vs. Gordon?

[didier.brest]

Originally Posted by Mark D.

The DAV sounds better. Any shootouts with DAV vs. Gordon?

I agree. Would you send to me a Gordon ?

What do you think about the other test Vovox vs crappy cable ?

[Mark D.]

Originally Posted by didier.brest

I agree. Would you send to me a Gordon ?
What do you think about the other test Vovox vs crappy cable ?

I would if I could but I can't so I won't. If I ever get one, I will. I heard improvements with the Vovox. There are several botique cables I've tried, they always sound better than Monster, which always sounds better than Mogami which always sounds better than regular cable. Then there is Zaolla which sounds better at first, but it has a 'tizziness' you can't mix out of it, making it unusable. The cost is worth it for folks like me who can hear a real difference, however slight, in the 'a lot of little improvements add up to one big improvement' theory group.

[didier.brest]

Originally Posted by Mark D.

I heard improvements with the Vovox.

I did not...

[mobius.media]

Originally Posted by didier.brest

I compared two cables:

Pure Sound. 10 m length. It's a brand from Stagg. Low cost (a few euros). It seems there is now only a 15m length microphone cable available from Stagg.

Vovox unshielded. 5 m length.

Microphones: Schoeps CMC6-MK21. Cables to the mic splitter: Monster Performer 500, 3 m length. Preamp and ADC: MR816X.

The Pure Sound track was about 1 dB quieter, which I compensated digitally, separately on left (+ 1.15 dB) and right channel (+ 0.9 dB).

I ran a null on the two with audio Diff Maker (which auto adjusts for volume) and the piano is still very clearly audible in the leftover. So either there is a difference between the cables, your splitter imparted something to one and not the other, or Diffmaker had a problem with the correcting the levels of each channel independently.

Not sure, but there's definitely some of the piano part left behind after. I'll split them to L/R and run each channel separate later if I have time. Or you could of course do the same if you like.

[didier.brest]

Originally Posted by mobius.media

I ran a null on the two with audio Diff Maker (which auto adjusts for volume) and the piano is still very clearly audible in the leftover.

Yes I did that already in Wavelab. That's why the level of both samples (Vovox and Pure Sound) are balanced within 0.01 dB. I just have adjusted again the right level and I've got left and right RMS levels for the difference of both sample equal to -89 dBFS and -93 dBFS and peak levels equal to -70 dBFS and - 72 dBFS . Yes by cranking up the playback gain, I can ear clearly the piano. But I could not tell which one is the Vovox by listening at both samples.

[mobius.media]

Originally Posted by didier.brest

Yes I did that already in Wavelab. That's why the level of both samples (Vovox and Pure Sound) are balanced within 0.01 dB. I just have adjusted again the right level and I've got left and right RMS levels for the difference of both sample equal to -89 dBFS and -93 dBFS and peak levels equal to -70 dBFS and - 72 dBFS . Yes by cranking up the playback gain, I can ear clearly the piano. But I could not tell which one is the Vovox by listening at both samples.

I find this interesting because I frequently hear people say "there is no audible difference between mic cables" or "there is no measurable difference between mic cables significant to the audio spectrum".

In this case, I think one could still debate whether or not the difference is important or worth it, but you could not debate that there is some aspect of music that is recorded with one and not the other.

I suppose theoretically this could still be an artifact of your splitter and not the cable - I don't know - but some difference is there either way. Thanks for sharing.

[dfahb44]

Sure there is. Comparing pre's is easy.


Step 1. Get a guitar amp
Step 2. Mic it up
Step 3. Record dry DI signal
Step 4. Reamp
Step 5. Unplug from preamp A, plug into preamp B, match the levels.

[Audiop]

Originally Posted by mobius.media

I find this interesting because I frequently hear people say "there is no audible difference between mic cables" or "there is no measurable difference between mic cables significant to the audio spectrum".

If the R-C-L and screen are close there is no proof to date that two cables impart different sound. I think it wold be cool if someone could prove it but I'm not liked in these thread anyway.. :-)

I suppose theoretically this could still be an artifact of your splitter and not the cable - I don't know - but some difference is there either way. Thanks for sharing.

Well this is the spec's of the splitter:

Frequency response: Equivalent Input Noise: THD: Gain:
Input Impedance: Maximum Input Level: Output Impedance: Load Impedance:
BLOCK DIAGRAM
20Hz to 20kHz (+0, -0.5dB) > -121dBu < 0.01 % (1kHz, 150R load) 0dB (-20dB with Pad)
2k 0dBu (+20dBu with Pad) < 50R >150R

Yes, it's very likely that a difference IS from the splitter. The only way to find out is to measure the two splitter sections and see how close they are.

