Preliminary notes


Last summer I performed a blind listening comparison between three different audio formats, all set for ~175 kbps encodings. The purpose of the test was to investigate about encoding quality with classical music (and only classical) and to see which format would be the most efficient (i.e. the closest to transparency at lowest bitrate possible) for this kind of music. As jumping-off place for bitrate I took MPC –standard preset which was indisputably recognized as the best encoding solution outputing at 175...190 kbps on average. And indeed, the test ended on musepack superiority. MPC was even superior to Vorbis and MP3 at presets presenting higher bitrate (~195 kbps for LAME, ~185 for Vorbis against ~175 for musepack). Consequently, MPC encodings appeared to sound better and to be smaller at the same time. Amazed by the existent gap between all contenders I conclude my specific test with these words: “I didn’t think that MPC –standard was so in advance”.

My vacation are now quite over. I performed during my free time a big checkup of lossy quality at 80 kbps and 96 kbps (this one has to be translated in english ), and it’s too late now to complete the 128 kbps I planned to do in the same silly conditions (150+35 tested samples). But I used my small remaining time to do again the listening test at 175 kbps I did last year, with the same 18 samples and the same hardware.


Why doing again the same test?


As a result of constant evolution of most audio encoders I consider my previous results as really outdated. I recall that Vorbis encodings were done with MEGAMIX I (hybrid encoder melting aoTuV beta 2, Garf Tuned 2 and Quantum Knot tunings). This encoder didn't subsist for a long time... and doesn’t exist anymore; it was replaced by MEGAMIX II, then official 1.1 with Impulse Trigger Profile + Impulse Noisetune switches, which was finally followed by aoTuV beta 3 and beta 4. The same goes for LAME: 3.97 alpha 3 was tested, and during this time LAME developers have submitted eight new versions of this alpha and a few other ones (lame_may, lame_june...)! MPC has also changed: from 1.14 beta to 1.15 alpha which is now considered as safe to use.
As a consequence of this evolution, problems audible last years (kind of ringing for LAME, or noise and coarseness for Vorbis) may be corrected or at least be lowered. The first purpose of my test is therefore to check the outcomes of recent tunings done for high bitrate settings.

There’s also a second point which stimulated me to do again the test and this point is called AAC. I haven’t tested AAC last year for technical and moral reasons. Technically, iTunes encoder couldn’t be set to ~175 kbps; Apple's AAC encoder wasn't also gapless and is for my purpose unsuitable for my conception of artefact-free encodings. I also felt as dishonest the inclusion of Nero AAC: it had recognized issues with classical first and a new encoder supposed to solve these problems was announced as imminent. Some readers suggest me to include faac as competitor, but I felt as unfair to test an encoder which was probably not the state of the art of AAC format and to compare it to the most advanced implementation of other formats (MEGAMIX and LAME 3.97).
I never regret my choice. But this absence of AAC frustrated my curiosity for a long time, because I had strictly no idea about comparative performance of this format with other contenders. That’s why I decided to absolutely include AAC this time. WMAPro will also be tested this time if possible.

The purpose of my test is therefore to obtain a fresh photography of the current performance of all modern lossy formats with classical music using the most advanced implementations for each of them.


I. Choosing the encoders


My purpose being to test most advanced encoders the choice of format hasn't to be controversial for most of them:

• MP3: LAME 3.97 alpha 11. Release date: July 2005. Note: --vbr-new encoding mode.
• MPC: mppenc 1.15v. Release date: march 2005.
• Vorbis: aoTuV beta 4. Release date: June 2005, updated in July 2005 (merged with SVN 1.1.1).
• WMAPro: no choice here: it's 9.1 or nothing. Release date: during 2004.

