Topic: Effect of MD compression on .wav comp

[madman]

Those are my comments.  I can hear the difference easily.  The ATRAC doesn't suck, but it's clearly worse.  I don't have time to fill out the whole shebang.

Hi-SP is not bit perfect, just the oppositem it is another lossy compression algo.  The question presented by Madman is whether (or not) Hi-SP sounds better than old SP, which would be interesting to know IF one is going to use any lossy compression in the first place, which of course is undesirable for live recording except under exceptional circumstances (i.e., its still better than nothing)>

Actually, my question is whether or not the human ear can hear the difference between Hi-SP and uncompressed in a blind test. 

[tms]

Hi-SP is not bit perfect, but Hi-MD recorded as PCM and recovered by HiMDRenderer is.

Have you done any controlled testing with HiMDRenderer to verify its bit-accuracy?  I'm sure the author has improved it from when I last checked into it, but...last I looked into it HiMDRenderer did not prove bit-accurate.

I have not b/c I don't own one.  But from what I've read it is very very close to bit perfect.  Apparently the software recovers the data in 2 minute windows (those are bit perfect) and then puts it all together.  At those 2 minute intervals the software author says it may or may not be bit perfect.

Depends on how much of a purist you are I guess, budget, what you have on hand, what your objectives are. 

If I had one I'd use it for taping, they're smaller than a JB3 and I like the removable media.

[tms]

The problem is not whether HiMDRender is bit accurate.  I think it is, and *a small number of* my tests confirm this.  The problem is that the MD *resamples* the digital input.  So, for the purist, you need a NJB3 for digital recording.

Can you expand on this?  How does resampling change the data?

I thought I had isolated just the effect of compression but apparently the resampling was also included in the comp too.

[tms]

Actually, my question is whether or not the human ear can hear the difference between Hi-SP and uncompressed in a blind test. 

This comparison is essentially the same as a Hi-SP vs. wav test.  Reading online (http://forums.minidisc.org/index.php?showtopic=11644&hl=)  I found the specs of SP and Hi-SP.   ATRAC SP is 292 kbps,  Hi-SP is 256 kbps.

SP and Hi-SP are essentially the same, so this comparison should answer this.  If you want to repeat with the Hi-SP that would be good too.

[Brian Skalinder]

Hi-SP is not bit perfect, but Hi-MD recorded as PCM and recovered by HiMDRenderer is.

Have you done any controlled testing with HiMDRenderer to verify its bit-accuracy?  I'm sure the author has improved it from when I last checked into it, but...last I looked into it HiMDRenderer did not prove bit-accurate.

I have not b/c I don't own one.  But from what I've read it is very very close to bit perfect.  Apparently the software recovers the data in 2 minute windows (those are bit perfect) and then puts it all together.  At those 2 minute intervals the software author says it may or may not be bit perfect.

Depends on how much of a purist you are I guess, budget, what you have on hand, what your objectives are. 

If I had one I'd use it for taping, they're smaller than a JB3 and I like the removable media.

Thanks for the update, TMS.  "Very very close to bit-perfect", and "may or may not be bit perfect" does not equal bit perfect.  I'm just trying to ensure people have access to accurate information.  I know I'd be upset if I bought that solution thinking it's bit-perfect, when in fact it is not.  And I agree, whether the solution proves useful depends on the individual case.

[zowie]

Hi-SP is not bit perfect, but Hi-MD recorded as PCM and recovered by HiMDRenderer is.

Have you done any controlled testing with HiMDRenderer to verify its bit-accuracy?  I'm sure the author has improved it from when I last checked into it, but...last I looked into it HiMDRenderer did not prove bit-accurate.

I have not b/c I don't own one.  But from what I've read it is very very close to bit perfect.  Apparently the software recovers the data in 2 minute windows (those are bit perfect) and then puts it all together.  At those 2 minute intervals the software author says it may or may not be bit perfect.

Depends on how much of a purist you are I guess, budget, what you have on hand, what your objectives are. 

If I had one I'd use it for taping, they're smaller than a JB3 and I like the removable media.

Thanks for the update, TMS.  "Very very close to bit-perfect", and "may or may not be bit perfect" does not equal bit perfect.  I'm just trying to ensure people have access to accurate information.  I know I'd be upset if I bought that solution thinking it's bit-perfect, when in fact it is not.  And I agree, whether the solution proves useful depends on the individual case.

Well, I'm thinking about buying a HiMD to run mic in strictly for unobtrusive recording.  So resampling isn't really an issue for me.  Bad transfer software could be.  A comparison of how much quality is lost with the new HiSP vs. PCM would be nice to know in order to make an informed judgment about using it on occassions where the set is too long to fit at 16/44.