Step no2 is a blind ABX or similar controlled test with significant statistical results.


/Peter

[James Lehmann]

Originally Posted by didier.brest

I compared two cables:

The Pure Sound track was about 1 dB quieter, which I compensated digitally, separately on left (+ 1.15 dB) and right channel (+ 0.9 dB).

Didier,

If you have added digital gain to one file and not the other, then the test is asymmetrical.

If there is a gain differential between the two cables then it can never be totally symmetrical in terms of preparing the files for a listening test that evaluates only tone and not volume. HOWEVER, you can level the playing field by at least applying some make-up gain to both clips - albeit differing amounts - to level match them; at least then they have both undergone an additional step of processing.

In terms of the splitter itself, it's easy enough to repatch and repeat the test to eliminate any linear differences between outputs, but for a decent quality brand like LA Audio, speaking for myself I am quite prepared to accept that it is doing it's job sufficiently well to not be the significant factor affecting the tone of each split.

In terms of an ABX test that reduces bias in open Forum listening tests, a number of us are contributing ideas to the development of just such a test that, crucially, includes an element of online data gathering making it much harder to skew results - do please follow this link to this thread and weigh in with your ideas on how this might work.

[didier.brest]

James,

I don't think that applying 1 dB digital gain to a signal would cause some audio quality loss that could be expained theoretically, except if this gain causes clipping. Theoretically the 1 dB difference between both signal levels causes a 1 dB difference between the signal to quantization noise ratio values. The quantization noise RMS level being below -100 dBFS for an ideal 16 bit coder, this difference should not be significant.

I think that the better way for balancing both levels is not the one I choose nor the one that you are proposing; it would be to apply a negative gain to the stronger signal because this would balancing also the quantization noise levels.

But this test is not symetrical by nature: it does not aim to compare two cables but to decide whether a high end cable provides an audible quality improvement with respect to a low end cable. The first unbalancing factor is that the length of the low end cable is twice the length of the high rend cable. The second one, not significant IMO, is the positive gain correction applied to the low end signal.

Despite of these factors, there is no difference audible to me. This personal conclusion is reinforced by the the measurement showing that the difference signal RMS level is below -90 dBFS. Good news for my cable budget !

[James Lehmann]

Originally Posted by didier.brest

Despite of these factors, there is no difference audible to me. This personal conclusion is reinforced by the the measurement showing that the difference signal RMS level is below -90 dBFS. Good news for my cable budget!

Fair enough Didier - these sorts of tests are far more useful to folks who do them themselves; you've done your tests and reached your own conclusions rather than relying exclusively on what it's possible to infer from the views or WAVs of others.

Interestingly, when I did my tests with Vovox cables I reached a different conclusion from you, ie that the cables were making a difference to what arrived at my A-D converter. (And I will be re-doing this test soon with a splitter rather than matched mics to investigate further.)

I've also done some tests with different gain plug-ins and I felt there was a difference between adding digital gain or not, as well as between plug-ins, for example between Logic Gainer plug-in and Metric Halo Channel Strip. These were very subtle though, barely enough to worry about, but worth taking into consideration IMHO and hence why I pointed it out. "Never turn your back on digital" as Bob Ludwig is quoted as saying in Bob Katz's book 'Mastering Audio' (Ed.1, p198).

[DeeVox]

I too borrrowed a vovox and did a shootout with a couple of matched mics, and the difference was clear. I immediately bought one. then I tried a pair for my speakers. I didn't even need to A/B, after 10 seconds I was thinking "what the... did it use to sound this good?" No it didn't and the cables stayed in place.
In my ears they are worth the money. The difference is not HUGE, but it's there and the extra definition they supply is not possible to add in mix.

[Audiop]

Would you mind sharing the test protocol and possibly some files?


/Peter

[audiologist]

Hi,

this is a very good test here, during the last weeks I came in contact with VOVOX many times, always wondering if there is a differece or not.

I calculated a difference Signal from the both files above - and with normal listening level there was nothing audible (with good headphones).
Only if I normalized the difference file I was able to hear a noisy piano signal (like the user above) - but I would say it is not possible to notice that differences with normal listening levels (regardless any masking effects).

From my side I am very open to learn that a cable makes a difference. I have also some cables in my Monitoring-setup where I would say that there is a difference. bu in the case of the piano recording I would say both recordings are equal.

Best Regards
audiologist