Choosing the good AAC encoder is much harder:
• Apple AAC: There's still no VBR mode with iTunes. Consequently it's currently impossible to use Apple's AAC encoder unless other contenders will output an average bitrate close to either 160 kbps or 192 kbps. It's unlikely...
• Nero Digital AAC: the most advanced VBR AAC encoder and therefore better placed to represent the AAC format. Big problem: should I use the 'high' and defaulted encoder or rather the 'fast' one which is really better at lower bitrate with classical music? The first one is still recommended by all Nero's developers and it's a valid reason to choose it instead of something they don't consider as stable enough (Garf, JohnV and Ivan Dimokovic). But the situation has maybe changed since their recommendation; I wouldn't also discard too quickly the possibility of using an encoder working better for the difficulties proper to the musical genre I'll test. The debate could be endless if a trivial but objective argument hadn't close the debate: the average bitrate of VBR mode of both encoders (see below).
• faac AAC: testing faac might also be interesting. And even for fun, it would give me the possibility to oppose four different open-source implementations of four different formats But such friendly comparison has a price: increasing the onerousness of the test which is anything but easy at this bitrate...


II. Targeting a bitrate


The purpose of my test is not to see what encoders could do with xxx kbps for each sample; I don't plan to force each encoding reaching a precise bitrate. My purpose is to stay close to the real usage of a vast majority of listeners (if not all...): using for every encoding one fixed setting which should statistically corresponds on average to the desired bitrate. That's why it's really fundamental to precisely know the average bitrate corresponding to a defined preset. And there's only one way to get it: encoding several tracks or albums.
Last year, I used as reference ~20 classical (+3 non-classical) albums. This year, I decided to be more methodical. I’m now using 150 different tracks (I mean full tracks) coming from 150 different CD in order to increase the variety of encoded tracks. It’s important to note that I didn’t choose randomly those tracks. I meticulously worked to get a representative microcosm of my full classical library, balanced between different grand ensemble (vocal, orchestral, chamber, soloist recording). This collection is nothing more than the 150 full tracks from which I’ve extracted 150 short samples in order to build a “catalogue raisonné” of musical situations occurring with classical music (see this test).

I genuinely expect from this methodically constructed library to be a highly representative panel of my classical collection. My assumption could be verified by checking the average bitrate of the entire collection encoded with WavPack -fx5 (all my >1000 CD digital library is encoded with this preset): 642 kbps for the selection of 150 tracks against 635 kbps for a complete set of more than 15000 tracks. The deviation is inferior to 1%! Statistics are really magical.


III. Observing bitrates


I started with MPC which must give the reference bitrate. All other competitors have to be set in order to get a similar value.


• MPC: --quality 5 corresponds precisely to 184,54 kbps. This is higher from what I expected first (~175 kbps). The 150 reference tracks are maybe not as representative as supposed. I also tried 1.14 (used last year) with the same preset and --xlevel: 176,28 kbps, much closer to the native average bitrate of --standard profile and reassuring me about the representativity of my collection of tracks. The bitrate has therefore inflated by 4.7% from 1.14 to 1.15v with classical.
=> I'll therefore try to get from all other encoders a setting which outputs to 184,5 kbps ±2% (180,5...188,1 kbps).

• MP3: I first tried -V2 --vbr-new, which corresponds to the former --preset fast standard. Average bitrate is 181,79 kbps. Now, this value is lower from what I estimated last year (and that's why I tested -V3 in addition to -V2)... Indeed, 3.97alpha3 -V2 would output to 192,99 kbps. Nice gain (-5.80%). Obviously LAME developpers also worked on efficiency. Gain is great enough that LAME --preset standard could now be fairly compared to MPC --standard. But I recall another time that it only applies for classical (I suppose that bitrate is higher with other musical suffering from sb21 issue).

• Vorbis: aoTuV beta 4 -q6,00 leads to 181,48 kbps. This is lower than what I expected, and it's also lower than MPC --standard bitrate. I get 186,99 kbps for the old MEGAMIX I. Bitrate has therefore be lowered with latest aoTuV.
-q6,00 could therefore be directly compared to MPC --standard and LAME --preset fast standard (for classical music).

• WMAPro: VBR75 leads to 150,24 kbps. The next available preset is VBR90 and it leads to 203,96 kbps. Both are very far for the range I fixed and consequently WMAPro can't compete in this test.

• Nero Digital AAC: Like LAME and WMAPro Nero Digital doesn't offer any precise VBR scale but seven presets. -internet leads to ~142 kbps for both 'high' and 'fast' encoders. -streaming high corresponds to 176,14 kbps and -streaming fast to 193,33 kbps. Consequently none of them is inside the fixed range; the closest one is -streaming high and is therefore the less unacceptable solution (I recall that the 'high' encoder is still the recommended one).