[tms]

Well, I'm thinking about buying a HiMD to run mic in strictly for stealthing.  So resampling isn't really an issue for me.  Bad transfer software could be.  A comparison of how much quality is lost with the new HiSP vs. PCM would be nice to know in order to make an informed judgment about using it on occassions where the set is too long to fit at 16/44.

If you record 16/44.1 PCM using the analog in, SONY allows you to get your wav files off the HiMD bit perfect over their software.  Analog in = no problem.

Digital recordings are the problem with SONY.  The transfer software, HiMDRenderer only comes into play if you record optical in.  SONY doesn't want you to be able to get the wavs directly from the HiMD.  HiMDRenderer supposedly gets around this, but may not be bit perfect.  So the transfer software problem of not being bit perfect only comes into play if you're trying to record dig-in.

What are you going to compare the HiMD A/D>compressed audio to?

[tms]

Thanks for the update, TMS.  "Very very close to bit-perfect", and "may or may not be bit perfect" does not equal bit perfect.  I'm just trying to ensure people have access to accurate information.  I know I'd be upset if I bought that solution thinking it's bit-perfect, when in fact it is not.  And I agree, whether the solution proves useful depends on the individual case.

It may be more appropriate to say that it hasn't been proven not bit perfect yet.  The only real testing I've read about is right here in this thread and the he says it was bit perfect in all his tests.

This brings up another even more off topic question, is there such a thing as a bit perfect recorder in a real world scenario?  It seems like I've read on ts.com that supposed bit perfect recorders like the JB3 have occasionally dropped bits.

Personally I don't even care, I don't own a HiMD and probably never will.  I just thought it would be fun to put up a comp using an MD and watch people smash it out of the park.  Seems like it would be fun.  So far no takers though! 

[Brian Skalinder]

It may be more appropriate to say that it hasn't been proven not bit perfect yet.

I believe it's wise to err on the side of caution, i.e. assume it's not bit-perfect until proven so.  If we assume it's bit-perfect and it isn't, the consequences are flawed recordings.  If we assume it is not bit-perfect and it is, then there are no downsides.
 

is there such a thing as a bit perfect recorder in a real world scenario?

Yes.  My D100 was bit-perfect - I tested it.  My JB3 setup of V3 > Hosa ODL-312 > optical > V3 and V3 w/optical mod > optical > JB3 proved bit-perfect in a large sampling of test data.  I know because I tested it.  I'm in the process of testing my MT2496 now, too.

It seems like I've read on ts.com that supposed bit perfect recorders like the JB3 have occasionally dropped bits.

The JB3 proves bit-perfect with some gear, and not with others.  Usually, the culprit has been a digital format converter, and not the JB3 proper.

[Zaphod]

is there such a thing as a bit perfect recorder in a real world scenario?

Yes.  My D100 was bit-perfect - I tested it.  My JB3 setup of V3 > Hosa ODL-312 > optical > V3 and V3 w/optical mod > optical > JB3 proved bit-perfect in a large sampling of test data.  I know because I tested it.  I'm in the process of testing my MT2496 now, too.

What is the testing process to ensure bit-accuracy?

I haven't seen it mentioned, and I'm curious about it.

[Brian Skalinder]

What is the testing process to ensure bit-accuracy?

Basically two ways:

OPTION 1

• play back a digital file from PC, through a bit-perfect soundcard, into the recording device
• transfer the newly recorded file from the recording device back to the PC
• perform a comparison¹ of the two WAVs - they should match identically, sample for sample
  

.
or

OPTION 2

• record a file from a single ADC output to two recording devices, a known bit-perfect recorder (A) and the recorder in question (B)
• transfer both A and B recordings to a PC
• perform a comparison¹ of the two WAVs - they should match identically, sample for sample

.
Of course, in both cases, we need a known bit-perfect device.  IMO, option 2 is better because some gear may prove bit-perfect when paired with certain gear, but not with others.  So, we should test with the gear we use to record in the field. 

¹ EAC, WaveLab, and other apps have WAV compare features.

Link to a specific test case I outlined for the CO2:

http://taperssection.com/index.php?topic=36146.msg463275#msg463275

[zowie]

Well, I'm thinking about buying a HiMD to run mic in strictly for stealthing.  So resampling isn't really an issue for me.  Bad transfer software could be.  A comparison of how much quality is lost with the new HiSP vs. PCM would be nice to know in order to make an informed judgment about using it on occassions where the set is too long to fit at 16/44.

If you record 16/44.1 PCM using the analog in, SONY allows you to get your wav files off the HiMD bit perfect over their software.  Analog in = no problem.