• faac AAC: this is the only encoder able to fit into the fixed bitrate range (thanks to the precise VBR scale alla vorbis & mpc). AAC faac –q 175 leads to 180,92 kbps. This –q setting won’t probably correspond to 180 kbps with other musical genre and that’s the occasion to recall another time that the whole test is specific to classical music and nothing else.


Recapitulative table



=> faac, LAME, aoTuV are very close each others (difference is inferior to 0,9 kbps!). MPC presents a higher bitrate (+3 kbps) and Nero Digital a lower one (-5 kbps). The gap between the extreme is worrying: approximately 5% corresponding to 8 kbps. That's not a huge difference but these eight missing kbps may lead to a significant difference in quality. I could discard Nero Digital for this test but I would consider this choice as a mistake. For my own curiosity I'm also very impatient to see how would perform an advanced implementation of AAC in comparison to other formats, even if bitrate are not fully comparable.

=> As a consequence I decided to test both Nero Digital AAC and faac AAC, and I will consider Nero Digital presence as a "bonus" interesting to watch rather than an entire competitor. That's why my final diagramme (plots) will graphically separate Nero AAC results from other contenders. I hope this will avoid unecessary debate about any kind of unfairness based on bitrate disparity.



SUMMARY

Are going to be test:

• AAC: faac 1.24.1. Release date: end 2004 (?). Setting: -q175
• AAC: Nero Digital aacenc32 v.3.2.0.15. Release date: June 2005. Setting: -streaming (high/default encoder).
• MP3: LAME 3.97 alpha 11. Release date: July 2005. Setting: -V2 --vbr-new
• MPC: mppenc 1.15v. Release date: march 2005. Setting: --quality 5
• Vorbis: aoTuV beta 4 based on 1.1.1. Release date: July 2005. Setting: -q6,00



IV. Additional information


I performed all my last listening tests on a Creative Audigy 2 soundcard, which resamples everything to 48000 KHz. Some people consider that internal resampling (transparent in my opinion) is treating unfairly musepack and would biased any listening test. To cut the controversial short, I installed my (better) Terratec DMX6Fire 24/96 which doesn't resample 44100 KHz files (I'm not using it anymore for daily listening because of interference with my VIA chipset).

HARDWARE & SOFTWARE SETTINGS:

• soundcard: Terratec DMX6Fire 24/96
• headphone: BeyerDynamic DT-531
• amp: Onkyo MT-5
• software player: Java ABC/HR 0.5 beta 5.
• software decoder: foobar2000 0.83 (in order to automatically get files free of offset and to solve my incompatibility issues occuring with Vorbis).

TESTING PRINCIPLES:

• ABX phase : To limit the listening fatigue and to end the test before I left my appartment, I restricted the ABX tests to the most transparent encodings (note > 4.00).
• Number of trials : eight trials as a minimum. I recall that schnofler's ABC/HR software doesn't reveal to score until the test is closed by the user (and it also can't be resume). Therefore the number of trails hasn't to be fixed: as long as score is hidden the pval isn't ruined. That's why I add more trials when I suspect bad results. I never exceed 16 trials: if something is really transparent I didn't persecute the encoding
• Notation : My notation was very severe last year, with a full dynamic range of notation (a lot of notes were inferior to 2.0). That's why I decided to add 10 points to each score (in order to disconnect the notation from the usual corresponding scale). This year, I tried to respect the ITU scale. When a difference is audible but not really annoying, the notation is at least equal to 4.0 and my hairs must stand on end to allow a notation inferior to 2.0 (from "annoying" to "very annoying"). Notation is still severe (I keep in mind that all encodings were set at 180 kbps) and that's why results I get here can't absolutely not be compared to other listening tests I done, especially those performed for low bitrate settings. By the way, there are no anchors in this test (high anchor is of course unecessary here).
• Samples: Same as last year. See this thread.
• Gain: I hadn't modify the gain of any file. All were played at their original volume.

V. Results and detailed comments







Sample 01: Krall
Short description: the only non-classical sample (Jazz). Cymbals, drums and voice.
Possible problems: smearing on cymbals and drums, distorted cymbals.
replaygain_sample_gain: = -5.50 dB (indicative only: files were tested at their original volume)



Cymbals are still a problem for lossy encoders at this bitrate: often smeared and sometimes distorted. Nero Digital is the worst¹, followed by LAME which suffers from strongest pre-echo; then faac. Contrary to other encodings MPC doesn’t have any smearing issue, but cymbals don't sound true. aoTuV is the best: I really had to insist in order to unmask the encoding, but I totally missed the ABX phase.
¹ Nero Digital allocate much less bits (150 kbps) than competitors, all at more than 200 kbps.



Sample 02: Fuga
Short description: harpsichord.
Possible problems: pre-echo, out of tune (tremolos, vibrating notes).
replaygain_sample_gain: +14.97 dB (indicative only: files were tested at their original volume)




Not a surprise for me: MP3 and especially LAME has still weakness with this instrument, which is distorted by the encoding (notes are trembling) and also has smearing issue. Nero Digital is excellent on the beginning despite of smallest bitrate (163 kbps); but quality suddenly drops on the second half. Faac is a bit better on average. Both aoTuV and MPC are excellent here, with only a very subtle smearing. MPC¹ was easier to ABX due to a distortion occurring on one short moment; for aoTuV, I changed the tested range during the ABX phase in order to catch an easier segment for comparison. Pval is > 0,05 < 0,10 on overall but < 0,05 on the second range I tested.
¹ very high bitrate (230 kbps) for MPC, used to bloat the bitrate with solo harpsichord.



Sample 03: Mahler
Short description: mixed chorus (without instrument). Very tonal (no attacks).
Possible problems: kind of ringing, distortions on sibilant consonant (“s”).
replaygain_sample_gain: +0.87 dB (indicative only: files were tested at their original volume)




I recognized faac, which is the only encoder to my knowledge betraying some warbling issues on this kind of (tonal) samples. In addition to this problem voice are distorted (maybe a consequence of low bitrate allocation¹). MPC has serious issue here, already noticed in the past. Voices are not natural, a bit ugly and also unstable. Less annoying are distortions audible with LAME and aoTuV. This last one seems to remove some noise/matter. Nero Digital is the best, close to perfection with problem occurring on a very small portion of the sample (and ABXed without difficulties).
¹ 157 kbps which corresponds to the faac lowest bitrate of the whole test.


Sample 04: Weihnachts-Oratorium (Oratorio de Noël)
Short description: exulting orchestra (period instruments), with brass, percussions and mixed chorus.
Possible problems: loss in details, distorted on voices or instruments (especially brass).
replaygain_sample_gain: -2.62 dB (indicative only: files were tested at their original volume)




faac and LAME are the worse and both have the same kind of distortions altering the orchestral part. Nero AAC was one step better, with minor problems on chorus part. I noticed something I associated to noise reduction. MPC and aoTuV are very good. Brass is slightly distorted with MPC which also removes some details in the background. Vorbis is even better than MPC. So good that I first rated the reference, but the notation was automatically cancelled by ABC/HR after a positive ABX test. It’s important to note that I’ve suspected Vorbis: for the first time I heard coarseness/fatness used to affect this format.



Sample 05: Dom Bedos
Short description: organ, with long continuous tonal notes
Possible problems: kind of ringing (constant tonal part being fluctuating).
replaygain_sample_gain: +5.10 dB (indicative only: files were tested at their original volume)




Nero Digital appears as the worst encoding. Distortions are ugly. The encoding of the problematic last note was better and it preserves Nero AAC from a biting notation¹. Not far from Nero Digital: faac² and LAME³. Both offers a really good sound on beginning but quality drops for each on the end of the sample. Again, MPC and aoTuV are on top. Both are excellent on the first half of this sample. But second half and especially last note is apparently much harder to encode for all encoders. MPC is here distorted. Last year, I heard something wrong but failed on ABX phase; this time, ABX was successful. aoTuV presents a different form of distortion, less annoying: something raucous, unconstant and hard to describe – subtle but not too hard to ABX. Like often with very tonal part encoded by MPC: bloated bitrate.
¹ This is exactly the kind of sample for which Nero ‘fast’ encoder is perfect (but with bloated bitrate as other side of the coin: 233 kbps for this sample with 'fast' encoder and -streaming preset against 165 kbps for the tested encoding!)
² faac bitrate is very low: 150 kbps. It’s 80 kbps less than MPC!
³ LAME quality is now much better than the one tested one year ago (ugly distortions are ringing are gone).