Digital recordings are the problem with SONY.  The transfer software, HiMDRenderer only comes into play if you record optical in.  SONY doesn't want you to be able to get the wavs directly from the HiMD.  HiMDRenderer supposedly gets around this, but may not be bit perfect.  So the transfer software problem of not being bit perfect only comes into play if you're trying to record dig-in.

What are you going to compare the HiMD A/D>compressed audio to?

To HiMD A/D non-compressed audio (16/44 pcm) and to old MD-SP compressed audio.

Re transferrs -- fine for me.  I'm not going to use MD with a front end ADC.  If I'm able to bring a larger package w/ outboard gear package I'd be brininging my NJB or laptop. 

[madman]

Actually, my question is whether or not the human ear can hear the difference between Hi-SP and uncompressed in a blind test. 

This comparison is essentially the same as a Hi-SP vs. wav test.  Reading online (http://forums.minidisc.org/index.php?showtopic=11644&hl=)  I found the specs of SP and Hi-SP.   ATRAC SP is 292 kbps,  Hi-SP is 256 kbps.

SP and Hi-SP are essentially the same, so this comparison should answer this.  If you want to repeat with the Hi-SP that would be good too.

ATRAC SP is indeed 292kbps, but it's using an older form (I think the oldest) of ATRAC and is available on the older MD format, Hi-MD formats don't even support this ancient encoding.  Hi-SP uses ATRAC3plus and is specific to Hi-MD technology only.  Bitrate to bitrate comparisons are not apples to apples when dealing with different encoding mechanisms.  In short, SP and Hi-SP are nowhere near the same.

[tms]

Actually, my question is whether or not the human ear can hear the difference between Hi-SP and uncompressed in a blind test. 

This comparison is essentially the same as a Hi-SP vs. wav test.  Reading online (http://forums.minidisc.org/index.php?showtopic=11644&hl=)  I found the specs of SP and Hi-SP.   ATRAC SP is 292 kbps,  Hi-SP is 256 kbps.

SP and Hi-SP are essentially the same, so this comparison should answer this.  If you want to repeat with the Hi-SP that would be good too.

ATRAC SP is indeed 292kbps, but it's using an older form (I think the oldest) of ATRAC and is available on the older MD format, Hi-MD formats don't even support this ancient encoding.  Hi-SP uses ATRAC3plus and is specific to Hi-MD technology only.  Bitrate to bitrate comparisons are not apples to apples when dealing with different encoding mechanisms.  In short, SP and Hi-SP are nowhere near the same.

So newer = better?  You sure you want to stick with that theory?

But great, another comp!  Do a SP to HiSP comparison and find out.  I bet you wouldn't be able to tell the difference between these two compressions in a blind test.

I was basing my statement that SP ~ HiSP on this from minidisc.org:

-------------------------------------------------

Ok so here's how it goes.

As far has HiMD is concerned Atrac 256kbps (also known as HiSP) is the best Atrac sounding compression you can use. While sonicstage does encode Atrac into 320kbps these are as of yet not supported on HiMD.

Here is a list of supported bitrates and my opinion on them.

PCM (1411kbps): CD quality, this has no compression and is by far the best sound quality available.
-My rating: ***** (5/5)

Atrac3+ HiSP (256kbps):Highest quality Atrac compression. Very comparable to the original Atrac (which had a bitrate of 292kbps), this compression is almost impossible to tell apart from the original PCM source.
-My rating: ***** (5/5 because the whole point of compression is to let you store more audio in the same space and for all intensive purposes this sounds the same as PCM)
<snip of lower bitrates>

[madman]

Actually, my question is whether or not the human ear can hear the difference between Hi-SP and uncompressed in a blind test. 

This comparison is essentially the same as a Hi-SP vs. wav test.  Reading online (http://forums.minidisc.org/index.php?showtopic=11644&hl=)  I found the specs of SP and Hi-SP.   ATRAC SP is 292 kbps,  Hi-SP is 256 kbps.

SP and Hi-SP are essentially the same, so this comparison should answer this.  If you want to repeat with the Hi-SP that would be good too.

ATRAC SP is indeed 292kbps, but it's using an older form (I think the oldest) of ATRAC and is available on the older MD format, Hi-MD formats don't even support this ancient encoding.  Hi-SP uses ATRAC3plus and is specific to Hi-MD technology only.  Bitrate to bitrate comparisons are not apples to apples when dealing with different encoding mechanisms.  In short, SP and Hi-SP are nowhere near the same.

So newer = better?  You sure you want to stick with that theory?

If you took a moment to read the improvements of ATRAC3plus over ATRAC, you might understand why.  I don't currently have access to a Hi-MD deck, but I remember it sounding much better than my old MD decks that use SP/ATRAC.