Sample 06: Platée
Short description: orchestra representing a rainstorm (period instruments).
Possible problems: background is detailed due to the presence of instruments ensuring continuo; loss in details is expected
replaygain_sample_gain: -0.07 dB (indicative only: files were tested at their original volume)




faac presents the most distorted sound on orchestra again (see sample_04, and later sample_10 & sample_15). Details are either softened or distorted. aoTuV occupies here an unusual last but one place. Sound is too imprecise, and edges of harpsichord are softened. LAME sounds in a similar manner, offering a bit more precisions than aoTuV but also slight distortions on continuo. I discovered the problem during ABX phase, and first rated the reference instead of encoded file. Nero Digital is very good, with a subtle loss in details and located distortions. It’s a remarkable performance for such low bitrate encoding¹. Same goes for MPC but without distortions this time: just a slight smoothing of details in background.
¹ 161 kbps, whereas all other contenders have a bitrate comprise between 194 [aoTuV] and 215 kbps [LAME]!


Sample 07: Marche Royale (00.00 – 12.00)
Short description: Chamber orchestra. First part of a sample divided in two. Here: drums, violin, trumpet, cello, clarinet.
Possible problems: pre-echo (drums) and usual distortions on instruments like violin or clarinet.
replaygain_sample_gain: +1.02 dB (indicative only: files were tested at their original volume)




Disappointing performance of Nero Digital which presents the worst pre-echo performance. The same issue also appeared on sample_01. faac has the same situation, but annoyance is clearly lower. The encoder is followed by LAME, which seems to be very slightly better, and then MPC, also (slightly) smeared with other minor problems (hollowed drums). aoTuV is the best, with very few pre-echo but rather a distortion coming from within the drums (boosted noise).


Sample 08: Marche Royale (12.00 – 29.00)
Short description: Chamber orchestra – 2nd part. Here: cymbals (orchestral ones, different from those heard in sample_O1) are introduced; I focused my rating on these.
Possible problems: smearing and distortions (swoosh).
replaygain_sample_gain: +1.02 dB (indicative only: files were tested at their original volume)




First, short remark: I ABXed everything here, because my feeling became so confuse and imprecise on ABCHR phase that I wasn’t unable anymore to structure the notation and the hierarchy.
LAME is the worse; it wins its worst note for the entire test with this instrument (cymbals). Distorted and really unpleasant. faac also presents worrying distortions (and audible smearing). Nero Digital is much less annoying but smearing as well as distortions are still moderately perceptible. Much better is MPC, harder to ABX (13/16) and barely distorted. I noticed a very small difference for aoTuV during ABCHR phase but I couldn’t confirm it on ABX module (8/16) and I therefore manually cancelled the notation (4.8) I give him first.


Sample 09: Orion II
Short description: trombones –one of the most difficult brass instrument to encode for transform encoders- here meticulously recorded. It corresponds to usual “micro-attacks” problems.
Possible problems: micro-attacks encoded with noise or -in worst case- distorted.
replaygain_sample_gain: -4.80 dB (indicative only: files were tested at their original volume)




Nero Digital is very bad. Artefact is so terrible that I gave (thoughtful decision) the lowest note (1.0) to this sample¹. LAME and faac are much better and they only present noise as defect. MPC is even better (few noise) but presents a short artefact I identified during the ABX phase (14/16). aoTuV sounded similar to MPC but after several listenings it appears as slightly cleaner and without artefact².
¹ This is the only notation inferior to 2.0 for the entire test.
² Direct blind comparison between MPC and aoTuV was easy: 12/12 pval < 0.001. It should be worth noticing that Vorbis has made clear progress in this area (micro-attacks, used to be smeared at this preset) compared to last test including MEGAMIX.




Sample 10: “Dover, giustizia”
Short description: mezzo-soprano voice accompanied by orchestra.
Possible problems: voice can be distorted and instrument may also be softened.
replaygain_sample_gain: -4.73 dB (indicative only: files were tested at their original volume)




faac gets the last place, but with a flattering notation corresponding to a small distortion on voice (3.5). This is the highest note I gave for a worst encoding. Obviously, this sample doesn’t contain anything excessively difficult for ours five competitors. Unusual place for MPC which presents distortion on voice and which smooth subtle details in the background. Then comes Nero Digital which doesn’t distort voice but which softens some details in orchestral background. But Nero Digital also clearly distorts the harpsichord on the very end. LAME is similar to Nero Digital but without the distortions noticed before on harpsichord. aoTuV is transparent to my ears (5/12) despite of its low bitrate (164 kbps).


Sample 11: Trumpet Voluntar
Short description: mezzo trumpet with organ – the recording is very noisy on loud moments only (maybe a consequence of the organ?).
Possible problems: trumpet sometimes trigger artefacts; noise may cause ringing.
replaygain_sample_gain: -4.98 dB (indicative only: files were tested at their original volume)




The inconstant noise of the reference file is very disturbing (to be honest, I don’t like this sample). I therefore ABXed everything in order to be sure to not being betray by the weird reference sound.
faac is the worse and offers distortions (trumpet) that were immediately perceptible. Nero Digital and then MPC both present slight problems with trumpet; MPC encodes trumpets with some irregularities (phenomenon already noticed in chorus/sample_03). LAME and aoTuV are really excellent. I heard a very subtle distortion with LAME on ABC/HR ABXed with 8/10. aoTuV has no distortion but it adds a small amount of noise (also noticed with sample_07).

Sample 12: Liebestod
Short description: Soprano voice and full orchestra (modern instruments) – submitted by harashin.
Possible problems: same as sample_10. Violins (in orchestra) are often difficult to encode and may introduce another encoding difficulty.
replaygain_sample_gain: -3.57 dB (indicative only: files were tested at their original volume)




Nero Digital is the worst and pay for distortions occurring in orchestral moments¹. faac offers similar performance but with less distortions on violins. MPC is better and presents some irregularities especially on voice (see sample_10 for similar problem). Note: last year I unmasked MPC on ABCHR phase but missed the ABX test. Has the encoding quality regressed or did my hearing improved a bit?
Two files are transparent to my ears (a record for this test): LAME (8/16) and aoTuV (10/16). Note: there were a clear ringing audible with alpha 3 of LAME; now it’s totally gone.
¹ I already noticed similar distortions in the past with violins/chorus with older version of Nero AAC; the new ‘fast’ encoder is used to perform better. The same problem is audible in similar conditions with sample_18.



Sample 13: LadyMacbeth
Short description: full and hysterical orchestra (modern instruments) with a lot of percussions and cymbals – submitted by harashin.
Possible problems: pre-echo is likely; distortions on other instruments (especially cymbals and brass) may also occur.
replaygain_sample_gain: -5.27 dB (indicative only: files were tested at their original volume)




LAME and Nero Digital both produce the worst sounding: cymbals are distorted and smearing is also audible. Problems occurring with faac are less disturbing, despite of an unusual form of artefact (hollowed sound on drums). MPC and aoTuV are both very good. Musepack distorts cymbals which don't sound true again (see sample_01 and to a less extend sample_08) whereas Vorbis adds noise (see sample_07 and sample_11) which is less unpleasant to my ears and also harder to detect. ABX: 8/8 for MPC and 10/12 for aoTuV.


Sample 14: Vivaldi RV93
Short description: small concertino performed on period instruments.
Possible problems: on mandolin, softened or distorted harpsichord in continuo.
replaygain_sample_gain: -1.32 dB (indicative only: files were tested at their original volume)




LAME is worst (problems are some distortions and a lack of precision) followed by faac and then Nero Digital which is –by far- the most efficient encoding¹. aoTuV has something strange: a kind of mix between thinness and fatness. The fatness issue also occurred with sample_04. MPC was transparent: I rated the reference and missed by hair’s breadth the ABX test (pval=0.054).
¹ 168 kbps against 201 kbps (vorbis) and 213 (mpc) which are better, and 193 (faac) and 187 (mp3) which are worse.


Sample 15: Troisème Ballet
Short description: Concerto for bagpipe played on period instruments and including some others rare and rustic instruments (like hurdy-gurdy). - See this page for more informations about these old instruments.
Possible problems: ringing for bagpipe, lack of details for hurdy-gurdy.
replaygain_sample_gain: +3.21 dB (indicative only: files were tested at their original volume)




Most encoders are able to encode this sample perfectly or very limited problems. Only exception is faac, which distorts the bagpipe and which also adds a slight warbling, typical of this encoder (see sample_03 for similar artefact)¹. Nero Digital and MPC are both² close to perfection to my ears and they only fail (very slightly) on tonal notes of the bagpipe (10/12 for MPC³ and 9/12 for Nero Digital). LAME and aoTuV are both perfect to my ears (10/16 & 5/12).
¹ It might be interesting to link this poor performance to the consummed bitrate: faac used 224 kbps for this short sample (the highest one for faac on 18 samples, and also higher than other competitors for this sample).
² But Nero Digital is by far the most efficient, using 154 kbps instead of 213 kbps for Musepack
³ As for sample_12 I succeed in ABXing MPC 1.15v whereas 1.14 was transparent last year: did my ears become more sensitive or has 1.15 regressed in some areas?


Sample 16: basson
Short description: solo bassoon with light instrumental accompaniment.
Possible problems: micro-attacks on one moment may be softened or distorted.
replaygain_sample_gain: +0.68 dB (indicative only: files were tested at their original volume)




There were no micro-attacks issue (less stronger than those produced by trombone on sample_09. But encoding problems occur elsewhere -and for all encoders- : during one precise range when bassoon is playing at higher volume. It may be interesting to note that all encoders also allocate a lower bitrate than average (150…160 kbps and 175 kbps for MPC).
faac presents the most distorted rendering. The encoder is followed by aoTuV which presents this time (I would say another time… ¹) a fat texture on the described range; this is slightly irritating. The three remaining encoders present non-annoying distortions: Nero Digital, LAME (slightly better) and MPC (even better²).
¹ Same problems on sample_04 and sample_14. Last year, MEGAMIX also obtained an unusually low note on this sample (and exactly the same range).
² ABX comparison opposing directly LAME to MPC: 11/12



Sample 17: Seminarist
Short description: baritone voice accompanied by piano; the sample contains a lot of sibilant consonants.
Possible problems: there’s a big risk of distortion on sibilant.
replaygain_sample_gain: -0.45 dB (indicative only: files were tested at their original volume)




faac is not really good and is betrayed by distortions on voice. MPC has presents distortions on sibilant but also on tonal part of voice. Musepack is now used to have problems with voice¹. Nero Digital is slightly better and is followed by LAME; both only show problem on sibilants, moderately annoying (reference is, as I already wrote it during the test, very aggressive). aoTuV is with MPC the one which offers the lesser problem on sibilant, but with a perfect quality on tonal part of the voice. It’s a nice progress since last year².
¹ See comments for sample_03 (chorus), sample_10 (mezzo-soprano) and sample_12 (soprano). Now it’s the turn to male voice to trigger distortion.
² Vorbis MEGAMIX obtained its worst note (2.3) and was clearly worse than LAME 3.97a3; Vorbis aoTuV beta 4 now corrects all problems that were previously audible on voice.



Sample 18: Butterfly Lovers
Short description:violin concerto. Violin is playing alone, then is intertwined with the orchestra
Possible problems: like Liebestod, orchestral strings may be risky for some encoders. Blending between violin and violins is possible.
replaygain_sample_gain: +3.34 dB (indicative only: files were tested at their original volume)




Nero Digital produces an irritating sound. It’s a consequence of encoding problems with violins, already heard and commented for sample_12. MPC offers a similar quality. It sounded fat & coarse, and I was ready to bet that this encoding corresponded to Vorbis, former specialist of this typical problem… faac is in comparison much better. I had nothing but “warm fuzzy feeling” during ABCHR phase, and it was clarified (and verified!) by ABX (12/12): minor distortions on strings. Same thing for LAME<